diff --git a/docs/source/devguide/statepoint.rst b/docs/source/devguide/statepoint.rst
index 2b486f12e0..9862a45282 100644
--- a/docs/source/devguide/statepoint.rst
+++ b/docs/source/devguide/statepoint.rst
@@ -4,6 +4,544 @@
 State Point Binary File Specifications
 ======================================
 
+-----------
+Revision 10 
+-----------
+
+**integer(4) FILETYPE_STATEPOINT**
+
+    Flags whether this file is a statepoint file or a particle restart file.
+
+**integer(4) REVISION_STATEPOINT**
+
+    Revision of the binary state point file. Any time a change is made in the
+    format of the state-point file, this integer is incremented.
+
+**integer(4) VERSION_MAJOR**
+
+    Major version number for OpenMC
+
+**integer(4) VERSION_MINOR**
+
+    Minor version number for OpenMC
+
+**integer(4) VERSION_RELEASE**
+
+    Release version number for OpenMC
+
+**character(19) time_stamp**
+
+    Date and time the state point was written.
+
+**character(255) path**
+
+    Absolute path to directory containing input files.
+
+**integer(8) seed**
+
+    Pseudo-random number generator seed.
+
+**integer(4) run_mode**
+
+    run mode used. The modes are described in constants.F90.
+
+**integer(8) n_particles**
+
+    Number of particles used per generation.
+
+**integer(4) n_batches**
+
+    Total number of batches (active + inactive).
+
+**integer(4) current_batch**
+
+    The number of batches already simulated.
+
+if (run_mode == MODE_EIGENVALUE)
+
+    **integer(4) n_inactive**
+
+        Number of inactive batches
+
+    **integer(4) gen_per_batch**
+
+        Number of generations per batch for criticality calculations
+
+    *do i = 1, current_batch \* gen_per_batch*
+
+        **real(8) k_generation(i)**
+
+             k-effective for the i-th total generation
+
+    *do i = 1, current_batch \* gen_per_batch*
+
+        **real(8) entropy(i)**
+
+            Shannon entropy for the i-th total generation
+
+    **real(8) k_col_abs**
+
+        Sum of product of collision/absorption estimates of k-effective
+
+    **real(8) k_col_tra**
+
+        Sum of product of collision/track-length estimates of k-effective
+
+    **real(8) k_abs_tra**
+
+        Sum of product of absorption/track-length estimates of k-effective
+
+    **real(8) k_combined(2)**
+
+        Mean and standard deviation of a combined estimate of k-effective
+
+    **integer(4) cmfd_on**
+
+        Flag that cmfd is on
+
+    if (cmfd_on)
+
+        **integer(4) cmfd % indices**
+
+            Indices for cmfd mesh (i,j,k,g)
+
+        **real(8) cmfd % k_cmfd(1:current_batch)**
+
+            CMFD eigenvalues
+
+        **real(8) cmfd % src(1:I,1:J,1:K,1:G)**
+
+            CMFD fission source
+
+        **real(8) cmfd % entropy(1:current_batch)**
+
+            CMFD estimate of Shannon entropy
+
+        **real(8) cmfd % balance(1:current_batch)**
+
+            RMS of the residual neutron balance equation on CMFD mesh
+
+        **real(8) cmfd % dom(1:current_batch)**
+
+            CMFD estimate of dominance ratio
+
+        **real(8) cmfd % scr_cmp(1:current_batch)**
+
+            RMS comparison of difference between OpenMC and CMFD fission source
+
+**integer(4) n_meshes**
+
+    Number of meshes in tallies.xml file
+
+*do i = 1, n_meshes*
+
+    **integer(4) meshes(i) % id**
+
+        Unique ID of mesh.
+
+    **integer(4) meshes(i) % type**
+
+        Type of mesh.
+
+    **integer(4) meshes(i) % n_dimension**
+
+        Number of dimensions for mesh (2 or 3).
+
+    **integer(4) meshes(i) % dimension(:)**
+
+        Number of mesh cells in each dimension.
+
+    **real(8) meshes(i) % lower_left(:)**
+
+        Coordinates of lower-left corner of mesh.
+
+    **real(8) meshes(i) % upper_right(:)**
+
+        Coordinates of upper-right corner of mesh.
+
+    **real(8) meshes(i) % width(:)**
+
+        Width of each mesh cell in each dimension.
+
+**integer(4) n_tallies**
+
+*do i = 1, n_tallies*
+
+    **integer(4) tallies(i) % id**
+
+        Unique ID of tally.
+
+    **integer(4) tallies(i) % n_realizations**
+
+        Number of realizations for the i-th tally.
+
+    **integer(4) size(tallies(i) % scores, 1)**
+
+        Total number of score bins for the i-th tally
+
+    **integer(4) size(tallies(i) % scores, 2)**
+
+        Total number of filter bins for the i-th tally
+
+    **integer(4) tallies(i) % n_filters**
+
+    *do j = 1, tallies(i) % n_filters*
+
+        **integer(4) tallies(i) % filter(j) % type**
+
+            Type of tally filter.
+
+        **integer(4) tallies(i) % filter(j) % n_bins**
+
+            Number of bins for filter.
+
+        **integer(4)/real(8) tallies(i) % filter(j) % bins(:)**
+
+            Value for each filter bin of this type.
+
+    **integer(4) tallies(i) % n_nuclide_bins**
+
+        Number of nuclide bins. If none are specified, this is just one.
+
+    *do j = 1, tallies(i) % n_nuclide_bins*
+
+        **integer(4) tallies(i) % nuclide_bins(j)**
+
+            Values of specified nuclide bins
+
+    **integer(4) tallies(i) % n_score_bins**
+
+        Number of scoring bins.
+
+    *do j = 1, tallies(i) % n_score_bins*
+
+        **integer(4) tallies(i) % score_bins(j)**
+
+            Values of specified scoring bins (e.g. SCORE_FLUX).
+            
+    *do j = 1, tallies(i) % n_score_bins*
+
+        **integer(4) tallies(i) % scatt_order(j)**
+
+            Scattering Order specified scoring bins.
+            
+    **integer(4) tallies(i) % n_score_bins**
+
+        Number of scoring bins without accounting for those added by
+        the scatter-pn command.
+
+**integer(4) n_realizations**
+
+    Number of realizations for global tallies.
+
+**integer(4) N_GLOBAL_TALLIES**
+
+    Number of global tally scores
+
+*do i = 1, N_GLOBAL_TALLIES*
+
+    **real(8) global_tallies(i) % sum**
+
+        Accumulated sum for the i-th global tally
+
+    **real(8) global_tallies(i) % sum_sq**
+
+        Accumulated sum of squares for the i-th global tally
+
+**integer(4) tallies_on**
+
+    Flag indicated if tallies are present in the file.
+
+if (tallies_on > 0)
+
+    *do i = 1, n_tallies*
+
+        *do k = 1, size(tallies(i) % scores, 2)*
+
+            *do j = 1, size(tallies(i) % scores, 1)*
+
+                **real(8) tallies(i) % scores(j,k) % sum**
+            
+                    Accumulated sum for the j-th score and k-th filter of the
+                    i-th tally
+
+                **real(8) tallies(i) % scores(j,k) % sum_sq**
+
+                    Accumulated sum of squares for the j-th score and k-th
+                    filter of the i-th tally
+
+if (run_mode == MODE_EIGENVALUE)
+
+    *do i = 1, n_particles*
+
+        **real(8) source_bank(i) % wgt**
+
+            Weight of the i-th source particle
+
+        **real(8) source_bank(i) % xyz(1:3)**
+
+            Coordinates of the i-th source particle.
+
+        **real(8) source_bank(i) % uvw(1:3)**
+
+            Direction of the i-th source particle
+
+        **real(8) source_bank(i) % E**
+
+            Energy of the i-th source particle.
+
+----------
+Revision 9
+----------
+
+**integer(4) FILETYPE_STATEPOINT**
+
+    Flags whether this file is a statepoint file or a particle restart file.
+
+**integer(4) REVISION_STATEPOINT**
+
+    Revision of the binary state point file. Any time a change is made in the
+    format of the state-point file, this integer is incremented.
+
+**integer(4) VERSION_MAJOR**
+
+    Major version number for OpenMC
+
+**integer(4) VERSION_MINOR**
+
+    Minor version number for OpenMC
+
+**integer(4) VERSION_RELEASE**
+
+    Release version number for OpenMC
+
+**character(19) time_stamp**
+
+    Date and time the state point was written.
+
+**character(255) path**
+
+    Absolute path to directory containing input files.
+
+**integer(8) seed**
+
+    Pseudo-random number generator seed.
+
+**integer(4) run_mode**
+
+    run mode used. The modes are described in constants.F90.
+
+**integer(8) n_particles**
+
+    Number of particles used per generation.
+
+**integer(4) n_batches**
+
+    Total number of batches (active + inactive).
+
+**integer(4) current_batch**
+
+    The number of batches already simulated.
+
+if (run_mode == MODE_EIGENVALUE)
+
+    **integer(4) n_inactive**
+
+        Number of inactive batches
+
+    **integer(4) gen_per_batch**
+
+        Number of generations per batch for criticality calculations
+
+    *do i = 1, current_batch \* gen_per_batch*
+
+        **real(8) k_generation(i)**
+
+             k-effective for the i-th total generation
+
+    *do i = 1, current_batch \* gen_per_batch*
+
+        **real(8) entropy(i)**
+
+            Shannon entropy for the i-th total generation
+
+    **real(8) k_col_abs**
+
+        Sum of product of collision/absorption estimates of k-effective
+
+    **real(8) k_col_tra**
+
+        Sum of product of collision/track-length estimates of k-effective
+
+    **real(8) k_abs_tra**
+
+        Sum of product of absorption/track-length estimates of k-effective
+
+    **real(8) k_combined(2)**
+
+        Mean and standard deviation of a combined estimate of k-effective
+
+**integer(4) n_meshes**
+
+    Number of meshes in tallies.xml file
+
+*do i = 1, n_meshes*
+
+    **integer(4) meshes(i) % id**
+
+        Unique ID of mesh.
+
+    **integer(4) meshes(i) % type**
+
+        Type of mesh.
+
+    **integer(4) meshes(i) % n_dimension**
+
+        Number of dimensions for mesh (2 or 3).
+
+    **integer(4) meshes(i) % dimension(:)**
+
+        Number of mesh cells in each dimension.
+
+    **real(8) meshes(i) % lower_left(:)**
+
+        Coordinates of lower-left corner of mesh.
+
+    **real(8) meshes(i) % upper_right(:)**
+
+        Coordinates of upper-right corner of mesh.
+
+    **real(8) meshes(i) % width(:)**
+
+        Width of each mesh cell in each dimension.
+
+**integer(4) n_tallies**
+
+*do i = 1, n_tallies*
+
+    **integer(4) tallies(i) % id**
+
+        Unique ID of tally.
+
+    **integer(4) tallies(i) % n_realizations**
+
+        Number of realizations for the i-th tally.
+
+    **integer(4) size(tallies(i) % scores, 1)**
+
+        Total number of score bins for the i-th tally
+
+    **integer(4) size(tallies(i) % scores, 2)**
+
+        Total number of filter bins for the i-th tally
+
+    **integer(4) tallies(i) % n_filters**
+
+    *do j = 1, tallies(i) % n_filters*
+
+        **integer(4) tallies(i) % filter(j) % type**
+
+            Type of tally filter.
+
+        **integer(4) tallies(i) % filter(j) % n_bins**
+
+            Number of bins for filter.
+
+        **integer(4)/real(8) tallies(i) % filter(j) % bins(:)**
+
+            Value for each filter bin of this type.
+
+    **integer(4) tallies(i) % n_nuclide_bins**
+
+        Number of nuclide bins. If none are specified, this is just one.
+
+    *do j = 1, tallies(i) % n_nuclide_bins*
+
+        **integer(4) tallies(i) % nuclide_bins(j)**
+
+            Values of specified nuclide bins
+
+    **integer(4) tallies(i) % n_score_bins**
+
+        Number of scoring bins.
+
+    *do j = 1, tallies(i) % n_score_bins*
+
+        **integer(4) tallies(i) % score_bins(j)**
+
+            Values of specified scoring bins (e.g. SCORE_FLUX).
+            
+    *do j = 1, tallies(i) % n_score_bins*
+
+        **integer(4) tallies(i) % scatt_order(j)**
+
+            Scattering Order specified scoring bins.
+            
+    **integer(4) tallies(i) % n_score_bins**
+
+        Number of scoring bins without accounting for those added by
+        the scatter-pn command.
+
+**integer(4) n_realizations**
+
+    Number of realizations for global tallies.
+
+**integer(4) N_GLOBAL_TALLIES**
+
+    Number of global tally scores
+
+*do i = 1, N_GLOBAL_TALLIES*
+
+    **real(8) global_tallies(i) % sum**
+
+        Accumulated sum for the i-th global tally
+
+    **real(8) global_tallies(i) % sum_sq**
+
+        Accumulated sum of squares for the i-th global tally
+
+**integer(4) tallies_on**
+
+    Flag indicated if tallies are present in the file.
+
+if (tallies_on > 0)
+
+    *do i = 1, n_tallies*
+
+        *do k = 1, size(tallies(i) % scores, 2)*
+
+            *do j = 1, size(tallies(i) % scores, 1)*
+
+                **real(8) tallies(i) % scores(j,k) % sum**
+            
+                    Accumulated sum for the j-th score and k-th filter of the
+                    i-th tally
+
+                **real(8) tallies(i) % scores(j,k) % sum_sq**
+
+                    Accumulated sum of squares for the j-th score and k-th
+                    filter of the i-th tally
+
+if (run_mode == MODE_EIGENVALUE)
+
+    *do i = 1, n_particles*
+
+        **real(8) source_bank(i) % wgt**
+
+            Weight of the i-th source particle
+
+        **real(8) source_bank(i) % xyz(1:3)**
+
+            Coordinates of the i-th source particle.
+
+        **real(8) source_bank(i) % uvw(1:3)**
+
+            Direction of the i-th source particle
+
+        **real(8) source_bank(i) % E**
+
+            Energy of the i-th source particle.
+
 ----------
 Revision 8
 ----------
diff --git a/docs/source/usersguide/input.rst b/docs/source/usersguide/input.rst
index 425df5029a..9ec51ca21e 100644
--- a/docs/source/usersguide/input.rst
+++ b/docs/source/usersguide/input.rst
@@ -1139,6 +1139,22 @@ The ``<begin>`` element controls what batch CMFD calculations should begin.
 
   *Default*: 1
 
+``<display>`` Element
+---------------------
+
+The ``<display>`` element sets one additional CMFD output column. Options are:
+
+* "balance" - prints the RMS [%] of the resdiual from the neutron balance equation
+  on CMFD tallies.
+* "dominance" - prints the estimated dominance ratio from the CMFD iterations.
+  **This will only work for power iteration eigensolver**.
+* "entropy" - prints the *entropy* of the CMFD predicted fission source.
+  **Can only be used if OpenMC entropy is active as well**.
+* "source" - prints the RMS [%] between the OpenMC fission source and CMFD
+  fission source.
+
+  *Default*: None
+
 ``<feedback>`` Element
 ----------------------
 
@@ -1158,14 +1174,14 @@ with "true" and off with "false"
 
   *Default*: true
 
-``<keff_tol>`` Element
-----------------------
+``<inactive_flush>`` Element
+----------------------------
 
-The ``<keff_tol>`` element specifies acceptance criteria of a CMFD eigenvalue.
-If the CMFD eigenvalue and OpenMC batch eigenvalue are within this tolerance, 
-CMFD is allowed to modify source neutron weights. 
+The ``<inactive_flush>`` element controls when CMFD tallies are reset during
+inactive batches. The integer set here is the interval at which this reset
+occurs. The amout of resets is controlled with the ``<num_flushes>`` element.
 
-  *Default*: 0.005
+  *Defualt*: 9999
 
 ``<ksp_monitor>`` Element
 -------------------------
@@ -1249,14 +1265,13 @@ not impact the calculation.
 
   *Default*: 1.0
 
-``<n_procs_cmfd>`` Element
---------------------------
+``<num_flushes>`` Element
+-------------------------
 
-The ``<n_procs_cmfd>`` element is used to set the number of processors used 
-for CMFD calculation. It should be less than or equal to the number of 
-processors used during OpenMC. 
+The ``<num_flushes>`` element controls the number of CMFD tally resets that
+occur during inactive CMFD batches.
 
-  *Default*: 1
+  *Default*: 9999
 
 ``<power_monitor>`` Element
 ---------------------------
@@ -1264,26 +1279,33 @@ processors used during OpenMC.
 The ``<power_monitor>`` element is used to view the convergence of power iteration. 
 This option can be turned on with "true" and turned off with "false".
 
+  *Default*: false
+
+``<run_adjoint>`` Element
+-------------------------
+
+The ``<run_adjoint>`` element can be turned on with "true" to have an adjoint
+calculation be performed on the last batch when CMFD is active.
 
   *Default*: false
 
-``<write_balance>`` Element
----------------------------
+``<snes_monitor>`` Element
+--------------------------
 
-The ``<write_balance>`` element is used to view the balance of OpenMC tally
-residuals for every coarse mesh region and energy group. This option can be 
-turned on with "true" and off with "false". 
+The ``<snes_monitor>`` element is used to view the convergence of the nonlinear SNES
+function in PETSc. This option can be turned on with "true" and turned off with "false".
 
 
   *Default*: false
 
-``<write_hdf5>`` Element
-------------------------
+``<solver>`` Element
+--------------------
 
-The ``<write_hdf5>`` element can be turned on with "true" to get an 
-HDF5 output file of CMFD results. 
+The ``<solver>`` element controls whether the CMFD eigenproblem is solved with
+standard power iteration or nonlinear Jacobian-free Newton Krylov (JFNK). 
+By setting "power", power iteration is used and by setting "jfnk", JFNK is used.
 
-  *Default*: false
+  *Default*: power
 
 ``<write_matrices>`` Element
 ----------------------------
diff --git a/docs/source/usersguide/install.rst b/docs/source/usersguide/install.rst
index 159c689e98..84f7633c2f 100644
--- a/docs/source/usersguide/install.rst
+++ b/docs/source/usersguide/install.rst
@@ -71,13 +71,29 @@ Prerequisites
       To compile with support for HDF5_ output (highly recommended), you will
       need to have HDF5 installed on your computer. The installed version will
       need to have been compiled with the same compiler you intend to compile
-      OpenMC with.
+      OpenMC with. HDF5_ must be built with parallel I/O features if you intend
+      to use HDF5_ with MPI. An example of configuring HDF5_ is listed below::
+
+           FC=/opt/mpich/3.0.4-gnu/bin/mpif90 CC=/opt/mpich/3.0.4-gnu/bin/mpicc \
+           ./configure --prefix=/opt/hdf5/1.8.11-gnu --enable-fortran \
+                       --enable-fortran2003 --enable-parallel
+
+      You may omit '--enable-parallel' if you want to compile HDF5_ in serial.
 
     * PETSc_ for CMFD acceleration
 
-      To enable CMFD acceleration, you will need to have PETSc_ installed on
-      your computer. The installed version will need to have been compiled with
-      the same compiler you intend to compile OpenMC with.
+      To enable CMFD acceleration, you will need to have PETSc_ (3.4.2 or higher)
+      installed on your computer. The installed version will need to have been
+      compiled with the same compiler you intend to compile OpenMC with. OpenMC
+      requires PETSc_ to be configured with Fortran datatypes. An example of
+      configuring PETSc_ is listed below::
+
+           ./configure --prefix=/opt/petsc/3.4.2-gnu --download-f-blas-lapack \
+                       --with-mpi-dir=/opt/mpich/3.0.4-gnu/ --with-shared-libraries=0 \
+                       --with-fortran-datatypes
+
+      The BLAS/LAPACK library is not required to be downloaded and can be linked
+      explicitly (e.g., Intel MLK library).
 
     * git_ version control software for obtaining source code
 
diff --git a/src/DEPENDENCIES b/src/DEPENDENCIES
index 59b68e7c74..eeb538fc16 100644
--- a/src/DEPENDENCIES
+++ b/src/DEPENDENCIES
@@ -1,37 +1,29 @@
-set_header.o: constants.o
-set_header.o: list_header.o
+ace.o: ace_header.o
+ace.o: constants.o
+ace.o: endf.o
+ace.o: error.o
+ace.o: fission.o
+ace.o: global.o
+ace.o: material_header.o
+ace.o: output.o
+ace.o: set_header.o
+ace.o: string.o
 
-energy_grid.o: constants.o
-energy_grid.o: global.o
-energy_grid.o: list_header.o
-energy_grid.o: output.o
+ace_header.o: constants.o
+ace_header.o: endf_header.o
 
-list_header.o: constants.o
-
-cmfd_slepc_solver.o: cmfd_loss_operator.o
-cmfd_slepc_solver.o: cmfd_prod_operator.o
-cmfd_slepc_solver.o: constants.o
-cmfd_slepc_solver.o: global.o
-
-cmfd_loss_operator.o: constants.o
-cmfd_loss_operator.o: global.o
-
-particle_restart.o: bank_header.o
-particle_restart.o: constants.o
-particle_restart.o: geometry_header.o
-particle_restart.o: global.o
-particle_restart.o: output.o
-particle_restart.o: output_interface.o
-particle_restart.o: particle_header.o
-particle_restart.o: random_lcg.o
-particle_restart.o: tracking.o
-
-doppler.o: constants.o
+cmfd_data.o: cmfd_header.o
+cmfd_data.o: constants.o
+cmfd_data.o: error.o
+cmfd_data.o: global.o
+cmfd_data.o: mesh.o
+cmfd_data.o: mesh_header.o
+cmfd_data.o: string.o
+cmfd_data.o: tally_header.o
 
 cmfd_execute.o: cmfd_data.o
-cmfd_execute.o: cmfd_message_passing.o
+cmfd_execute.o: cmfd_jfnk_solver.o
 cmfd_execute.o: cmfd_power_solver.o
-cmfd_execute.o: cmfd_snes_solver.o
 cmfd_execute.o: constants.o
 cmfd_execute.o: error.o
 cmfd_execute.o: global.o
@@ -41,27 +33,99 @@ cmfd_execute.o: output.o
 cmfd_execute.o: search.o
 cmfd_execute.o: tally.o
 
+cmfd_header.o: constants.o
+
+cmfd_input.o: cmfd_header.o
+cmfd_input.o: error.o
+cmfd_input.o: global.o
+cmfd_input.o: mesh_header.o
+cmfd_input.o: output.o
+cmfd_input.o: string.o
+cmfd_input.o: tally.o
+cmfd_input.o: tally_header.o
+cmfd_input.o: tally_initialize.o
+cmfd_input.o: templates/cmfd_t.o
+
+cmfd_jfnk_solver.o: cmfd_loss_operator.o
+cmfd_jfnk_solver.o: cmfd_power_solver.o
+cmfd_jfnk_solver.o: cmfd_prod_operator.o
+cmfd_jfnk_solver.o: constants.o
+cmfd_jfnk_solver.o: global.o
+cmfd_jfnk_solver.o: matrix_header.o
+cmfd_jfnk_solver.o: solver_interface.o
+cmfd_jfnk_solver.o: vector_header.o
+
+cmfd_loss_operator.o: constants.o
+cmfd_loss_operator.o: global.o
+cmfd_loss_operator.o: matrix_header.o
+
 cmfd_power_solver.o: cmfd_loss_operator.o
 cmfd_power_solver.o: cmfd_prod_operator.o
 cmfd_power_solver.o: constants.o
 cmfd_power_solver.o: global.o
-cmfd_power_solver.o: string.o
+cmfd_power_solver.o: matrix_header.o
+cmfd_power_solver.o: solver_interface.o
+cmfd_power_solver.o: vector_header.o
 
-cmfd_message_passing.o: cmfd_header.o
-cmfd_message_passing.o: global.o
+cmfd_prod_operator.o: constants.o
+cmfd_prod_operator.o: global.o
+cmfd_prod_operator.o: matrix_header.o
 
-random_lcg.o: global.o
+cmfd_slepc_solver.o: cmfd_loss_operator.o
+cmfd_slepc_solver.o: cmfd_prod_operator.o
+cmfd_slepc_solver.o: constants.o
+cmfd_slepc_solver.o: global.o
 
-plot.o: constants.o
-plot.o: error.o
-plot.o: geometry.o
-plot.o: geometry_header.o
-plot.o: global.o
-plot.o: output.o
-plot.o: particle_header.o
-plot.o: plot_header.o
-plot.o: ppmlib.o
-plot.o: string.o
+cross_section.o: ace_header.o
+cross_section.o: constants.o
+cross_section.o: error.o
+cross_section.o: fission.o
+cross_section.o: global.o
+cross_section.o: material_header.o
+cross_section.o: particle_header.o
+cross_section.o: random_lcg.o
+cross_section.o: search.o
+
+doppler.o: constants.o
+
+eigenvalue.o: cmfd_execute.o
+eigenvalue.o: constants.o
+eigenvalue.o: error.o
+eigenvalue.o: global.o
+eigenvalue.o: math.o
+eigenvalue.o: mesh.o
+eigenvalue.o: mesh_header.o
+eigenvalue.o: output.o
+eigenvalue.o: particle_header.o
+eigenvalue.o: random_lcg.o
+eigenvalue.o: search.o
+eigenvalue.o: source.o
+eigenvalue.o: state_point.o
+eigenvalue.o: string.o
+eigenvalue.o: tally.o
+eigenvalue.o: tracking.o
+
+endf.o: constants.o
+endf.o: string.o
+
+energy_grid.o: constants.o
+energy_grid.o: global.o
+energy_grid.o: list_header.o
+energy_grid.o: output.o
+
+error.o: global.o
+
+finalize.o: global.o
+finalize.o: hdf5_interface.o
+finalize.o: output.o
+finalize.o: tally.o
+
+fission.o: ace_header.o
+fission.o: constants.o
+fission.o: error.o
+fission.o: global.o
+fission.o: interpolation.o
+fission.o: search.o
 
 fixed_source.o: constants.o
 fixed_source.o: global.o
@@ -74,114 +138,15 @@ fixed_source.o: string.o
 fixed_source.o: tally.o
 fixed_source.o: tracking.o
 
-source.o: bank_header.o
-source.o: constants.o
-source.o: error.o
-source.o: geometry_header.o
-source.o: global.o
-source.o: math.o
-source.o: output.o
-source.o: particle_header.o
-source.o: random_lcg.o
-source.o: string.o
-
-cmfd_prod_operator.o: constants.o
-cmfd_prod_operator.o: global.o
-
-ace_header.o: constants.o
-ace_header.o: endf_header.o
-
-fission.o: ace_header.o
-fission.o: constants.o
-fission.o: error.o
-fission.o: global.o
-fission.o: interpolation.o
-fission.o: search.o
-
-cmfd_jacobian_operator.o: cmfd_loss_operator.o
-cmfd_jacobian_operator.o: cmfd_prod_operator.o
-cmfd_jacobian_operator.o: constants.o
-cmfd_jacobian_operator.o: global.o
-
-cmfd_snes_solver.o: cmfd_jacobian_operator.o
-cmfd_snes_solver.o: cmfd_loss_operator.o
-cmfd_snes_solver.o: cmfd_power_solver.o
-cmfd_snes_solver.o: cmfd_prod_operator.o
-cmfd_snes_solver.o: constants.o
-cmfd_snes_solver.o: global.o
-cmfd_snes_solver.o: string.o
-
-cmfd_input.o: cmfd_message_passing.o
-cmfd_input.o: error.o
-cmfd_input.o: global.o
-cmfd_input.o: mesh_header.o
-cmfd_input.o: output.o
-cmfd_input.o: string.o
-cmfd_input.o: tally.o
-cmfd_input.o: tally_header.o
-cmfd_input.o: tally_initialize.o
-cmfd_input.o: templates/cmfd_t.o
-
-input_xml.o: cmfd_input.o
-input_xml.o: constants.o
-input_xml.o: dict_header.o
-input_xml.o: error.o
-input_xml.o: geometry_header.o
-input_xml.o: global.o
-input_xml.o: list_header.o
-input_xml.o: mesh_header.o
-input_xml.o: output.o
-input_xml.o: plot_header.o
-input_xml.o: random_lcg.o
-input_xml.o: string.o
-input_xml.o: tally_header.o
-input_xml.o: tally_initialize.o
-input_xml.o: templates/cross_sections_t.o
-input_xml.o: templates/geometry_t.o
-input_xml.o: templates/materials_t.o
-input_xml.o: templates/plots_t.o
-input_xml.o: templates/settings_t.o
-input_xml.o: templates/tallies_t.o
-
-main.o: constants.o
-main.o: eigenvalue.o
-main.o: finalize.o
-main.o: fixed_source.o
-main.o: global.o
-main.o: initialize.o
-main.o: particle_restart.o
-main.o: plot.o
-
-particle_restart_write.o: bank_header.o
-particle_restart_write.o: global.o
-particle_restart_write.o: output_interface.o
-particle_restart_write.o: particle_header.o
-particle_restart_write.o: string.o
-
-timer_header.o: constants.o
-
-math.o: constants.o
-math.o: random_lcg.o
-
-interpolation.o: constants.o
-interpolation.o: endf_header.o
-interpolation.o: error.o
-interpolation.o: global.o
-interpolation.o: search.o
-interpolation.o: string.o
-
-cmfd_data.o: cmfd_header.o
-cmfd_data.o: constants.o
-cmfd_data.o: error.o
-cmfd_data.o: global.o
-cmfd_data.o: mesh.o
-cmfd_data.o: mesh_header.o
-cmfd_data.o: tally_header.o
-
-cmfd_output.o: cmfd_data.o
-cmfd_output.o: cmfd_header.o
-cmfd_output.o: constants.o
-cmfd_output.o: global.o
+geometry.o: constants.o
+geometry.o: error.o
+geometry.o: geometry_header.o
+geometry.o: global.o
+geometry.o: output.o
+geometry.o: particle_header.o
+geometry.o: particle_restart_write.o
+geometry.o: string.o
+geometry.o: tally.o
 
 global.o: ace_header.o
 global.o: bank_header.o
@@ -198,51 +163,17 @@ global.o: source_header.o
 global.o: tally_header.o
 global.o: timer_header.o
 
-string.o: constants.o
-string.o: error.o
-string.o: global.o
-
-finalize.o: cmfd_output.o
-finalize.o: global.o
-finalize.o: hdf5_interface.o
-finalize.o: output.o
-finalize.o: tally.o
-
-tally_initialize.o: constants.o
-tally_initialize.o: global.o
-tally_initialize.o: tally_header.o
-
-tally.o: ace_header.o
-tally.o: constants.o
-tally.o: error.o
-tally.o: global.o
-tally.o: math.o
-tally.o: mesh.o
-tally.o: mesh_header.o
-tally.o: output.o
-tally.o: particle_header.o
-tally.o: search.o
-tally.o: string.o
-tally.o: tally_header.o
-
-particle_header.o: constants.o
-particle_header.o: geometry_header.o
-
-output_interface.o: constants.o
-output_interface.o: error.o
-output_interface.o: global.o
-output_interface.o: hdf5_interface.o
-output_interface.o: mpiio_interface.o
-output_interface.o: tally_header.o
-
-mesh.o: constants.o
-mesh.o: global.o
-mesh.o: mesh_header.o
-mesh.o: particle_header.o
-mesh.o: search.o
-
-endf.o: constants.o
-endf.o: string.o
+hdf5_summary.o: ace_header.o
+hdf5_summary.o: constants.o
+hdf5_summary.o: endf.o
+hdf5_summary.o: geometry_header.o
+hdf5_summary.o: global.o
+hdf5_summary.o: material_header.o
+hdf5_summary.o: mesh_header.o
+hdf5_summary.o: output.o
+hdf5_summary.o: output_interface.o
+hdf5_summary.o: string.o
+hdf5_summary.o: tally_header.o
 
 initialize.o: ace.o
 initialize.o: bank_header.o
@@ -265,97 +196,56 @@ initialize.o: string.o
 initialize.o: tally_header.o
 initialize.o: tally_initialize.o
 
-cross_section.o: ace_header.o
-cross_section.o: constants.o
-cross_section.o: error.o
-cross_section.o: fission.o
-cross_section.o: global.o
-cross_section.o: material_header.o
-cross_section.o: particle_header.o
-cross_section.o: random_lcg.o
-cross_section.o: search.o
+input_xml.o: cmfd_input.o
+input_xml.o: constants.o
+input_xml.o: dict_header.o
+input_xml.o: error.o
+input_xml.o: geometry_header.o
+input_xml.o: global.o
+input_xml.o: list_header.o
+input_xml.o: mesh_header.o
+input_xml.o: output.o
+input_xml.o: plot_header.o
+input_xml.o: random_lcg.o
+input_xml.o: string.o
+input_xml.o: tally_header.o
+input_xml.o: tally_initialize.o
+input_xml.o: templates/cross_sections_t.o
+input_xml.o: templates/geometry_t.o
+input_xml.o: templates/materials_t.o
+input_xml.o: templates/plots_t.o
+input_xml.o: templates/settings_t.o
+input_xml.o: templates/tallies_t.o
 
-state_point.o: constants.o
-state_point.o: error.o
-state_point.o: global.o
-state_point.o: output.o
-state_point.o: output_interface.o
-state_point.o: string.o
-state_point.o: tally_header.o
+interpolation.o: constants.o
+interpolation.o: endf_header.o
+interpolation.o: error.o
+interpolation.o: global.o
+interpolation.o: search.o
+interpolation.o: string.o
 
-eigenvalue.o: cmfd_execute.o
-eigenvalue.o: constants.o
-eigenvalue.o: error.o
-eigenvalue.o: global.o
-eigenvalue.o: math.o
-eigenvalue.o: mesh.o
-eigenvalue.o: mesh_header.o
-eigenvalue.o: output.o
-eigenvalue.o: particle_header.o
-eigenvalue.o: random_lcg.o
-eigenvalue.o: search.o
-eigenvalue.o: source.o
-eigenvalue.o: state_point.o
-eigenvalue.o: string.o
-eigenvalue.o: tally.o
-eigenvalue.o: tracking.o
+list_header.o: constants.o
 
-search.o: error.o
-search.o: global.o
+main.o: constants.o
+main.o: eigenvalue.o
+main.o: finalize.o
+main.o: fixed_source.o
+main.o: global.o
+main.o: initialize.o
+main.o: particle_restart.o
+main.o: plot.o
 
-tracking.o: cross_section.o
-tracking.o: error.o
-tracking.o: geometry.o
-tracking.o: geometry_header.o
-tracking.o: global.o
-tracking.o: output.o
-tracking.o: particle_header.o
-tracking.o: physics.o
-tracking.o: random_lcg.o
-tracking.o: string.o
-tracking.o: tally.o
+math.o: constants.o
+math.o: random_lcg.o
 
-ace.o: ace_header.o
-ace.o: constants.o
-ace.o: endf.o
-ace.o: error.o
-ace.o: fission.o
-ace.o: global.o
-ace.o: material_header.o
-ace.o: output.o
-ace.o: set_header.o
-ace.o: string.o
+matrix_header.o: constants.o
+matrix_header.o: vector_header.o
 
-geometry.o: constants.o
-geometry.o: error.o
-geometry.o: geometry_header.o
-geometry.o: global.o
-geometry.o: output.o
-geometry.o: particle_header.o
-geometry.o: particle_restart_write.o
-geometry.o: string.o
-geometry.o: tally.o
-
-plot_header.o: constants.o
-
-cmfd_header.o: constants.o
-
-tally_header.o: constants.o
-
-hdf5_summary.o: ace_header.o
-hdf5_summary.o: constants.o
-hdf5_summary.o: endf.o
-hdf5_summary.o: geometry_header.o
-hdf5_summary.o: global.o
-hdf5_summary.o: hdf5_interface.o
-hdf5_summary.o: material_header.o
-hdf5_summary.o: mesh_header.o
-hdf5_summary.o: output.o
-hdf5_summary.o: output_interface.o
-hdf5_summary.o: string.o
-hdf5_summary.o: tally_header.o
-
-error.o: global.o
+mesh.o: constants.o
+mesh.o: global.o
+mesh.o: mesh_header.o
+mesh.o: particle_header.o
+mesh.o: search.o
 
 output.o: ace_header.o
 output.o: constants.o
@@ -371,6 +261,32 @@ output.o: plot_header.o
 output.o: string.o
 output.o: tally_header.o
 
+output_interface.o: constants.o
+output_interface.o: error.o
+output_interface.o: global.o
+output_interface.o: hdf5_interface.o
+output_interface.o: mpiio_interface.o
+output_interface.o: tally_header.o
+
+particle_header.o: constants.o
+particle_header.o: geometry_header.o
+
+particle_restart.o: bank_header.o
+particle_restart.o: constants.o
+particle_restart.o: geometry_header.o
+particle_restart.o: global.o
+particle_restart.o: output.o
+particle_restart.o: output_interface.o
+particle_restart.o: particle_header.o
+particle_restart.o: random_lcg.o
+particle_restart.o: tracking.o
+
+particle_restart_write.o: bank_header.o
+particle_restart_write.o: global.o
+particle_restart_write.o: output_interface.o
+particle_restart_write.o: particle_header.o
+particle_restart_write.o: string.o
+
 physics.o: ace_header.o
 physics.o: constants.o
 physics.o: endf.o
@@ -388,3 +304,87 @@ physics.o: random_lcg.o
 physics.o: search.o
 physics.o: string.o
 
+plot.o: constants.o
+plot.o: error.o
+plot.o: geometry.o
+plot.o: geometry_header.o
+plot.o: global.o
+plot.o: output.o
+plot.o: particle_header.o
+plot.o: plot_header.o
+plot.o: ppmlib.o
+plot.o: string.o
+
+plot_header.o: constants.o
+
+random_lcg.o: global.o
+
+search.o: error.o
+search.o: global.o
+
+set_header.o: constants.o
+set_header.o: list_header.o
+
+solver_interface.o: error.o
+solver_interface.o: global.o
+solver_interface.o: matrix_header.o
+solver_interface.o: vector_header.o
+
+source.o: bank_header.o
+source.o: constants.o
+source.o: error.o
+source.o: geometry_header.o
+source.o: global.o
+source.o: math.o
+source.o: output.o
+source.o: particle_header.o
+source.o: random_lcg.o
+source.o: string.o
+
+state_point.o: constants.o
+state_point.o: error.o
+state_point.o: global.o
+state_point.o: output.o
+state_point.o: output_interface.o
+state_point.o: string.o
+state_point.o: tally_header.o
+
+string.o: constants.o
+string.o: error.o
+string.o: global.o
+
+tally.o: ace_header.o
+tally.o: constants.o
+tally.o: error.o
+tally.o: global.o
+tally.o: math.o
+tally.o: mesh.o
+tally.o: mesh_header.o
+tally.o: output.o
+tally.o: particle_header.o
+tally.o: search.o
+tally.o: string.o
+tally.o: tally_header.o
+
+tally_header.o: constants.o
+
+tally_initialize.o: constants.o
+tally_initialize.o: global.o
+tally_initialize.o: tally_header.o
+
+timer_header.o: constants.o
+
+tracking.o: cross_section.o
+tracking.o: error.o
+tracking.o: geometry.o
+tracking.o: geometry_header.o
+tracking.o: global.o
+tracking.o: output.o
+tracking.o: particle_header.o
+tracking.o: physics.o
+tracking.o: random_lcg.o
+tracking.o: string.o
+tracking.o: tally.o
+
+vector_header.o: constants.o
+
diff --git a/src/Makefile b/src/Makefile
index a1aa11c0c5..4300d42e6e 100644
--- a/src/Makefile
+++ b/src/Makefile
@@ -42,12 +42,12 @@ GIT_SHA1 = $(shell git log -1 2>/dev/null | head -n 1 | awk '{print $$2}')
 
 ifeq ($(COMPILER),gnu)
   F90 = gfortran
-  F90FLAGS := -cpp -fbacktrace
+  F90FLAGS := -cpp -std=f2008 -fbacktrace
   LDFLAGS =
 
   # Debugging
   ifeq ($(DEBUG),yes)
-    F90FLAGS += -g -Wall -pedantic -std=f2008 -fbounds-check \
+    F90FLAGS += -g -Wall -pedantic -fbounds-check \
                 -ffpe-trap=invalid,overflow,underflow
     LDFLAGS  += -g
   endif
@@ -70,12 +70,12 @@ endif
 
 ifeq ($(COMPILER),intel)
   F90 = ifort
-  F90FLAGS := -fpp -warn -assume byterecl -traceback
+  F90FLAGS := -fpp -std08 -assume byterecl -traceback
   LDFLAGS =
 
   # Debugging
   ifeq ($(DEBUG),yes)
-    F90FLAGS += -g -ftrapuv -fp-stack-check -check all -fpe0
+    F90FLAGS += -g -warn -ftrapuv -fp-stack-check -check all -fpe0
     LDFLAGS  += -g
   endif
 
diff --git a/src/cmfd_data.F90 b/src/cmfd_data.F90
index a552b9d000..d0e80aaa90 100644
--- a/src/cmfd_data.F90
+++ b/src/cmfd_data.F90
@@ -15,34 +15,31 @@ module cmfd_data
 contains
 
 !==============================================================================
-! SET_UP_CMFD
+! SET_UP_CMFD configures cmfd object for a CMFD eigenvalue calculation
 !==============================================================================
 
   subroutine set_up_cmfd() 
 
     use cmfd_header,         only: allocate_cmfd
     use constants,           only: CMFD_NOACCEL
-    use global,              only: cmfd, cmfd_coremap, cmfd_run_2grp
+    use global,              only: cmfd, cmfd_coremap, cmfd_downscatter
 
-    ! initialize cmfd object
-    if (.not.allocated(cmfd%flux)) call allocate_cmfd(cmfd)
-
-    ! check for core map and set it up
+    ! Check for core map and set it up
     if ((cmfd_coremap) .and. (cmfd%mat_dim == CMFD_NOACCEL)) call set_coremap()
 
-    ! calculate all cross sections based on reaction rates from last batch
-    if (.not. cmfd_run_2grp) call compute_xs()
+    ! Calculate all cross sections based on reaction rates from last batch
+    call compute_xs()
 
-    ! check neutron balance
-!   call neutron_balance(670)
+    ! Compute effective downscatter cross section
+    if (cmfd_downscatter) call compute_effective_downscatter()
 
-    ! fix 2 grp cross sections
-    if (cmfd_run_2grp) call fix_2_grp()
+    ! Check neutron balance
+    call neutron_balance()
 
-    ! calculate dtilde
+    ! Calculate dtilde
     call compute_dtilde()
 
-    ! calucate dhat
+    ! Calculate dhat
     call compute_dhat()
 
   end subroutine set_up_cmfd 
@@ -55,12 +52,14 @@ contains
 
     use constants,    only: FILTER_MESH, FILTER_ENERGYIN, FILTER_ENERGYOUT,     &
                             FILTER_SURFACE, IN_RIGHT, OUT_RIGHT, IN_FRONT,      &
-                            OUT_FRONT, IN_TOP, OUT_TOP, CMFD_NOACCEL, ZERO, ONE
+                            OUT_FRONT, IN_TOP, OUT_TOP, CMFD_NOACCEL, ZERO,     &
+                            ONE, TINY_BIT
     use error,        only: fatal_error
     use global,       only: cmfd, message, n_cmfd_tallies, cmfd_tallies, meshes,&
                             matching_bins
     use mesh,         only: mesh_indices_to_bin
     use mesh_header,  only: StructuredMesh
+    use string,       only: to_str
     use tally_header, only: TallyObject
 
     integer :: nx            ! number of mesh cells in x direction
@@ -84,30 +83,30 @@ contains
     type(TallyObject),    pointer :: t => null() ! pointer for tally object
     type(StructuredMesh), pointer :: m => null() ! pointer for mesh object
 
-    ! extract spatial and energy indices from object
+    ! Extract spatial and energy indices from object
     nx = cmfd % indices(1)
     ny = cmfd % indices(2)
     nz = cmfd % indices(3)
     ng = cmfd % indices(4)
 
-    ! set flux object and source distribution to all zeros
+    ! Set flux object and source distribution to all zeros
     cmfd % flux = ZERO
     cmfd % openmc_src = ZERO
 
-    ! associate tallies and mesh
+    ! Associate tallies and mesh
     t => cmfd_tallies(1)
     i_mesh = t % filters(t % find_filter(FILTER_MESH)) % int_bins(1)
     m => meshes(i_mesh)
 
-    ! set mesh widths
+    ! Set mesh widths
     cmfd % hxyz(1,:,:,:) = m % width(1) ! set x width
     cmfd % hxyz(2,:,:,:) = m % width(2) ! set y width
     cmfd % hxyz(3,:,:,:) = m % width(3) ! set z width
 
-   ! begin loop around tallies
+   ! Begin loop around tallies
    TAL: do ital = 1, n_cmfd_tallies
 
-     ! associate tallies and mesh
+     ! Associate tallies and mesh
      t => cmfd_tallies(ital)
      i_mesh = t % filters(t % find_filter(FILTER_MESH)) % int_bins(1)
      m => meshes(i_mesh)
@@ -117,98 +116,99 @@ contains
      i_filter_eout = t % find_filter(FILTER_ENERGYOUT)
      i_filter_surf = t % find_filter(FILTER_SURFACE)
 
-     ! begin loop around space
+     ! Begin loop around space
      ZLOOP: do k = 1,nz
 
        YLOOP: do j = 1,ny
 
           XLOOP: do i = 1,nx
  
-            ! check for active mesh cell
+            ! Check for active mesh cell
             if (allocated(cmfd%coremap)) then
               if (cmfd%coremap(i,j,k) == CMFD_NOACCEL) then
                 cycle
               end if
             end if
 
-            ! loop around energy groups
+            ! Loop around energy groups
             OUTGROUP: do h = 1,ng
 
-              ! start tally 1
+              ! Start tally 1
               TALLY: if (ital == 1) then
 
-                ! reset all bins to 1
+                ! Reset all bins to 1
                 matching_bins(1:t%n_filters) = 1
 
-                ! set ijk as mesh indices
+                ! Set ijk as mesh indices
                 ijk = (/ i, j, k /)
 
-                ! get bin number for mesh indices
+                ! Get bin number for mesh indices
                 matching_bins(i_filter_mesh) = mesh_indices_to_bin(m,ijk)
 
-                ! apply energy in filter
+                ! Apply energy in filter
                 if (i_filter_ein > 0) then
                   matching_bins(i_filter_ein) = ng - h + 1
                 end if
 
-                ! calculate score index from bins
+                ! Calculate score index from bins
                 score_index = sum((matching_bins(1:t%n_filters) - 1) * t%stride) + 1
 
-                ! get flux
+                ! Get flux
                 flux = t % results(1,score_index) % sum
                 cmfd % flux(h,i,j,k) = flux
 
-                ! detect zero flux
-                if ((flux - 0.0D0) < 1.0E-10_8) then
-                  print *,h,i,j,k,flux
-                  message = 'Detected zero flux without coremap overlay'
+                ! Detect zero flux, abort if located
+                if ((flux - ZERO) < TINY_BIT) then
+                  message = 'Detected zero flux without coremap overlay at: (' &
+                          // to_str(i) // ',' // to_str(j) // ',' // to_str(k) &
+                          // ') in group ' // to_str(h)
                   call fatal_error()
                 end if
 
-                ! get total rr and convert to total xs
+                ! Get total rr and convert to total xs
                 cmfd % totalxs(h,i,j,k) = t % results(2,score_index) % sum / flux
 
-                ! get p1 scatter rr and convert to p1 scatter xs
+                ! Get p1 scatter rr and convert to p1 scatter xs
                 cmfd % p1scattxs(h,i,j,k) = t % results(3,score_index) % sum / flux
 
-                ! calculate diffusion coefficient
+                ! Calculate diffusion coefficient
                 cmfd % diffcof(h,i,j,k) = ONE/(3.0_8*(cmfd % totalxs(h,i,j,k) - &
                      cmfd % p1scattxs(h,i,j,k)))
 
               else if (ital == 2) then
 
-                ! begin loop to get energy out tallies
+                ! Begin loop to get energy out tallies
                 INGROUP: do g = 1, ng
 
-                  ! reset all bins to 1
+                  ! Reset all bins to 1
                   matching_bins(1:t%n_filters) = 1
 
-                  ! set ijk as mesh indices
+                  ! Set ijk as mesh indices
                   ijk = (/ i, j, k /)
 
-                  ! get bin number for mesh indices
+                  ! Get bin number for mesh indices
                   matching_bins(i_filter_mesh) = mesh_indices_to_bin(m,ijk)
 
                   if (i_filter_ein > 0) then
-                    ! apply energy in filter
+                    ! Apply energy in filter
                     matching_bins(i_filter_ein) = ng - h + 1
 
-                    ! set energy out bin
+                    ! Set energy out bin
                     matching_bins(i_filter_eout) = ng - g + 1
                   end if
 
-                  ! calculate score index from bins
+                  ! Calculate score index from bins
                   score_index = sum((matching_bins(1:t%n_filters) - 1) * t%stride) + 1
 
-                  ! get scattering
+                  ! Get scattering
                   cmfd % scattxs(h,g,i,j,k) = t % results(1,score_index) % sum /&
                        cmfd % flux(h,i,j,k)
 
-                  ! get nu-fission
+                  ! Get nu-fission
                   cmfd % nfissxs(h,g,i,j,k) = t % results(2,score_index) % sum /&
                        cmfd % flux(h,i,j,k)
 
-                  ! bank source
+                  ! Bank source
                   cmfd % openmc_src(g,i,j,k) = cmfd % openmc_src(g,i,j,k) + &
                        t % results(2,score_index) % sum
 
@@ -216,13 +216,13 @@ contains
 
               else if (ital == 3) then
 
-                ! initialize and filter for energy
+                ! Initialize and filter for energy
                 matching_bins(1:t%n_filters) = 1
                 if (i_filter_ein > 0) then
                   matching_bins(i_filter_ein) = ng - h + 1
                 end if
 
-                ! left surface
+                ! Left surface
                 matching_bins(i_filter_mesh) = mesh_indices_to_bin(m, &
                      (/ i-1, j, k /) + 1, .true.)
                 matching_bins(i_filter_surf) = IN_RIGHT
@@ -232,7 +232,7 @@ contains
                 score_index = sum((matching_bins(1:t%n_filters) - 1) * t % stride) + 1 ! incoming
                 cmfd % current(2,h,i,j,k) = t % results(1,score_index) % sum
 
-                ! right surface
+                ! Right surface
                 matching_bins(i_filter_mesh) = mesh_indices_to_bin(m, &
                      (/ i, j, k /) + 1, .true.)
                 matching_bins(i_filter_surf) = IN_RIGHT
@@ -242,7 +242,7 @@ contains
                 score_index = sum((matching_bins(1:t%n_filters) - 1) * t % stride) + 1 ! outgoing
                 cmfd % current(4,h,i,j,k) = t % results(1,score_index) % sum
 
-                ! back surface
+                ! Back surface
                 matching_bins(i_filter_mesh) = mesh_indices_to_bin(m, &
                      (/ i, j-1, k /) + 1, .true.)
                 matching_bins(i_filter_surf) = IN_FRONT
@@ -252,7 +252,7 @@ contains
                 score_index = sum((matching_bins(1:t%n_filters) - 1) * t % stride) + 1 ! incoming
                 cmfd % current(6,h,i,j,k) = t % results(1,score_index) % sum
 
-                ! front surface
+                ! Front surface
                 matching_bins(i_filter_mesh) = mesh_indices_to_bin(m, &
                      (/ i, j, k /) + 1, .true.)
                 matching_bins(i_filter_surf) = IN_FRONT
@@ -262,7 +262,7 @@ contains
                 score_index = sum((matching_bins(1:t%n_filters) - 1) * t % stride) + 1 ! outgoing
                 cmfd % current(8,h,i,j,k) = t % results(1,score_index) % sum
 
-                ! bottom surface
+                ! Bottom surface
                 matching_bins(i_filter_mesh) = mesh_indices_to_bin(m, &
                      (/ i, j, k-1 /) + 1, .true.)
                 matching_bins(i_filter_surf) = IN_TOP
@@ -272,7 +272,7 @@ contains
                 score_index = sum((matching_bins(1:t%n_filters) - 1) * t % stride) + 1 ! incoming
                 cmfd % current(10,h,i,j,k) = t % results(1,score_index) % sum
 
-                ! top surface
+                ! Top surface
                 matching_bins(i_filter_mesh) = mesh_indices_to_bin(m, &
                      (/ i, j, k /) + 1, .true.)
                 matching_bins(i_filter_surf) = IN_TOP
@@ -294,10 +294,10 @@ contains
 
     end do TAL
 
-    ! normalize openmc source distribution
+    ! Normalize openmc source distribution
     cmfd % openmc_src = cmfd % openmc_src/sum(cmfd % openmc_src)*cmfd%norm
 
-    ! nullify all pointers
+    ! Nullify all pointers
     if (associated(t)) nullify(t)
     if (associated(m)) nullify(m)
 
@@ -312,47 +312,47 @@ contains
     use constants,  only: CMFD_NOACCEL
     use global,     only: cmfd
 
-    integer :: kount=1           ! counter for unique fuel assemblies
-    integer :: nx                ! number of mesh cells in x direction
-    integer :: ny                ! number of mesh cells in y direction
-    integer :: nz                ! number of mesh cells in z direction
-    integer :: i                 ! iteration counter for x
-    integer :: j                 ! iteration counter for y
-    integer :: k                 ! iteration counter for z
+    integer :: counter=1 ! counter for unique fuel assemblies
+    integer :: nx        ! number of mesh cells in x direction
+    integer :: ny        ! number of mesh cells in y direction
+    integer :: nz        ! number of mesh cells in z direction
+    integer :: i         ! iteration counter for x
+    integer :: j         ! iteration counter for y
+    integer :: k         ! iteration counter for z
 
-    ! extract spatial indices from object
+    ! Extract spatial indices from object
     nx = cmfd % indices(1)
     ny = cmfd % indices(2)
     nz = cmfd % indices(3)
 
-    ! count how many fuel assemblies exist
+    ! Count how many fuel assemblies exist
     cmfd % mat_dim = sum(cmfd % coremap - 1)
 
-    ! allocate indexmap
+    ! Allocate indexmap
     if (.not. allocated(cmfd % indexmap)) &
          allocate(cmfd % indexmap(cmfd % mat_dim,3))
 
-    ! begin loops over spatial indices
+    ! Begin loops over spatial indices
     ZLOOP: do k = 1, nz
 
       YLOOP: do j = 1, ny
 
         XLOOP: do i = 1, nx
 
-          ! check for reflector
+          ! Check for reflector
           if (cmfd % coremap(i,j,k) == 1) then
 
-            ! reset value to 99999
+            ! reset value to CMFD no acceleration constant 
             cmfd % coremap(i,j,k) = CMFD_NOACCEL 
 
           else
 
-            ! must be a fuel --> give unique id number
-            cmfd % coremap(i,j,k) = kount
-            cmfd % indexmap(kount,1) = i
-            cmfd % indexmap(kount,2) = j
-            cmfd % indexmap(kount,3) = k
-            kount = kount + 1
+            ! Must be a fuel --> give unique id number
+            cmfd % coremap(i,j,k) = counter
+            cmfd % indexmap(counter,1) = i
+            cmfd % indexmap(counter,2) = j
+            cmfd % indexmap(counter,3) = k
+            counter = counter + 1
 
           end if
 
@@ -368,10 +368,10 @@ contains
 ! NEUTRON_BALANCE writes a file that contains n. bal. info for all cmfd mesh
 !===============================================================================
 
-  subroutine neutron_balance(uid)
+  subroutine neutron_balance()
 
     use constants,    only: ONE, ZERO, CMFD_NOACCEL, CMFD_NORES
-    use global,       only: cmfd, keff
+    use global,       only: cmfd, keff, current_batch
 
     integer :: nx           ! number of mesh cells in x direction
     integer :: ny           ! number of mesh cells in y direction
@@ -383,26 +383,28 @@ contains
     integer :: g            ! iteration counter for g
     integer :: h            ! iteration counter for outgoing groups
     integer :: l            ! iteration counter for leakage
-    integer :: uid
+    integer :: cnt          ! number of locations to count neutron balance
     real(8) :: leakage      ! leakage term in neutron balance
     real(8) :: interactions ! total number of interactions in balance
     real(8) :: scattering   ! scattering term in neutron balance
     real(8) :: fission      ! fission term in neutron balance
     real(8) :: res          ! residual of neutron balance
+    real(8) :: rms          ! RMS of the residual
 
-    ! check if keff is close to 0 (happens on first active batch)
-    if (keff < 1e-8_8) return
-
-    ! extract spatial and energy indices from object
+    ! Extract spatial and energy indices from object
     nx = cmfd % indices(1)
     ny = cmfd % indices(2)
     nz = cmfd % indices(3)
     ng = cmfd % indices(4)
 
-    ! allocate res dataspace
+    ! Allocate res dataspace
     if (.not. allocated(cmfd%resnb)) allocate(cmfd%resnb(ng,nx,ny,nz))
 
-    ! begin loop around space and energy groups
+    ! Reset rms and cnt
+    rms = ZERO
+    cnt = 0
+
+    ! Begin loop around space and energy groups
     ZLOOP: do k = 1, nz
 
       YLOOP: do j = 1, ny
@@ -411,7 +413,7 @@ contains
 
           GROUPG: do g = 1, ng
 
-            ! check for active mesh
+            ! Check for active mesh
             if (allocated(cmfd%coremap)) then
               if (cmfd%coremap(i,j,k) == CMFD_NOACCEL) then
                 cmfd%resnb(g,i,j,k) = CMFD_NORES
@@ -419,7 +421,7 @@ contains
               end if
             end if
 
-            ! get leakage
+            ! Get leakage
             leakage = ZERO
             LEAK: do l = 1, 3
 
@@ -430,10 +432,10 @@ contains
 
             end do LEAK
 
-            ! interactions
+            ! Interactions
             interactions = cmfd % totalxs(g,i,j,k) * cmfd % flux(g,i,j,k)
 
-            ! get scattering and fission
+            ! Get scattering and fission
             scattering = ZERO
             fission = ZERO
             GROUPH: do h = 1, ng
@@ -446,25 +448,18 @@ contains
 
             end do GROUPH
 
-            ! compute residual
+            ! Compute residual
             res = leakage + interactions - scattering - (ONE/keff)*fission
 
-            ! normalize by flux
+            ! Normalize by flux
             res = res/cmfd%flux(g,i,j,k)
 
-            ! bank res in cmfd object
+            ! Bank res in cmfd object
             cmfd%resnb(g,i,j,k) = res
 
-            ! write out info to file
-            write(uid,'(A,1X,I0,1X,I0,1X,I0,1X,A,1X,I0,1X,A,1PE11.4)') &
-                          'Location',i,j,k,' Group:',g,'Balance:',res
-            write(uid,100) 'Leakage:',leakage
-            write(uid,100) 'Interactions:',interactions
-            write(uid,100) 'Scattering:',scattering
-            write(uid,100) 'Fission:',fission
-            write(uid,100) 'k-eff:',keff
-            write(uid,100) 'k-eff balanced:',fission/(leakage+interactions-scattering)
-            write(uid,100) 'Balance:',keff - fission/(leakage+interactions-scattering)
+            ! Take square for RMS calculation
+            rms = rms + res**2
+            cnt = cnt + 1
 
           end do GROUPG
 
@@ -474,12 +469,13 @@ contains
 
     end do ZLOOP
 
- 100 FORMAT(A,1X,1PE11.4)
+    ! Calculate RMS and record in vector for this batch
+    cmfd % balance(current_batch) = sqrt(ONE/dble(cnt)*rms)
 
   end subroutine neutron_balance
 
 !===============================================================================
-! COMPUTE_DTILDE computes the diffusion coupling coefficient
+! COMPUTE_DTILDE precomputes the diffusion coupling coefficient
 !===============================================================================
 
   subroutine compute_dtilde()
@@ -487,44 +483,44 @@ contains
     use constants,  only: CMFD_NOACCEL, ZERO_FLUX, TINY_BIT
     use global,     only: cmfd, cmfd_coremap
 
-    integer :: nx                 ! maximum number of cells in x direction
-    integer :: ny                 ! maximum number of cells in y direction
-    integer :: nz                 ! maximum number of cells in z direction
-    integer :: ng                 ! maximum number of energy groups
-    integer :: nxyz(3,2)          ! single vector containing boundary locations
-    integer :: i                  ! iteration counter for x
-    integer :: j                  ! iteration counter for y
-    integer :: k                  ! iteration counter for z
-    integer :: g                  ! iteration counter for groups
-    integer :: l                  ! iteration counter for leakages
-    integer :: xyz_idx            ! index for determining if x,y or z leakage
-    integer :: dir_idx            ! index for determining - or + face of cell
-    integer :: shift_idx          ! parameter to shift index by +1 or -1
-    integer :: neig_idx(3)        ! spatial indices of neighbour
-    integer :: bound(6)           ! vector containing indices for boudary check
-    real(8) :: albedo(6)          ! albedo vector with global boundaries
-    real(8) :: cell_dc            ! diffusion coef of current cell
-    real(8) :: cell_hxyz(3)       ! cell dimensions of current ijk cell
-    real(8) :: neig_dc            ! diffusion coefficient of neighbor cell
-    real(8) :: neig_hxyz(3)       ! cell dimensions of neighbor cell
-    real(8) :: dtilde             ! finite difference coupling parameter 
-    real(8) :: ref_albedo         ! albedo to reflector
+    integer :: nx           ! maximum number of cells in x direction
+    integer :: ny           ! maximum number of cells in y direction
+    integer :: nz           ! maximum number of cells in z direction
+    integer :: ng           ! maximum number of energy groups
+    integer :: nxyz(3,2)    ! single vector containing boundary locations
+    integer :: i            ! iteration counter for x
+    integer :: j            ! iteration counter for y
+    integer :: k            ! iteration counter for z
+    integer :: g            ! iteration counter for groups
+    integer :: l            ! iteration counter for leakages
+    integer :: xyz_idx      ! index for determining if x,y or z leakage
+    integer :: dir_idx      ! index for determining - or + face of cell
+    integer :: shift_idx    ! parameter to shift index by +1 or -1
+    integer :: neig_idx(3)  ! spatial indices of neighbour
+    integer :: bound(6)     ! vector containing indices for boudary check
+    real(8) :: albedo(6)    ! albedo vector with global boundaries
+    real(8) :: cell_dc      ! diffusion coef of current cell
+    real(8) :: cell_hxyz(3) ! cell dimensions of current ijk cell
+    real(8) :: neig_dc      ! diffusion coefficient of neighbor cell
+    real(8) :: neig_hxyz(3) ! cell dimensions of neighbor cell
+    real(8) :: dtilde       ! finite difference coupling parameter 
+    real(8) :: ref_albedo   ! albedo to reflector
 
-    ! get maximum of spatial and group indices
+    ! Get maximum of spatial and group indices
     nx = cmfd%indices(1)
     ny = cmfd%indices(2)
     nz = cmfd%indices(3)
     ng = cmfd%indices(4)
 
-    ! create single vector of these indices for boundary calculation
+    ! Create single vector of these indices for boundary calculation
     nxyz(1,:) = (/1,nx/)
     nxyz(2,:) = (/1,ny/)
     nxyz(3,:) = (/1,nz/)
 
-    ! get boundary condition information
+    ! Get boundary condition information
     albedo = cmfd%albedo
 
-    ! geting loop over group and spatial indices
+    ! Loop over group and spatial indices
     ZLOOP: do k = 1, nz
 
       YLOOP: do j = 1, ny
@@ -533,63 +529,63 @@ contains
 
           GROUP: do g = 1, ng
 
-            ! check for active mesh cell
+            ! Check for active mesh cell
             if (allocated(cmfd%coremap)) then
               if (cmfd%coremap(i,j,k) == CMFD_NOACCEL) cycle
             end if
 
-            ! get cell data
+            ! Get cell data
             cell_dc = cmfd%diffcof(g,i,j,k)
             cell_hxyz = cmfd%hxyz(:,i,j,k)
 
-            ! setup of vector to identify boundary conditions
+            ! Setup of vector to identify boundary conditions
             bound = (/i,i,j,j,k,k/)
 
-            ! begin loop around sides of cell for leakage
+            ! Begin loop around sides of cell for leakage
             LEAK: do l = 1, 6
 
-              ! define xyz and +/- indices
+              ! Define xyz and +/- indices
               xyz_idx = int(ceiling(real(l)/real(2)))  ! x=1, y=2, z=3
               dir_idx = 2 - mod(l,2) ! -=1, +=2
-              shift_idx = -2*mod(l,2) + 1          ! shift neig by -1 or +1
+              shift_idx = -2*mod(l,2) + 1 ! shift neig by -1 or +1
 
-              ! check if at a boundary
+              ! Check if at a boundary
               if (bound(l) == nxyz(xyz_idx,dir_idx)) then
 
-                ! compute dtilde
+                ! Compute dtilde with albedo boundary condition
                 dtilde = (2*cell_dc*(1-albedo(l)))/(4*cell_dc*(1+albedo(l)) + &
                      (1-albedo(l))*cell_hxyz(xyz_idx))
 
-                ! check for zero flux
+                ! Check for zero flux
                 if (abs(albedo(l) - ZERO_FLUX) < TINY_BIT) dtilde = 2*cell_dc / &
                      cell_hxyz(xyz_idx)
 
               else  ! not a boundary
 
-                ! compute neighboring cell indices
+                ! Compute neighboring cell indices
                 neig_idx = (/i,j,k/)                ! begin with i,j,k
                 neig_idx(xyz_idx) = shift_idx + neig_idx(xyz_idx)
 
-                ! get neigbor cell data
+                ! Get neigbor cell data
                 neig_dc = cmfd%diffcof(g,neig_idx(1),neig_idx(2),neig_idx(3))
                 neig_hxyz = cmfd%hxyz(:,neig_idx(1),neig_idx(2),neig_idx(3))
 
-                ! check for fuel-reflector interface
+                ! Check for fuel-reflector interface
                 if (cmfd_coremap) then
 
                   if (cmfd % coremap(neig_idx(1),neig_idx(2),neig_idx(3)) == &
                        CMFD_NOACCEL .and. cmfd % coremap(i,j,k) /= CMFD_NOACCEL) then
 
-                    ! get albedo
+                    ! Get albedo
                     ref_albedo = get_reflector_albedo(l,g,i,j,k)
 
-                    ! compute dtilde
+                    ! Compute dtilde
                     dtilde = (2*cell_dc*(1-ref_albedo))/(4*cell_dc*(1+ &
                          ref_albedo)+(1-ref_albedo)*cell_hxyz(xyz_idx))
 
-                  else ! not next to a reflector or no core map
+                  else ! Not next to a reflector or no core map
 
-                    ! compute dtilde
+                    ! Compute dtilde to neighbor cell
                     dtilde = (2*cell_dc*neig_dc)/(neig_hxyz(xyz_idx)*cell_dc + &
                          cell_hxyz(xyz_idx)*neig_dc)
 
@@ -597,7 +593,7 @@ contains
 
                 else ! no core map
 
-                  ! compute dtilde
+                  ! Compute dtilde to neighbor cell
                   dtilde = (2*cell_dc*neig_dc)/(neig_hxyz(xyz_idx)*cell_dc + &
                        cell_hxyz(xyz_idx)*neig_dc)
 
@@ -605,7 +601,7 @@ contains
 
               end if
 
-              ! record dtilde in cmfd object
+              ! Record dtilde in cmfd object
               cmfd%dtilde(l,g,i,j,k) = dtilde
 
             end do LEAK
@@ -629,40 +625,40 @@ contains
     use constants,  only: CMFD_NOACCEL, ZERO
     use global,     only: cmfd, cmfd_coremap
 
-    integer :: nx                 ! maximum number of cells in x direction
-    integer :: ny                 ! maximum number of cells in y direction
-    integer :: nz                 ! maximum number of cells in z direction
-    integer :: ng                 ! maximum number of energy groups
-    integer :: nxyz(3,2)          ! single vector containing boundary locations
-    integer :: i                  ! iteration counter for x
-    integer :: j                  ! iteration counter for y
-    integer :: k                  ! iteration counter for z
-    integer :: g                  ! iteration counter for groups
-    integer :: l                  ! iteration counter for leakages
-    integer :: xyz_idx            ! index for determining if x,y or z leakage
-    integer :: dir_idx            ! index for determining - or + face of cell
-    integer :: shift_idx          ! parameter to shift index by +1 or -1
-    integer :: neig_idx(3)        ! spatial indices of neighbour
-    integer :: bound(6)           ! vector containing indices for boudary check
-    real(8) :: cell_dtilde(6)     ! cell dtilde for each face
-    real(8) :: cell_flux          ! flux in current cell
-    real(8) :: current(12)        ! area integrated cell current at each face
-    real(8) :: net_current        ! net current on a face
-    real(8) :: neig_flux          ! flux in neighbor cell
-    real(8) :: dhat               ! dhat equivalence parameter
+    integer :: nx             ! maximum number of cells in x direction
+    integer :: ny             ! maximum number of cells in y direction
+    integer :: nz             ! maximum number of cells in z direction
+    integer :: ng             ! maximum number of energy groups
+    integer :: nxyz(3,2)      ! single vector containing boundary locations
+    integer :: i              ! iteration counter for x
+    integer :: j              ! iteration counter for y
+    integer :: k              ! iteration counter for z
+    integer :: g              ! iteration counter for groups
+    integer :: l              ! iteration counter for leakages
+    integer :: xyz_idx        ! index for determining if x,y or z leakage
+    integer :: dir_idx        ! index for determining - or + face of cell
+    integer :: shift_idx      ! parameter to shift index by +1 or -1
+    integer :: neig_idx(3)    ! spatial indices of neighbour
+    integer :: bound(6)       ! vector containing indices for boudary check
+    real(8) :: cell_dtilde(6) ! cell dtilde for each face
+    real(8) :: cell_flux      ! flux in current cell
+    real(8) :: current(12)    ! area integrated cell current at each face
+    real(8) :: net_current    ! net current on a face
+    real(8) :: neig_flux      ! flux in neighbor cell
+    real(8) :: dhat           ! dhat equivalence parameter
 
-    ! get maximum of spatial and group indices
+    ! Get maximum of spatial and group indices
     nx = cmfd%indices(1)
     ny = cmfd%indices(2)
     nz = cmfd%indices(3)
     ng = cmfd%indices(4)
 
-    ! create single vector of these indices for boundary calculation
+    ! Create single vector of these indices for boundary calculation
     nxyz(1,:) = (/1,nx/)
     nxyz(2,:) = (/1,ny/)
     nxyz(3,:) = (/1,nz/)
 
-    ! geting loop over group and spatial indices
+    ! Geting loop over group and spatial indices
     ZLOOP:  do k = 1,nz
 
       YLOOP: do j = 1,ny
@@ -671,51 +667,51 @@ contains
 
           GROUP: do g = 1,ng
 
-            ! check for active mesh cell
+            ! Check for active mesh cell
             if (allocated(cmfd%coremap)) then
               if (cmfd%coremap(i,j,k) == CMFD_NOACCEL) then
                 cycle
               end if
             end if
 
-            ! get cell data
+            ! Get cell data
             cell_dtilde = cmfd%dtilde(:,g,i,j,k)
             cell_flux = cmfd%flux(g,i,j,k)/product(cmfd%hxyz(:,i,j,k))
             current = cmfd%current(:,g,i,j,k)
 
-            ! setup of vector to identify boundary conditions
+            ! Setup of vector to identify boundary conditions
             bound = (/i,i,j,j,k,k/)
 
-            ! begin loop around sides of cell for leakage
+            ! Begin loop around sides of cell for leakage
             LEAK: do l = 1,6
 
-              ! define xyz and +/- indices
+              ! Define xyz and +/- indices
               xyz_idx = int(ceiling(real(l)/real(2)))  ! x=1, y=2, z=3
               dir_idx = 2 - mod(l,2) ! -=1, +=2
               shift_idx = -2*mod(l,2) +1          ! shift neig by -1 or +1
 
-              ! calculate net current on l face (divided by surf area)
+              ! Calculate net current on l face (divided by surf area)
               net_current = (current(2*l) - current(2*l-1)) / &
                    product(cmfd%hxyz(:,i,j,k)) * cmfd%hxyz(xyz_idx,i,j,k)
 
-              ! check if at a boundary
+              ! Check if at a boundary
               if (bound(l) == nxyz(xyz_idx,dir_idx)) then
 
-                ! compute dhat
+                ! Compute dhat
                 dhat = (net_current - shift_idx*cell_dtilde(l)*cell_flux) / &
                      cell_flux
 
               else  ! not a boundary
 
-                ! compute neighboring cell indices
+                ! Compute neighboring cell indices
                 neig_idx = (/i,j,k/)                ! begin with i,j,k
                 neig_idx(xyz_idx) = shift_idx + neig_idx(xyz_idx)
 
-                ! get neigbor flux 
+                ! Get neigbor flux 
                 neig_flux = cmfd%flux(g,neig_idx(1),neig_idx(2),neig_idx(3)) / &
                      product(cmfd%hxyz(:,neig_idx(1),neig_idx(2),neig_idx(3)))
 
-                ! check for fuel-reflector interface
+                ! Check for fuel-reflector interface
                 if (cmfd_coremap) then
 
                   if (cmfd % coremap(neig_idx(1),neig_idx(2),neig_idx(3)) == &
@@ -727,7 +723,7 @@ contains
 
                   else ! not a fuel-reflector interface
 
-                    ! compute dhat 
+                    ! Compute dhat 
                     dhat = (net_current + shift_idx*cell_dtilde(l)* &
                          (neig_flux - cell_flux))/(neig_flux + cell_flux)
 
@@ -735,7 +731,7 @@ contains
 
                 else ! not for fuel-reflector case
 
-                  ! compute dhat 
+                  ! Compute dhat 
                   dhat = (net_current + shift_idx*cell_dtilde(l)* &
                        (neig_flux - cell_flux))/(neig_flux + cell_flux)
 
@@ -743,13 +739,6 @@ contains
 
               end if
 
-              ! check for zero net current
-!             if ((abs(current(2*l)-current(2*l-1)) < 1e-8_8).and.xyz_idx/=3) then
-!               print *,'Zero net current interface',g,i,j,k
-!               print *,current(2*l),current(2*l-1),net_current
-!               dhat = ZERO
-!             end if
-
               ! record dhat in cmfd object
               cmfd%dhat(l,g,i,j,k) = dhat
 
@@ -778,23 +767,23 @@ contains
     use global,     only: cmfd, cmfd_hold_weights
 
     real(8) :: get_reflector_albedo ! reflector albedo
-    integer :: i                    ! iteration counter for x
-    integer :: j                    ! iteration counter for y
-    integer :: k                    ! iteration counter for z
-    integer :: g                    ! iteration counter for groups
-    integer :: l                    ! iteration counter for leakages
+    integer, intent(in) :: i ! iteration counter for x
+    integer, intent(in) :: j ! iteration counter for y
+    integer, intent(in) :: k ! iteration counter for z
+    integer, intent(in) :: g ! iteration counter for groups
+    integer, intent(in) :: l ! iteration counter for leakages
 
-    integer :: shift_idx            ! parameter to shift index by +1 or -1
-    real(8) :: current(12)          ! partial currents for all faces of mesh cell            
-    real(8) :: albedo               ! the albedo
+    integer :: shift_idx   ! parameter to shift index by +1 or -1
+    real(8) :: current(12) ! partial currents for all faces of mesh cell            
+    real(8) :: albedo      ! the albedo
 
-    ! get partial currents from object
+    ! Get partial currents from object
     current = cmfd%current(:,g,i,j,k)
 
-    ! define xyz and +/- indices
+    ! Define xyz and +/- indices
     shift_idx = -2*mod(l,2) + 1          ! shift neig by -1 or +1
 
-    ! calculate albedo
+    ! Calculate albedo
     if ((shift_idx ==  1 .and. current(2*l  ) < 1.0e-10_8) .or. &
         (shift_idx == -1 .and. current(2*l-1) < 1.0e-10_8)) then
       albedo = ALBEDO_REJECT 
@@ -803,43 +792,13 @@ contains
       albedo = (current(2*l-1)/current(2*l))**(shift_idx)
     end if
 
-    ! assign to function variable
+    ! Assign to function variable
     get_reflector_albedo = albedo
 
   end function get_reflector_albedo
 
 !===============================================================================
-! FIX_2_GRP modifies xs for benchmark with stand alone (homogeneous)
-!===============================================================================
-
-  subroutine fix_2_grp()
-
-    use global,  only: cmfd
-
-    ! overwrite cross sections
-    cmfd % totalxs(1,:,:,:) = 0.02597_8
-    cmfd % totalxs(2,:,:,:) = 0.06669_8
-    cmfd % scattxs(1,1,:,:,:) = 0.0_8
-    cmfd % scattxs(1,2,:,:,:) = 0.01742_8
-    cmfd % scattxs(2,1,:,:,:) = 0.0_8
-    cmfd % scattxs(2,2,:,:,:) = 0.0_8
-    cmfd % nfissxs(1,1,:,:,:) = 0.00536_8
-    cmfd % nfissxs(1,2,:,:,:) = 0.0_8
-    cmfd % nfissxs(2,1,:,:,:) = 0.10433_8
-    cmfd % nfissxs(2,2,:,:,:) = 0.0_8
-    cmfd % diffcof(1,:,:,:) = 1.4176_8
-    cmfd % diffcof(2,:,:,:) = 0.37336_8
-    cmfd % hxyz(1,:,:,:) = 0.5_8
-    cmfd % hxyz(2,:,:,:) = 0.5_8
-    cmfd % hxyz(3,:,:,:) = 0.5_8
-
-    ! set dhat reset to true
-    dhat_reset = .true.
-
-  end subroutine fix_2_grp
-
-!===============================================================================
-! FIX_NEUTRON_BALANCE
+! FIX_NEUTRON_BALANCE is a method to adjust parameters to have perfect balance
 !===============================================================================
 
   subroutine fix_neutron_balance()
@@ -848,54 +807,54 @@ contains
     use global,     only: cmfd, keff
     use, intrinsic :: ISO_FORTRAN_ENV
 
-    integer :: nx                ! number of mesh cells in x direction
-    integer :: ny                ! number of mesh cells in y direction
-    integer :: nz                ! number of mesh cells in z direction
-    integer :: ng                ! number of energy groups
-    integer :: i                 ! iteration counter for x
-    integer :: j                 ! iteration counter for y
-    integer :: k                 ! iteration counter for z
-    integer :: l                 ! iteration counter for surface
-    real(8) :: leak1             ! leakage rate in group 1
-    real(8) :: leak2             ! leakage rate in group 2
-    real(8) :: flux1             ! group 1 volume int flux
-    real(8) :: flux2             ! group 2 volume int flux
-    real(8) :: sigt1             ! group 1 total xs
-    real(8) :: sigt2             ! group 2 total xs
-    real(8) :: sigs11            ! scattering transfer 1 --> 1
-    real(8) :: sigs21            ! scattering transfer 2 --> 1
-    real(8) :: sigs12            ! scattering transfer 1 --> 2
-    real(8) :: sigs22            ! scattering transfer 2 --> 2
-    real(8) :: nsigf11           ! fission transfer 1 --> 1
-    real(8) :: nsigf21           ! fission transfer 2 --> 1
-    real(8) :: nsigf12           ! fission transfer 1 --> 2
-    real(8) :: nsigf22           ! fission transfer 2 --> 2
-    real(8) :: siga1             ! group 1 abs xs
-    real(8) :: siga2             ! group 2 abs xs
-    real(8) :: sigs12_eff        ! effective downscatter xs
+    integer :: nx         ! number of mesh cells in x direction
+    integer :: ny         ! number of mesh cells in y direction
+    integer :: nz         ! number of mesh cells in z direction
+    integer :: ng         ! number of energy groups
+    integer :: i          ! iteration counter for x
+    integer :: j          ! iteration counter for y
+    integer :: k          ! iteration counter for z
+    integer :: l          ! iteration counter for surface
+    real(8) :: leak1      ! leakage rate in group 1
+    real(8) :: leak2      ! leakage rate in group 2
+    real(8) :: flux1      ! group 1 volume int flux
+    real(8) :: flux2      ! group 2 volume int flux
+    real(8) :: sigt1      ! group 1 total xs
+    real(8) :: sigt2      ! group 2 total xs
+    real(8) :: sigs11     ! scattering transfer 1 --> 1
+    real(8) :: sigs21     ! scattering transfer 2 --> 1
+    real(8) :: sigs12     ! scattering transfer 1 --> 2
+    real(8) :: sigs22     ! scattering transfer 2 --> 2
+    real(8) :: nsigf11    ! fission transfer 1 --> 1
+    real(8) :: nsigf21    ! fission transfer 2 --> 1
+    real(8) :: nsigf12    ! fission transfer 1 --> 2
+    real(8) :: nsigf22    ! fission transfer 2 --> 2
+    real(8) :: siga1      ! group 1 abs xs
+    real(8) :: siga2      ! group 2 abs xs
+    real(8) :: sigs12_eff ! effective downscatter xs
 
-    ! extract spatial and energy indices from object
+    ! Extract spatial and energy indices from object
     nx = cmfd % indices(1)
     ny = cmfd % indices(2)
     nz = cmfd % indices(3)
     ng = cmfd % indices(4)
 
-    ! return if not two groups
+    ! Return if not two groups
     if (ng /= 2) return
 
-    ! begin loop around space and energy groups
+    ! Begin loop around space and energy groups
     ZLOOP: do k = 1, nz
 
       YLOOP: do j = 1, ny
 
         XLOOP: do i = 1, nx
 
-          ! check for active mesh
+          ! Check for active mesh
           if (allocated(cmfd%coremap)) then
             if (cmfd%coremap(i,j,k) == CMFD_NOACCEL) cycle 
           end if
 
-          ! compute leakage in groups 1 and 2
+          ! Compute leakage in groups 1 and 2
           leak1 = ZERO 
           leak2 = ZERO
           LEAK: do l = 1, 3
@@ -913,7 +872,7 @@ contains
 
           end do LEAK
 
-          ! extract cross sections and flux from object
+          ! Extract cross sections and flux from object
           flux1 = cmfd % flux(1,i,j,k)
           flux2 = cmfd % flux(2,i,j,k)
           sigt1 = cmfd % totalxs(1,i,j,k)
@@ -927,37 +886,37 @@ contains
           nsigf12 = cmfd % nfissxs(1,2,i,j,k)
           nsigf22 = cmfd % nfissxs(2,2,i,j,k)
 
-          ! check for no fission into group 2
+          ! Check for no fission into group 2
           if (.not.(nsigf12 < 1e-6_8 .and. nsigf22 < 1e-6_8)) then
             write(OUTPUT_UNIT,'(A,1PE11.4,1X,1PE11.4)') 'Fission in G=2', &
                   nsigf12,nsigf22
           end if
 
-          ! compute absorption xs
+          ! Compute absorption xs
           siga1 = sigt1 - sigs11 - sigs12
           siga2 = sigt2 - sigs22 - sigs21
 
-          ! compute effective downscatter xs
+          ! Compute effective downscatter xs
           sigs12_eff = (ONE/keff*nsigf11*flux1 - leak1 - siga1*flux1 &
                - ONE/keff*nsigf21/siga2*leak2 ) / ( flux1*(ONE &
                - ONE/keff*nsigf21/siga2))
 
-          ! redefine flux 2
+          ! Redefine flux 2
           flux2 = (sigs12_eff*flux1 - leak2)/siga2
           cmfd % flux(2,i,j,k) = flux2
  
-          ! recompute total cross sections (use effective and no upscattering)
+          ! Recompute total cross sections (use effective and no upscattering)
           sigt1 = siga1 + sigs11 + sigs12_eff
           sigt2 = siga2 + sigs22
 
-          ! record total xs
+          ! Record total xs
           cmfd % totalxs(1,i,j,k) = sigt1
           cmfd % totalxs(2,i,j,k) = sigt2
 
-          ! record effective downscatter xs
+          ! Record effective downscatter xs
           cmfd % scattxs(1,2,i,j,k) = sigs12_eff
 
-          ! zero out upscatter cross section
+          ! Zero out upscatter cross section
           cmfd % scattxs(2,1,i,j,k) = ZERO 
 
         end do XLOOP
@@ -969,7 +928,7 @@ contains
   end subroutine fix_neutron_balance
 
 !===============================================================================
-! FIX_NEUTRON_BALANCE
+! COMPUTE_EFFECTIVE_DOWNSCATTER changes downscatter rate for zero upscatter
 !===============================================================================
 
   subroutine compute_effective_downscatter()
@@ -996,31 +955,28 @@ contains
     real(8) :: siga2             ! group 2 abs xs
     real(8) :: sigs12_eff        ! effective downscatter xs
 
-    ! check for balance
-    if (.not. cmfd_downscatter) return
-
-    ! extract spatial and energy indices from object
+    ! Extract spatial and energy indices from object
     nx = cmfd % indices(1)
     ny = cmfd % indices(2)
     nz = cmfd % indices(3)
     ng = cmfd % indices(4)
 
-    ! return if not two groups
+    ! Return if not two groups
     if (ng /= 2) return
 
-    ! begin loop around space and energy groups
+    ! Begin loop around space and energy groups
     ZLOOP: do k = 1, nz
 
       YLOOP: do j = 1, ny
 
         XLOOP: do i = 1, nx
 
-          ! check for active mesh
+          ! Check for active mesh
           if (allocated(cmfd%coremap)) then
             if (cmfd%coremap(i,j,k) == CMFD_NOACCEL) cycle
           end if
 
-          ! extract cross sections and flux from object
+          ! Extract cross sections and flux from object
           flux1 = cmfd % flux(1,i,j,k)
           flux2 = cmfd % flux(2,i,j,k)
           sigt1 = cmfd % totalxs(1,i,j,k)
@@ -1030,25 +986,25 @@ contains
           sigs12 = cmfd % scattxs(1,2,i,j,k)
           sigs22 = cmfd % scattxs(2,2,i,j,k)
 
-          ! compute absorption xs
+          ! Compute absorption xs
           siga1 = sigt1 - sigs11 - sigs12
           siga2 = sigt2 - sigs22 - sigs21
 
-          ! compute effective downscatter xs
+          ! Compute effective downscatter xs
           sigs12_eff = sigs12 - sigs21*flux2/flux1 
 
-          ! recompute total cross sections (use effective and no upscattering)
+          ! Recompute total cross sections (use effective and no upscattering)
           sigt1 = siga1 + sigs11 + sigs12_eff
           sigt2 = siga2 + sigs22
     
-          ! record total xs
+          ! Record total xs
           cmfd % totalxs(1,i,j,k) = sigt1
           cmfd % totalxs(2,i,j,k) = sigt2
 
-          ! record effective downscatter xs
+          ! Record effective downscatter xs
           cmfd % scattxs(1,2,i,j,k) = sigs12_eff
 
-          ! zero out upscatter cross section
+          ! Zero out upscatter cross section
           cmfd % scattxs(2,1,i,j,k) = ZERO 
 
         end do XLOOP
diff --git a/src/cmfd_execute.F90 b/src/cmfd_execute.F90
index 44e94a2aa6..7a35a12144 100644
--- a/src/cmfd_execute.F90
+++ b/src/cmfd_execute.F90
@@ -5,81 +5,57 @@ module cmfd_execute
 ! cross section generation, diffusion calculation, and source re-weighting
 !==============================================================================
 
+  use global
+
   implicit none
   private
   public :: execute_cmfd, cmfd_init_batch
 
-# ifdef PETSC
-#   include <finclude/petsc.h90>
-# endif
-
 contains
 
 !==============================================================================
-! EXECUTE_CMFD
+! EXECUTE_CMFD runs the CMFD calculation
 !==============================================================================
 
   subroutine execute_cmfd()
 
-# ifdef PETSC
-
     use cmfd_data,              only: set_up_cmfd
-    use cmfd_message_passing,   only: petsc_init_mpi, cmfd_bcast
     use cmfd_power_solver,      only: cmfd_power_execute
-    use cmfd_snes_solver,       only: cmfd_snes_execute
+    use cmfd_jfnk_solver,       only: cmfd_jfnk_execute
     use error,                  only: warning, fatal_error 
-    use global,                 only: n_procs_cmfd, cmfd,                       &
-                                      cmfd_solver_type, time_cmfd,              &
-                                      cmfd_run_adjoint, cmfd_write_hdf5,        &
-                                      cmfd_feedback,cmfd_hold_weights,          &
-                                      cmfd_inact_flush, cmfd_keff_tol,          &
-                                      cmfd_act_flush, current_batch, keff,      &
-                                      n_batches, message, master, mpi_err, rank
 
-    ! stop cmfd timer
+    ! CMFD single processor on master
     if (master) then
+
+      ! Start cmfd timer
       call time_cmfd % start()
-    end if
 
-    ! filter processors (lowest PETSc group)
-    if (rank < n_procs_cmfd) then
+      ! Create cmfd data from OpenMC tallies 
+      call set_up_cmfd()
 
-      ! set up cmfd data (master only)
-      if (master) call set_up_cmfd()
-
-      ! broadcast cmfd to all petsc procs
-      call cmfd_bcast()
-
-      ! process solver options
+      ! Process solver options
       call process_cmfd_options()
 
-    end if
-
-    ! filter processors (lowest PETSc group)
-    if (rank < n_procs_cmfd) then
-
-      ! call solver
+      ! Call solver
       if (trim(cmfd_solver_type) == 'power') then
         call cmfd_power_execute()
       elseif (trim(cmfd_solver_type) == 'jfnk') then
-        call cmfd_snes_execute()
+        call cmfd_jfnk_execute()
       else
         message = 'solver type became invalid after input processing'
         call fatal_error() 
       end if
 
-      ! perform any last batch tasks 
-      if (current_batch == n_batches) then
+      ! Save k-effective
+      cmfd % k_cmfd(current_batch) = cmfd % keff
 
-        ! check for adjoint run
-        if (cmfd_run_adjoint) then
-          if (trim(cmfd_solver_type) == 'power') then
-            call cmfd_power_execute(adjoint = .true.)
-          elseif (trim(cmfd_solver_type) == 'jfnk') then
-            call cmfd_snes_execute(adjoint = .true.)
-          end if
+      ! check to perform adjoint on last batch
+      if (current_batch == n_batches .and. cmfd_run_adjoint) then
+        if (trim(cmfd_solver_type) == 'power') then
+          call cmfd_power_execute(adjoint = .true.)
+        elseif (trim(cmfd_solver_type) == 'jfnk') then
+          call cmfd_jfnk_execute(adjoint = .true.)
         end if
-
       end if
 
     end if
@@ -91,19 +67,12 @@ contains
     if (cmfd_feedback) call cmfd_reweight(.true.)
 
     ! stop cmfd timer
-    if (master) then
-      call time_cmfd % stop()
-    end if
-
-    ! wait here for all procs
-    call MPI_Barrier(MPI_COMM_WORLD, mpi_err)
-
-# endif
+    if (master) call time_cmfd % stop()
 
   end subroutine execute_cmfd
 
 !==============================================================================
-! CMFD_INIT_BATCH
+! CMFD_INIT_BATCH handles cmfd options at the start of every batch
 !==============================================================================
 
   subroutine cmfd_init_batch()
@@ -112,7 +81,7 @@ contains
                                  cmfd_inact_flush, cmfd_act_flush, cmfd_run, &
                                  current_batch, cmfd_hold_weights
 
-    ! check to activate CMFD diffusion and possible feedback
+    ! Check to activate CMFD diffusion and possible feedback
     ! this guarantees that when cmfd begins at least one batch of tallies are
     ! accumulated
     if (cmfd_run .and. cmfd_begin == current_batch) then
@@ -120,18 +89,19 @@ contains
       cmfd_tally_on = .true.
     end if
 
-    ! check to flush cmfd tallies for active batches, no more inactive flush
+    ! If this is a restart run and we are just replaying batches leave
+    if (restart_run .and. current_batch <= restart_batch) return
+
+    ! Check to flush cmfd tallies for active batches, no more inactive flush
     if (cmfd_run .and. cmfd_act_flush == current_batch) then
       call cmfd_tally_reset()
       cmfd_tally_on = .true.
       cmfd_inact_flush(2) = -1
     end if
 
-    ! check to flush cmfd tallies during inactive batches (>= on number of
+    ! Check to flush cmfd tallies during inactive batches (>= on number of
     ! flushes important as the code will flush on the first batch which we
     ! dont want to count)
-!    if (cmfd_run .and. current_batch < n_inactive .and. mod(current_batch-1,cmfd_inact_flush(1))   &
-!       == 0 .and. cmfd_inact_flush(2) >= 0) then
     if (cmfd_run .and. mod(current_batch,cmfd_inact_flush(1))   &
        == 0 .and. cmfd_inact_flush(2) > 0 .and. cmfd_begin < current_batch) then
         cmfd_hold_weights = .true.
@@ -141,23 +111,23 @@ contains
 
   end subroutine cmfd_init_batch
 
-# ifdef PETSC
-
 !==============================================================================
-! PROCESS_CMFD_OPTIONS 
+! PROCESS_CMFD_OPTIONS processes user options that interface with PETSc 
 !==============================================================================
 
   subroutine process_cmfd_options()
 
     use global,       only: cmfd_snes_monitor, cmfd_ksp_monitor, mpi_err
 
-    ! check for snes monitor
+#ifdef PETSC
+    ! Check for snes monitor
     if (cmfd_snes_monitor) call PetscOptionsSetValue("-snes_monitor", &
          "stdout", mpi_err)
 
-    ! check for ksp monitor
+    ! Check for ksp monitor
     if (cmfd_ksp_monitor) call PetscOptionsSetValue("-ksp_monitor", &
          "stdout", mpi_err)
+#endif
 
     end subroutine process_cmfd_options
 
@@ -168,39 +138,44 @@ contains
   subroutine calc_fission_source()
 
     use constants,  only: CMFD_NOACCEL, ZERO, TWO
-    use global,     only: cmfd, cmfd_coremap, master, mpi_err, entropy_on
+    use global,     only: cmfd, cmfd_coremap, master, mpi_err, entropy_on, &
+                          current_batch
+#ifdef MPI
+    use mpi
+#endif
 
-    integer :: nx ! maximum number of cells in x direction
-    integer :: ny ! maximum number of cells in y direction
-    integer :: nz ! maximum number of cells in z direction
-    integer :: ng ! maximum number of energy groups
-    integer :: n  ! total size
-    integer :: i ! iteration counter for x
-    integer :: j ! iteration counter for y
-    integer :: k ! iteration counter for z
-    integer :: g ! iteration counter for groups
-    integer :: idx ! index in vector
+    integer :: nx      ! maximum number of cells in x direction
+    integer :: ny      ! maximum number of cells in y direction
+    integer :: nz      ! maximum number of cells in z direction
+    integer :: ng      ! maximum number of energy groups
+    integer :: n       ! total size
+    integer :: i       ! iteration counter for x
+    integer :: j       ! iteration counter for y
+    integer :: k       ! iteration counter for z
+    integer :: g       ! iteration counter for groups
+    integer :: idx     ! index in vector
     real(8) :: hxyz(3) ! cell dimensions of current ijk cell
     real(8) :: vol     ! volume of cell
     real(8),allocatable :: source(:,:,:,:)  ! tmp source array for entropy
 
-    ! get maximum of spatial and group indices
-    nx = cmfd%indices(1)
-    ny = cmfd%indices(2)
-    nz = cmfd%indices(3)
-    ng = cmfd%indices(4)
+    ! Get maximum of spatial and group indices
+    nx = cmfd % indices(1)
+    ny = cmfd % indices(2)
+    nz = cmfd % indices(3)
+    ng = cmfd % indices(4)
     n  = ng*nx*ny*nz
 
-    ! allocate cmfd source if not already allocated and allocate buffer
-    if (.not. allocated(cmfd%cmfd_src)) allocate(cmfd%cmfd_src(ng,nx,ny,nz))
+    ! Allocate cmfd source if not already allocated and allocate buffer
+    if (.not. allocated(cmfd % cmfd_src)) &
+       allocate(cmfd % cmfd_src(ng,nx,ny,nz))
 
-    ! reset cmfd source to 0
-    cmfd%cmfd_src = ZERO
+    ! Reset cmfd source to 0
+    cmfd % cmfd_src = ZERO
 
-    ! only perform for master
+    ! Only perform for master
     if (master) then
 
-      ! loop around indices to map to cmfd object
+      ! Loop around indices to map to cmfd object
       ZLOOP: do k = 1, nz
 
         YLOOP: do j = 1, ny
@@ -209,25 +184,25 @@ contains
 
             GROUP: do g = 1, ng
 
-              ! check for core map
+              ! Check for core map
               if (cmfd_coremap) then
-                if (cmfd%coremap(i,j,k) == CMFD_NOACCEL) then
+                if (cmfd % coremap(i,j,k) == CMFD_NOACCEL) then
                   cycle
                 end if
               end if
 
-              ! get dimensions of cell
-              hxyz = cmfd%hxyz(:,i,j,k)
+              ! Get dimensions of cell
+              hxyz = cmfd % hxyz(:,i,j,k)
 
-              ! calculate volume
+              ! Calculate volume
               vol = hxyz(1)*hxyz(2)*hxyz(3)
 
-              ! get first index
+              ! Get first index
               idx = get_matrix_idx(1,i,j,k,ng,nx,ny)
 
-              ! compute fission source
-              cmfd%cmfd_src(g,i,j,k) = sum(cmfd%nfissxs(:,g,i,j,k) * &
-                   cmfd%phi(idx:idx+(ng-1)))*vol
+              ! Compute fission source
+              cmfd % cmfd_src(g,i,j,k) = sum(cmfd % nfissxs(:,g,i,j,k) * &
+                     cmfd % phi(idx:idx + (ng - 1)))*vol
 
             end do GROUP
 
@@ -237,43 +212,49 @@ contains
 
       end do ZLOOP
 
-      ! normalize source such that it sums to 1.0
-      cmfd%cmfd_src = cmfd%cmfd_src/sum(cmfd%cmfd_src)
+      ! Normalize source such that it sums to 1.0
+      cmfd % cmfd_src = cmfd % cmfd_src/sum(cmfd % cmfd_src)
 
-      ! compute entropy
+      ! Compute entropy
       if (entropy_on) then
 
-        ! allocate tmp array
+        ! Allocate tmp array
         if (.not.allocated(source)) allocate(source(ng,nx,ny,nz))
 
-        ! initialize the source
+        ! Initialize the source
         source = ZERO
 
-        ! compute log
-        where (cmfd%cmfd_src > ZERO)
-          source = cmfd%cmfd_src*log(cmfd%cmfd_src)/log(TWO)
+        ! Compute log
+        where (cmfd % cmfd_src > ZERO)
+          source = cmfd % cmfd_src*log(cmfd % cmfd_src)/log(TWO)
         end where
 
-        ! sum that source
-        cmfd%entropy = -sum(source)
+        ! Sum that source
+        cmfd % entropy(current_batch) = -sum(source)
 
-        ! deallocate tmp array
+        ! Deallocate tmp array
         if (allocated(source)) deallocate(source)
 
       end if
 
-      ! normalize source so average is 1.0
-      cmfd%cmfd_src = cmfd%cmfd_src/sum(cmfd%cmfd_src)*cmfd%norm
+      ! Normalize source so average is 1.0
+      cmfd % cmfd_src = cmfd % cmfd_src/sum(cmfd % cmfd_src)*cmfd % norm
+
+      ! Calculate differences between normalized sources
+      cmfd % src_cmp(current_batch) = sqrt(ONE/cmfd % norm * &
+             sum((cmfd % cmfd_src - cmfd % openmc_src)**2))
 
     end if
 
-    ! broadcast full source to all procs
-    call MPI_BCAST(cmfd%cmfd_src, n, MPI_REAL8, 0, MPI_COMM_WORLD, mpi_err)
+#ifdef MPI
+    ! Broadcast full source to all procs
+    call MPI_BCAST(cmfd % cmfd_src, n, MPI_REAL8, 0, MPI_COMM_WORLD, mpi_err)
+#endif
 
   end subroutine calc_fission_source
 
 !===============================================================================
-! CMFD_REWEIGHT
+! CMFD_REWEIGHT performs weighting of particles in the source bank
 !===============================================================================
 
   subroutine cmfd_reweight(new_weights)
@@ -286,110 +267,116 @@ contains
     use mesh,        only: count_bank_sites, get_mesh_indices
     use search,      only: binary_search
 
-    ! local variables
-    integer :: nx ! maximum number of cells in x direction
-    integer :: ny ! maximum number of cells in y direction
-    integer :: nz ! maximum number of cells in z direction
-    integer :: ng ! maximum number of energy groups
-    integer :: i ! iteration counter
-    integer :: ijk(3) ! spatial bin location
-    integer :: e_bin ! energy bin of source particle
-    integer :: n_groups ! number of energy groups
-    logical :: outside ! any source sites outside mesh
-    logical :: in_mesh ! source site is inside mesh
-    logical :: new_weights ! calcualte new weights
-    type(StructuredMesh), pointer :: m ! point to mesh
-    real(8), allocatable :: egrid(:)
+#ifdef MPI
+    use mpi
+#endif
 
-    ! associate pointer
+    logical, intent(in) :: new_weights ! calcualte new weights
+
+    integer :: nx       ! maximum number of cells in x direction
+    integer :: ny       ! maximum number of cells in y direction
+    integer :: nz       ! maximum number of cells in z direction
+    integer :: ng       ! maximum number of energy groups
+    integer :: i        ! iteration counter
+    integer :: ijk(3)   ! spatial bin location
+    integer :: e_bin    ! energy bin of source particle
+    integer :: n_groups ! number of energy groups
+    logical :: outside  ! any source sites outside mesh
+    logical :: in_mesh  ! source site is inside mesh
+
+    type(StructuredMesh), pointer :: m ! point to mesh
+    real(8), allocatable :: egrid(:)   ! energy grid
+
+    ! Associate pointer
     m => meshes(n_user_meshes + 1)
 
-    ! get maximum of spatial and group indices
-    nx = cmfd%indices(1)
-    ny = cmfd%indices(2)
-    nz = cmfd%indices(3)
-    ng = cmfd%indices(4)
+    ! Get maximum of spatial and group indices
+    nx = cmfd % indices(1)
+    ny = cmfd % indices(2)
+    nz = cmfd % indices(3)
+    ng = cmfd % indices(4)
 
     ! allocate arrays in cmfd object (can take out later extend to multigroup)
     if (.not.allocated(cmfd%sourcecounts)) then 
       allocate(cmfd%sourcecounts(ng,nx,ny,nz))
       cmfd % sourcecounts = 0
     end if
-    if (.not.allocated(cmfd%weightfactors)) then 
-      allocate(cmfd%weightfactors(ng,nx,ny,nz))
+    if (.not.allocated(cmfd % weightfactors)) then 
+      allocate(cmfd % weightfactors(ng,nx,ny,nz))
       cmfd % weightfactors = ONE
     end if
 
-    ! allocate energy grid and reverse cmfd energy grid
-    if (.not. allocated(egrid)) allocate(egrid(ng+1))
-    egrid = (/(cmfd%egrid(ng-i+2),i = 1,ng+1)/)
+    ! Allocate energy grid and reverse cmfd energy grid
+    if (.not. allocated(egrid)) allocate(egrid(ng + 1))
+    egrid = (/(cmfd % egrid(ng - i + 2), i = 1, ng + 1)/)
 
-    ! compute new weight factors
+    ! Compute new weight factors
     if (new_weights) then
 
-      ! zero out weights
+      ! Zero out weights
       cmfd%weightfactors = ZERO
 
-      ! count bank sites in mesh
+      ! Count bank sites in mesh
       call count_bank_sites(m, source_bank, cmfd%sourcecounts, egrid, &
            sites_outside=outside, size_bank=work)
 
-      ! check for sites outside of the mesh
+      ! Check for sites outside of the mesh
       if (master .and. outside) then
         message = "Source sites outside of the CMFD mesh!"
         call fatal_error()
       end if
 
-      ! have master compute weight factors
+      ! Have master compute weight factors (watch for 0s)
       if (master) then
-        where(cmfd%cmfd_src > ZERO .and. cmfd%sourcecounts > ZERO)
-          cmfd%weightfactors = cmfd%cmfd_src/sum(cmfd%cmfd_src)* &
-                               sum(cmfd%sourcecounts) / cmfd%sourcecounts
+        where(cmfd % cmfd_src > ZERO .and. cmfd % sourcecounts > ZERO)
+          cmfd % weightfactors = cmfd % cmfd_src/sum(cmfd % cmfd_src)* &
+                               sum(cmfd % sourcecounts) / cmfd % sourcecounts
         end where
       end if
 
-      ! broadcast weight factors to all procs
-      call MPI_BCAST(cmfd%weightfactors, ng*nx*ny*nz, MPI_REAL8, 0, &
+      ! Broadcast weight factors to all procs
+#ifdef MPI
+      call MPI_BCAST(cmfd % weightfactors, ng*nx*ny*nz, MPI_REAL8, 0, &
            MPI_COMM_WORLD, mpi_err)
-
+#endif
    end if
 
     ! begin loop over source bank
     do i = 1, int(work,4)
 
-      ! determine spatial bin
-      call get_mesh_indices(m, source_bank(i)%xyz, ijk, in_mesh)
+      ! Determine spatial bin
+      call get_mesh_indices(m, source_bank(i) % xyz, ijk, in_mesh)
 
-      ! determine energy bin
-      n_groups = size(cmfd%egrid) - 1
-      if (source_bank(i) % E < cmfd%egrid(1)) then
+      ! Determine energy bin
+      n_groups = size(cmfd % egrid) - 1
+      if (source_bank(i) % E < cmfd % egrid(1)) then
         e_bin = 1
-        message = 'source pt below energy grid'
+        message = 'Source pt below energy grid'
         call warning()
-      elseif (source_bank(i) % E > cmfd%egrid(n_groups+1)) then
+      elseif (source_bank(i) % E > cmfd % egrid(n_groups + 1)) then
         e_bin = n_groups
-        message = 'source pt above energy grid'
+        message = 'Source pt above energy grid'
         call warning()
       else
-        e_bin = binary_search(cmfd%egrid, n_groups + 1, source_bank(i) % E)
+        e_bin = binary_search(cmfd % egrid, n_groups + 1, source_bank(i) % E)
       end if
 
-      ! reverese energy bin (lowest grp is highest energy bin)
+      ! Reverese energy bin (lowest grp is highest energy bin)
       e_bin = n_groups - e_bin + 1
 
-      ! check for outside of mesh
+      ! Check for outside of mesh
       if (.not. in_mesh) then
-        message = 'Source site found outside of CMFD mesh!'
+        message = 'Source site found outside of CMFD mesh'
         call fatal_error()
       end if
 
-      ! reweight particle
-      source_bank(i)%wgt = source_bank(i)%wgt * &
-           cmfd%weightfactors(e_bin,ijk(1),ijk(2),ijk(3))
+      ! Reweight particle
+      source_bank(i) % wgt = source_bank(i) % wgt * &
+             cmfd % weightfactors(e_bin, ijk(1), ijk(2), ijk(3))
 
     end do
 
-    ! deallocate
+    ! Deallocate all
     if (allocated(egrid)) deallocate(egrid)
 
   end subroutine cmfd_reweight
@@ -403,33 +390,31 @@ contains
     use global, only: cmfd, cmfd_coremap
 
     integer :: matidx ! the index location in matrix
-    integer :: i      ! current x index
-    integer :: j      ! current y index
-    integer :: k      ! current z index
-    integer :: g      ! current group index
-    integer :: nx     ! maximum number of cells in x direction
-    integer :: ny     ! maximum number of cells in y direction
-    integer :: ng     ! maximum number of energy groups
+    integer, intent(in) :: i  ! current x index
+    integer, intent(in) :: j  ! current y index
+    integer, intent(in) :: k  ! current z index
+    integer, intent(in) :: g  ! current group index
+    integer, intent(in) :: nx ! maximum number of cells in x direction
+    integer, intent(in) :: ny ! maximum number of cells in y direction
+    integer, intent(in) :: ng ! maximum number of energy groups
 
-    ! check if coremap is used
+    ! Check if coremap is used
     if (cmfd_coremap) then
 
-      ! get idx from core map
+      ! Get idx from core map
       matidx = ng*(cmfd % coremap(i,j,k)) - (ng - g)
 
     else
 
-      ! compute index
+      ! Compute index
       matidx = g + ng*(i - 1) + ng*nx*(j - 1) + ng*nx*ny*(k - 1)
 
     end if
 
   end function get_matrix_idx
 
-# endif
-
 !===============================================================================
-! CMFD_TALLY_RESET
+! CMFD_TALLY_RESET resets all cmfd tallies
 !===============================================================================
 
   subroutine cmfd_tally_reset()
@@ -440,14 +425,14 @@ contains
 
     integer :: i ! loop counter
 
-    ! print message
+    ! Print message
     message = "CMFD tallies reset"
     call write_message(7)
 
-    ! begin loop around CMFD tallies
+    ! Begin loop around CMFD tallies
     do i = 1, n_cmfd_tallies
 
-      ! reset that tally
+      ! Reset that tally
       cmfd_tallies(i) % n_realizations = 0
       call reset_result(cmfd_tallies(i) % results)
 
diff --git a/src/cmfd_header.F90 b/src/cmfd_header.F90
index c4dad04f2a..3d7786ab8e 100644
--- a/src/cmfd_header.F90
+++ b/src/cmfd_header.F90
@@ -8,93 +8,106 @@ module cmfd_header
 
   type, public :: cmfd_type
 
-    ! indices for problem
+    ! Indices for problem
     integer :: indices(4)
 
-    ! albedo boundary condition
+    ! Albedo boundary condition
     real(8) :: albedo(6)
 
-    ! core overlay map
+    ! Core overlay map
     integer, allocatable :: coremap(:,:,:)
     integer, allocatable :: indexmap(:,:)
     integer :: mat_dim = CMFD_NOACCEL 
 
-    ! energy grid
+    ! Energy grid
     real(8), allocatable :: egrid(:)
 
-    ! cross sections
+    ! Cross sections
     real(8), allocatable :: totalxs(:,:,:,:)
     real(8), allocatable :: p1scattxs(:,:,:,:)
     real(8), allocatable :: scattxs(:,:,:,:,:)
     real(8), allocatable :: nfissxs(:,:,:,:,:)
 
-    ! diffusion coefficient
+    ! Diffusion coefficient
     real(8), allocatable :: diffcof(:,:,:,:)
 
-    ! current
+    ! Current
     real(8), allocatable :: current(:,:,:,:,:)
 
-    ! flux
+    ! Flux
     real(8), allocatable :: flux(:,:,:,:)
 
-    ! coupling coefficients and equivalence parameters
+    ! Coupling coefficients and equivalence parameters
     real(8), allocatable :: dtilde(:,:,:,:,:)
     real(8), allocatable :: dhat(:,:,:,:,:)
 
-    ! dimensions of mesh cells ([hu,hv,hw],xloc,yloc,zloc)
+    ! Dimensions of mesh cells ([hu,hv,hw],xloc,yloc,zloc)
     real(8), allocatable :: hxyz(:,:,:,:)
 
-    ! source distributions
+    ! Source distributions
     real(8), allocatable :: cmfd_src(:,:,:,:)
     real(8), allocatable :: openmc_src(:,:,:,:)
 
-    ! source sites in each mesh box
+    ! Source sites in each mesh box
     real(8), allocatable :: sourcecounts(:,:,:,:)
 
-    ! weight adjustment factors 
+    ! Weight adjustment factors 
     real(8), allocatable :: weightfactors(:,:,:,:)
 
-    ! eigenvector/eigenvalue from cmfd run
+    ! Eigenvector/eigenvalue from cmfd run
     real(8), allocatable :: phi(:)
     real(8) :: keff = ZERO
 
-    ! eigenvector/eigenvalue from adjoint run
+    ! Eigenvector/eigenvalue from adjoint run
     real(8), allocatable :: adj_phi(:)
     real(8) :: adj_keff = ZERO
 
-    ! residual for neutron balance
+    ! Residual for neutron balance
     real(8), allocatable :: resnb(:,:,:,:)
 
-    ! openmc source normalization factor
+    ! Openmc source normalization factor
     real(8) :: norm = ONE
 
-    ! Shannon entropy from cmfd fission source
-    real(8) :: entropy 
+    ! "Shannon entropy" from cmfd fission source
+    real(8), allocatable :: entropy(:)
+
+    ! RMS of neutron balance equations
+    real(8), allocatable :: balance(:)
+
+    ! RMS deviation of OpenMC and CMFD normalized source
+    real(8), allocatable :: src_cmp(:)
+
+    ! Dominance ratio
+    real(8), allocatable :: dom(:)
+
+    ! List of CMFD k
+    real(8), allocatable :: k_cmfd(:)
 
   end type cmfd_type
 
 contains
 
 !==============================================================================
-! ALLOCATE_CMFD
+! ALLOCATE_CMFD allocates all data in of cmfd type
 !==============================================================================
 
-  subroutine allocate_cmfd(this)
+  subroutine allocate_cmfd(this, n_batches)
 
-    type(cmfd_type) :: this 
+    integer, intent(in)            :: n_batches ! number of batches in calc
+    type(cmfd_type), intent(inout) :: this      ! cmfd instance
 
     integer :: nx  ! number of mesh cells in x direction
     integer :: ny  ! number of mesh cells in y direction
     integer :: nz  ! number of mesh cells in z direction
     integer :: ng  ! number of energy groups
 
-   ! extract spatial and energy indices from object
+   ! Extract spatial and energy indices from object
     nx = this % indices(1)
     ny = this % indices(2)
     nz = this % indices(3)
     ng = this % indices(4)
 
-    ! allocate flux, cross sections and diffusion coefficient
+    ! Allocate flux, cross sections and diffusion coefficient
     if (.not. allocated(this % flux))       allocate(this % flux(ng,nx,ny,nz))
     if (.not. allocated(this % totalxs))    allocate(this % totalxs(ng,nx,ny,nz))
     if (.not. allocated(this % p1scattxs))  allocate(this % p1scattxs(ng,nx,ny,nz))
@@ -102,25 +115,32 @@ contains
     if (.not. allocated(this % nfissxs))    allocate(this % nfissxs(ng,ng,nx,ny,nz))
     if (.not. allocated(this % diffcof))    allocate(this % diffcof(ng,nx,ny,nz))
 
-    ! allocate dtilde and dhat 
+    ! Allocate dtilde and dhat 
     if (.not. allocated(this % dtilde))     allocate(this % dtilde(6,ng,nx,ny,nz))
     if (.not. allocated(this % dhat))       allocate(this % dhat(6,ng,nx,ny,nz))
 
-    ! allocate dimensions for each box (here for general case)
+    ! Allocate dimensions for each box (here for general case)
     if (.not. allocated(this % hxyz))       allocate(this % hxyz(3,nx,ny,nz))
 
-    ! allocate surface currents
+    ! Allocate surface currents
     if (.not. allocated(this % current))    allocate(this % current(12,ng,nx,ny,nz))
 
-    ! allocate source distributions
+    ! Allocate source distributions
     if (.not. allocated(this % cmfd_src)) allocate(this % cmfd_src(ng,nx,ny,nz))
     if (.not. allocated(this % openmc_src)) allocate(this % openmc_src(ng,nx,ny,nz))
 
-    ! allocate source weight modification vars
+    ! Allocate source weight modification vars
     if (.not. allocated(this % sourcecounts)) allocate(this % sourcecounts(ng,nx,ny,nz))
     if (.not. allocated(this % weightfactors)) allocate(this % weightfactors(ng,nx,ny,nz))
 
-    ! set everthing to 0 except weight multiply factors if feedback isnt on
+    ! Allocate batchwise parameters
+    if (.not. allocated(this % entropy)) allocate(this % entropy(n_batches))
+    if (.not. allocated(this % balance)) allocate(this % balance(n_batches))
+    if (.not. allocated(this % src_cmp)) allocate(this % src_cmp(n_batches))
+    if (.not. allocated(this % dom)) allocate(this % dom(n_batches))
+    if (.not. allocated(this % k_cmfd)) allocate(this % k_cmfd(n_batches))
+
+    ! Set everthing to 0 except weight multiply factors if feedback isnt on
     this % flux          = ZERO
     this % totalxs       = ZERO
     this % p1scattxs     = ZERO
@@ -135,16 +155,21 @@ contains
     this % openmc_src    = ZERO
     this % sourcecounts  = ZERO
     this % weightfactors = ONE
+    this % balance       = ZERO
+    this % src_cmp       = ZERO
+    this % dom           = ZERO
+    this % k_cmfd        = ZERO
+    this % entropy       = ZERO
 
   end subroutine allocate_cmfd
 
 !===============================================================================
-! DEALLOCATE_CMFD 
+! DEALLOCATE_CMFD frees all memory of cmfd type 
 !===============================================================================
 
   subroutine deallocate_cmfd(this)
 
-    type(cmfd_type) :: this
+    type(cmfd_type), intent(inout) :: this ! cmfd instance
 
     if (allocated(this % egrid))         deallocate(this % egrid)
     if (allocated(this % totalxs))       deallocate(this % totalxs)
@@ -164,6 +189,11 @@ contains
     if (allocated(this % weightfactors)) deallocate(this % weightfactors)
     if (allocated(this % cmfd_src))      deallocate(this % cmfd_src)
     if (allocated(this % openmc_src))    deallocate(this % openmc_src)
+    if (allocated(this % balance))       deallocate(this % balance)
+    if (allocated(this % src_cmp))       deallocate(this % src_cmp)
+    if (allocated(this % dom))           deallocate(this % dom)
+    if (allocated(this % k_cmfd))        deallocate(this % k_cmfd)
+    if (allocated(this % entropy))       deallocate(this % entropy)
 
   end subroutine deallocate_cmfd
 
diff --git a/src/cmfd_input.F90 b/src/cmfd_input.F90
index c228890d1e..d893d7bc74 100644
--- a/src/cmfd_input.F90
+++ b/src/cmfd_input.F90
@@ -1,5 +1,11 @@
 module cmfd_input
 
+  use global
+
+#ifdef PETSC
+  use petscsys
+#endif
+
   implicit none
   private
   public :: configure_cmfd 
@@ -7,23 +13,47 @@ module cmfd_input
 contains
 
 !===============================================================================
-! CONFIGURE_CMFD
+! CONFIGURE_CMFD initializes PETSc and CMFD parameters
 !===============================================================================
 
   subroutine configure_cmfd()
 
-# ifdef PETSC
-    use cmfd_message_passing,   only: petsc_init_mpi
-# endif
+    use cmfd_header,  only: allocate_cmfd
 
-    ! read in cmfd input file
+    integer :: new_comm ! new mpi communicator
+    integer :: color    ! color group of processor
+
+    ! Read in cmfd input file
     call read_cmfd_xml()
 
-    ! initialize petsc on mpi
-# ifdef PETSC
-    call petsc_init_mpi()
+    ! Assign color
+    if (master) then
+      color = 1
+    else
+      color = 2
+    end if
+
+    ! Split up procs
+# ifdef PETSC 
+    call MPI_COMM_SPLIT(MPI_COMM_WORLD, color, 0, new_comm, mpi_err)
 # endif
 
+    ! assign to PETSc
+# ifdef PETSC
+    PETSC_COMM_WORLD = new_comm
+
+    ! Initialize PETSc on all procs
+    call PetscInitialize(PETSC_NULL_CHARACTER, mpi_err)
+# endif
+
+    ! Initialize timers
+    call time_cmfd % reset()
+    call time_cmfdbuild % reset()
+    call time_cmfdsolve % reset()
+
+    ! Allocate cmfd object
+    call allocate_cmfd(cmfd, n_batches)
+
   end subroutine configure_cmfd
 
 !===============================================================================
@@ -32,18 +62,17 @@ contains
 
   subroutine read_cmfd_xml()
     
-    use error,   only: fatal_error
-    use global
+    use error,   only: fatal_error, warning
     use output,  only: write_message
     use string,  only: lower_case
     use xml_data_cmfd_t
     use, intrinsic :: ISO_FORTRAN_ENV
 
-    integer :: ng
-    logical :: file_exists ! does cmfd.xml exist?
-    character(MAX_LINE_LEN) :: filename
+    integer :: ng                       ! number of energy groups
+    logical :: file_exists              ! does cmfd.xml exist?
+    character(MAX_LINE_LEN) :: filename ! name of input file
 
-    ! read cmfd infput file
+    ! Read cmfd input file
     filename = trim(path_input) // "cmfd.xml"
     inquire(FILE=filename, EXIST=file_exists)
     if (.not. file_exists) then
@@ -55,79 +84,78 @@ contains
       return
     else
 
-      ! tell user
+      ! Tell user
       message = "Reading CMFD XML file..."
       call write_message(5)
 
     end if
 
-    ! parse cmfd.xml file
+    ! Parse cmfd.xml file
     call read_xml_file_cmfd_t(filename)
 
-    ! set spatial dimensions in cmfd object
-    cmfd % indices(1:3) = mesh_ % dimension(1:3) ! sets spatial dimensions
+    ! Set spatial dimensions in cmfd object (structed Cartesian mesh)
+    cmfd % indices(1:3) = mesh_ % dimension(1:3)
 
-    ! get number of energy groups
+    ! Get number of energy groups or set to 1 group default
     if (associated(mesh_ % energy)) then
       ng = size(mesh_ % energy)
-      if(.not.allocated(cmfd%egrid)) allocate(cmfd%egrid(ng))
-      cmfd%egrid = mesh_ % energy 
+      if(.not.allocated(cmfd % egrid)) allocate(cmfd % egrid(ng))
+      cmfd % egrid = mesh_ % energy 
       cmfd % indices(4) = ng - 1 ! sets energy group dimension
     else
-      if(.not.allocated(cmfd%egrid)) allocate(cmfd%egrid(2))
-      cmfd%egrid = (/0.0_8,20.0_8/)
+      if(.not.allocated(cmfd % egrid)) allocate(cmfd % egrid(2))
+      cmfd % egrid = (/0.0_8,20.0_8/)
       cmfd % indices(4) = 1 ! one energy group
     end if
     
-    ! set global albedo
+    ! Set global albedo, these can be overwritten by coremap
     if (associated(mesh_ % albedo)) then
       cmfd % albedo = mesh_ % albedo
     else
       cmfd % albedo = (/1.0, 1.0, 1.0, 1.0, 1.0, 1.0/)
     end if
 
-    ! get acceleration map
+    ! Get core map overlay for a subset mesh for CMFD 
     if (associated(mesh_ % map)) then
+
+      ! Allocate a core map with appropriate dimensions
       allocate(cmfd % coremap(cmfd % indices(1), cmfd % indices(2), &
            cmfd % indices(3)))
+
+      ! Check and make sure it is of correct size
       if (size(mesh_ % map) /= product(cmfd % indices(1:3))) then
-        message = 'FATAL==>CMFD coremap not to correct dimensions'
+        message = 'CMFD coremap not to correct dimensions'
         call fatal_error() 
       end if
-      cmfd % coremap = reshape(mesh_ % map,(cmfd % indices(1:3)))
-      cmfd_coremap = .true.
-   end if
 
-    ! check for core map activation by printing note
-    if (cmfd_coremap .and. master) then
+      ! Reshape core map vector into (x,y,z) array
+      cmfd % coremap = reshape(mesh_ % map,(cmfd % indices(1:3)))
+
+      ! Indicate to cmfd that a core map overlay is active
+      cmfd_coremap = .true.
+
+      ! Write to stdout that a core map overlay is active
       message = "Core Map Overlay Activated"
-      call write_message()
+      call write_message(10)
+
     end if
 
-    ! check for normalization constant
+    ! Get normalization constant for source (default is 1.0 from XML) 
     cmfd % norm = norm_
 
-    ! set feedback logical
+    ! Set feedback logical
     call lower_case(feedback_)
     if (feedback_ == 'true' .or. feedback_ == '1') cmfd_feedback = .true.
 
-    ! set balance logical
-    ! call lower_case(balance_)
-    ! if (balance_ == 'true' .or. balance == '1') cmfd_balance = .true.
+    ! Set downscatter logical
+    call lower_case(downscatter_)
+    if (downscatter_ == 'true' .or. downscatter_ == '1') &
+         cmfd_downscatter = .true.
 
-    ! set downscatter logical
-    ! call lower_case(downscatter_)
-    ! if (downscatter_ == 'true' .or. downscatter == '1') &
-    !      cmfd_downscatter = downscatter_
-
-    ! set 2 group fix
-    call lower_case(run_2grp_)
-    if (run_2grp_ == 'true' .or. run_2grp_ == '1') cmfd_run_2grp = .true.
-
-    ! set the solver type
+    ! Set the solver type (default power from XML)
     cmfd_solver_type = solver_(1:10)
 
-    ! set monitoring 
+    ! Set convergence monitoring 
     call lower_case(snes_monitor_)
     call lower_case(ksp_monitor_)
     call lower_case(power_monitor_)
@@ -138,46 +166,41 @@ contains
     if (power_monitor_ == 'true' .or. power_monitor_ == '1') &
          cmfd_power_monitor = .true.
 
-    ! output logicals
-    call lower_case(write_balance_)
+    ! Output logicals
     call lower_case(write_matrices_)
-    ! call lower_case(write_hdf5_)
-    if (write_balance_ == 'true' .or. write_balance_ == '1') &
-         cmfd_write_balance = .true.
     if (write_matrices_ == 'true' .or. write_matrices_ == '1') &
          cmfd_write_matrices = .true.
-    ! if (write_hdf5_ == 'true' .or. write_hdf5_ == '1') &
-    !      cmfd_write_hdf5 = .true.
 
-    ! run an adjoint calc
+    ! Run an adjoint calc
     call lower_case(run_adjoint_)
     if (run_adjoint_ == 'true' .or. run_adjoint_ == '1') &
          cmfd_run_adjoint = .true.
 
-    ! batch to begin cmfd
+    ! Batch to begin cmfd (default is 1 from XML)
     cmfd_begin = begin_
 
-    ! tally during inactive batches
+    ! Tally during inactive batches (by default we will always tally from 1)
     call lower_case(inactive_)
     if (inactive_ == 'false' .or. inactive_ == '0') cmfd_tally_on = .false.
 
-    ! inactive batch flush window
-    cmfd_inact_flush(1) = inactive_flush_
-    cmfd_inact_flush(2) = num_flushes_
+    ! Inactive batch flush window
+    cmfd_inact_flush(1) = inactive_flush_ ! the interval of batches
+    cmfd_inact_flush(2) = num_flushes_    ! number of times to do this
 
-    ! last flush before active batches
+    ! Last flush before active batches
     cmfd_act_flush = active_flush_
 
-    ! tolerance on keff
-    cmfd_keff_tol = keff_tol_
-    
-    ! create tally objects
-    call create_cmfd_tally()
+    ! Get display
+    cmfd_display = display_
+    if (trim(cmfd_display) == 'dominance' .and. &
+        trim(cmfd_solver_type) /= 'power') then
+      message = 'Dominance Ratio only aviable with power iteration solver'
+      call warning()
+      cmfd_display = ''
+    end if
 
-    ! set number of CMFD processors and report to user
-    n_procs_cmfd = n_cmfd_procs_ 
-    if (master) write(OUTPUT_UNIT,'(A,1X,I0,1X,A)') "CMFD Running on", &
-       n_procs_cmfd," processors."
+    ! Create tally objects
+    call create_cmfd_tally()
 
   end subroutine read_cmfd_xml
 
@@ -193,7 +216,6 @@ contains
   subroutine create_cmfd_tally()
 
     use error,            only: fatal_error, warning
-    use global
     use mesh_header,      only: StructuredMesh
     use string
     use tally,            only: setup_active_cmfdtallies
@@ -206,27 +228,27 @@ contains
     integer :: ng          ! number of energy groups (default 1)
     integer :: n_filters   ! number of filters
     integer :: i_filter_mesh ! index for mesh filter
-    character(MAX_LINE_LEN) :: filename
+    character(MAX_LINE_LEN) :: filename ! name of cmfd file
     type(TallyObject),    pointer :: t => null()
     type(StructuredMesh), pointer :: m => null()
     type(TallyFilter) :: filters(N_FILTER_TYPES) ! temporary filters
 
-    ! parse cmfd.xml file
+    ! Parse cmfd.xml file
      filename = trim(path_input) // "cmfd.xml"
      call read_xml_file_cmfd_t(filename)
 
-    ! set global variables if they are 0 (this can happen if there is no tally
+    ! Set global variables if they are 0 (this can happen if there is no tally
     ! file)
     if (n_meshes == 0) n_meshes = n_user_meshes + n_cmfd_meshes
 
-    ! allocate mesh
+    ! Allocate mesh
     if (.not. allocated(meshes)) allocate(meshes(n_meshes))
     m => meshes(n_user_meshes+1)
 
-    ! set mesh id
+    ! Set mesh id
     m % id = n_user_meshes + 1 
 
-    ! set mesh type to rectangular
+    ! Set mesh type to rectangular
     m % type = LATTICE_RECT
 
     ! Determine number of dimensions for mesh
@@ -321,20 +343,20 @@ contains
     ! Add mesh to dictionary
     call mesh_dict % add_key(m % id, n_user_meshes + 1)
 
-    ! allocate tallies
+    ! Allocate tallies
     call add_tallies("cmfd", n_cmfd_tallies)
 
-    ! begin loop around tallies
+    ! Begin loop around tallies
     do i = 1, n_cmfd_tallies
 
-      ! point t to tally variable
+      ! Point t to tally variable
       t => cmfd_tallies(i)
 
-      ! set reset property
+      ! Set reset property
       call lower_case(reset_)
       if (reset_ == 'true' .or. reset_ == '1') t % reset = .true.
 
-      ! set up mesh filter
+      ! Set up mesh filter
       n_filters = 1
       filters(n_filters) % type = FILTER_MESH
       filters(n_filters) % n_bins = product(m % dimension)
@@ -342,7 +364,7 @@ contains
       filters(n_filters) % int_bins(1) = n_user_meshes + 1
       t % find_filter(FILTER_MESH) = n_filters
 
-      ! read and set incoming energy mesh filter
+      ! Read and set incoming energy mesh filter
       if (associated(mesh_ % energy)) then
         n_filters = n_filters + 1
         filters(n_filters) % type = FILTER_ENERGYIN
@@ -353,40 +375,40 @@ contains
         t % find_filter(FILTER_ENERGYIN) = n_filters
       end if
 
-      ! set number of nucilde bins
+      ! Set number of nucilde bins
       allocate(t % nuclide_bins(1))
       t % nuclide_bins(1) = -1
       t % n_nuclide_bins = 1
 
-      ! record tally id which is equivalent to loop number
+      ! Record tally id which is equivalent to loop number
       t % id = i_cmfd_tallies + i
 
       if (i == 1) then
 
-        ! set label
+        ! Set label
         t % label = "CMFD flux, total, scatter-1"
 
-        ! set tally estimator to analog
+        ! Set tally estimator to analog
         t % estimator = ESTIMATOR_ANALOG
 
-        ! set tally type to volume
+        ! Set tally type to volume
         t % type = TALLY_VOLUME
 
-        ! allocate and set filters
+        ! Allocate and set filters
         t % n_filters = n_filters
         allocate(t % filters(n_filters))
         t % filters = filters(1:n_filters)
 
-        ! allocate scoring bins 
+        ! Allocate scoring bins 
         allocate(t % score_bins(3))
         t % n_score_bins = 3
         t % n_user_score_bins = 3
 
-        ! allocate scattering order data
+        ! Allocate scattering order data
         allocate(t % scatt_order(3))
         t % scatt_order = 0
         
-        ! set macro_bins
+        ! Set macro_bins
         t % score_bins(1)  = SCORE_FLUX
         t % score_bins(2)  = SCORE_TOTAL
         t % score_bins(3)  = SCORE_SCATTER_N
@@ -394,16 +416,16 @@ contains
 
       else if (i == 2) then
 
-        ! set label
+        ! Set label
         t % label = "CMFD neutron production"
 
-        ! set tally estimator to analog
+        ! Set tally estimator to analog
         t % estimator = ESTIMATOR_ANALOG
 
-        ! set tally type to volume
+        ! Set tally type to volume
         t % type = TALLY_VOLUME
 
-        ! read and set outgoing energy mesh filter
+        ! Read and set outgoing energy mesh filter
         if (associated(mesh_ % energy)) then
           n_filters = n_filters + 1
           filters(n_filters) % type = FILTER_ENERGYOUT
@@ -414,34 +436,34 @@ contains
           t % find_filter(FILTER_ENERGYOUT) = n_filters
         end if
 
-        ! allocate and set filters
+        ! Allocate and set filters
         t % n_filters = n_filters
         allocate(t % filters(n_filters))
         t % filters = filters(1:n_filters)
 
-        ! deallocate filters bins array
+        ! Deallocate filters bins array
         if (associated(mesh_ % energy)) &
              deallocate(filters(n_filters) % real_bins)
 
-        ! allocate macro reactions
+        ! Allocate macro reactions
         allocate(t % score_bins(2))
         t % n_score_bins = 2
         t % n_user_score_bins = 2
 
-        ! allocate scattering order data
+        ! Allocate scattering order data
         allocate(t % scatt_order(2))
         t % scatt_order = 0
 
-        ! set macro_bins
+        ! Set macro_bins
         t % score_bins(1) = SCORE_NU_SCATTER
         t % score_bins(2) = SCORE_NU_FISSION
 
       else if (i == 3) then
 
-        ! set label
+        ! Set label
         t % label = "CMFD surface currents"
 
-        ! set tally estimator to analog
+        ! Set tally estimator to analog
         t % estimator = ESTIMATOR_ANALOG
 
         ! Add extra filter for surface
@@ -458,40 +480,41 @@ contains
         end if
         t % find_filter(FILTER_SURFACE) = n_filters
 
-        ! allocate and set filters
+        ! Allocate and set filters
         t % n_filters = n_filters
         allocate(t % filters(n_filters))
         t % filters = filters(1:n_filters)
 
-        ! deallocate filters bins array
+        ! Deallocate filters bins array
         deallocate(filters(n_filters) % int_bins)
 
-        ! allocate macro reactions
+        ! Allocate macro reactions
         allocate(t % score_bins(1))
         t % n_score_bins = 1
         t % n_user_score_bins = 1
 
-        ! allocate scattering order data
+        ! Allocate scattering order data
         allocate(t % scatt_order(1))
         t % scatt_order = 0
 
-        ! set macro bins
+        ! Set macro bins
         t % score_bins(1) = SCORE_CURRENT
         t % type = TALLY_SURFACE_CURRENT
 
-        ! we need to increase the dimension by one since we also need
+        ! We need to increase the dimension by one since we also need
         ! currents coming into and out of the boundary mesh cells.
         i_filter_mesh = t % find_filter(FILTER_MESH)
         t % filters(i_filter_mesh) % n_bins = product(m % dimension + 1)
 
       end if
 
-      ! deallocate filter bins
+      ! Deallocate filter bins
       deallocate(filters(1) % int_bins)
       if (associated(mesh_ % energy)) deallocate(filters(2) % real_bins)
 
     end do
 
+    ! Put cmfd tallies into active tally array and turn tallies on
     call setup_active_cmfdtallies()
     tallies_on = .true.
 
diff --git a/src/cmfd_jacobian_operator.F90 b/src/cmfd_jacobian_operator.F90
deleted file mode 100644
index b5532b66a7..0000000000
--- a/src/cmfd_jacobian_operator.F90
+++ /dev/null
@@ -1,357 +0,0 @@
-module cmfd_jacobian_operator
-
-# ifdef PETSC
-
-  use cmfd_loss_operator, only: loss_operator,init_M_operator, &
-                                build_loss_matrix,destroy_M_operator
-  use cmfd_prod_operator, only: prod_operator,init_F_operator, &
-                                build_prod_matrix,destroy_F_operator
-  implicit none
-  private
-  public :: init_J_operator, build_jacobian_matrix, destroy_J_operator
-
-# include <finclude/petsc.h90>
-
-  integer  :: nx   ! maximum number of x cells
-  integer  :: ny   ! maximum number of y cells
-  integer  :: nz   ! maximum number of z cells
-  integer  :: ng   ! maximum number of groups
-  integer  :: ierr ! petsc error code
-    
-  type, public :: jacobian_operator
-    Mat      :: J    ! petsc matrix for neutronic prod operator
-    integer  :: n    ! dimensions of matrix
-    integer  :: nnz  ! max number of nonzeros
-    integer  :: localn ! local size on proc
-    integer, allocatable :: d_nnz(:) ! vector of diagonal preallocation
-    integer, allocatable :: o_nnz(:) ! vector of off-diagonal preallocation
-  end type jacobian_operator
-
-  type, public :: operators 
-    type(loss_operator) :: loss
-    type(prod_operator) :: prod
-  end type operators
-
-contains
-
-!==============================================================================
-! INIT_J_OPERATOR
-!==============================================================================
-
-  subroutine init_J_operator(this,ctx)
-
-    type(jacobian_operator) :: this
-    type(operators)         :: ctx
-
-    ! get indices
-    call get_J_indices(this)
-
-    ! get preallocation
-    call preallocate_jacobian_matrix(this,ctx)
-
-    ! set up M operator
-    call MatCreateAIJ(PETSC_COMM_WORLD, this%localn, this%localn, PETSC_DECIDE,&
-         PETSC_DECIDE, PETSC_NULL_INTEGER, this%d_nnz, PETSC_NULL_INTEGER, &
-         this%o_nnz, this%J,ierr)
-    call MatSetOption(this%J, MAT_NEW_NONZERO_LOCATIONS, PETSC_TRUE, ierr)
-    call MatSetOption(this%J, MAT_IGNORE_ZERO_ENTRIES, PETSC_TRUE, ierr)
-
-  end subroutine init_J_operator
-
-!==============================================================================
-! GET_J_INDICES
-!==============================================================================
-
-  subroutine get_J_indices(this)
-
-    use global,  only: cmfd, cmfd_coremap
-
-    type(jacobian_operator) :: this
-
-    ! get maximum number of cells in each direction
-    nx = cmfd%indices(1)
-    ny = cmfd%indices(2)
-    nz = cmfd%indices(3)
-    ng = cmfd%indices(4)
-
-    ! get number of nonzeros
-    this%nnz = 7 + ng - 1
-
-    ! calculate dimensions of matrix
-    this%n = nx*ny*nz*ng
-
-    ! calculate dimensions of matrix
-    if (cmfd_coremap) then
-      this%n = cmfd % mat_dim * ng
-    else
-      this%n = nx*ny*nz*ng
-    end if
-
-    ! add 1 for eigenvalue row
-!   this%n = this%n + 1
-
-  end subroutine get_J_indices
-
-!===============================================================================
-! PREALLOCATE_JACOBIAN_MATRIX
-!===============================================================================
-
-  subroutine preallocate_jacobian_matrix(this,ctx)
-
-    use global,      only: cmfd, n_procs_cmfd, rank
-
-    type(jacobian_operator) :: this
-    type(operators)         :: ctx
-
-    integer :: n             ! the extent of the matrix
-    integer :: row_start     ! index of local starting row
-    integer :: row_end       ! index of local final row
-
-    ! get local problem size
-    n = this%n
-
-    ! determine local size, divide evenly between all other procs
-    this%localn = n/(n_procs_cmfd)
-
-    ! add 1 more if less proc id is less than mod
-    if (rank < mod(n,n_procs_cmfd)) this%localn = this%localn + 1
-
-    ! add another 1 on last proc
-    if (rank == n_procs_cmfd - 1) this%localn = this%localn + 1
-
-    ! determine local starting row
-    row_start = 0
-    if (rank < mod(n,n_procs_cmfd)) then
-      row_start = rank*(n/n_procs_cmfd+1)
-    else
-      row_start = min(mod(n,n_procs_cmfd)*(n/n_procs_cmfd+1) + (rank - &
-           mod(n,n_procs_cmfd))*(n/n_procs_cmfd),n)
-    end if
-
-    ! determine local final row
-    row_end = row_start + this%localn - 1
-
-    ! allocate counters
-    if (.not. allocated(this%d_nnz)) allocate(this%d_nnz(row_start:row_end))
-    if (.not. allocated(this%o_nnz)) allocate(this%o_nnz(row_start:row_end))
-    this % d_nnz = 0
-    this % o_nnz = 0
-
-    ! start with pattern from loss matrix
-    if (rank == n_procs_cmfd - 1) then
-      this % d_nnz(row_start:row_end-1) = ctx%loss%d_nnz
-      this % o_nnz(row_start:row_end-1) = ctx%loss%o_nnz
-    else
-      this % d_nnz = ctx%loss%d_nnz
-      this % o_nnz = ctx%loss%o_nnz
-    end if
-
-    ! append -F*phi term for last processor will take care of 1 for lambda
-    if (rank == n_procs_cmfd - 1) then
-      this%d_nnz = this%d_nnz + 1
-    else
-      this%o_nnz = this%o_nnz + 1
-    end if
-
-    ! do last row which has all filled (already did lower left corner above)
-    if (rank == n_procs_cmfd - 1) then
-      this % d_nnz(row_end) = this % d_nnz(row_end) + (row_end - row_start)
-      this % o_nnz(row_end) = this % o_nnz(row_end) + (this%n  - (row_end - &
-           row_start))
-    end if
-
-  end subroutine preallocate_jacobian_matrix
-
-!===============================================================================
-! BUILD_JACOBIAN_MATRIX creates the matrix representing loss of neutrons
-!===============================================================================
-
-  subroutine build_jacobian_matrix(snes,x,jac,jac_prec,flag,ctx,ierr)
-
-    use constants,    only: ZERO, ONE
-    use global,       only: n_procs_cmfd, cmfd_write_matrices, rank
-
-    SNES            :: snes      ! the snes context
-    Vec             :: x         ! the solution vector
-    Mat             :: jac       ! the jacobian matrix
-    Mat             :: jac_prec  ! the jacobian preconditioner
-    MatStructure    :: flag      ! not used
-    type(operators) :: ctx       ! not used
-    integer         :: ierr      ! petsc error flag
-
-    Vec                  :: phi      ! flux vector
-    Vec                  :: source   ! source vector
-    integer              :: n        ! problem size
-    integer              :: k        ! implied do loop counter
-    integer              :: ncols    ! number of nonzeros in cols
-    integer              :: irow     ! row counter
-    integer              :: row_start! starting local row on process
-    integer              :: row_end  ! ending local row on process
-    integer, allocatable :: dims(:)  ! vec of starting and ending rows
-    integer, allocatable :: dims1(:) ! vec of sizes on each proc
-    integer, allocatable :: cols(:)  ! vector of column numbers
-    real(8)              :: lambda   ! eigenvalue
-    real(8), pointer     :: xptr(:)  ! pointer to solution vector
-    real(8), pointer     :: sptr(:)  ! pointer to source vector
-    real(8), allocatable :: vals(:)  ! vector of row values
-    real(8), allocatable :: phi_tmp(:) ! temp buffer for flux
-
-    ! create operators
-    call build_loss_matrix(ctx%loss)
-    call build_prod_matrix(ctx%prod)
-
-    ! get problem size
-    n = ctx%loss%n
-
-    ! get local size on each processor
-    call MatGetOwnershipRange(jac_prec, row_start, row_end, ierr)
-
-    ! allocate cols and  initialize to zero
-    if (.not. allocated(cols)) allocate(cols(&
-         maxval(ctx%loss%d_nnz + ctx%loss%o_nnz)))
-    if (.not. allocated(vals)) allocate(vals(&
-         maxval(ctx%loss%d_nnz + ctx%loss%o_nnz)))
-    cols = 0
-    vals = ZERO
-
-    ! get pointers to residual vector 
-    call VecGetArrayF90(x, xptr, ierr)
-
-    ! create petsc vector for flux
-    call VecCreateMPI(PETSC_COMM_WORLD, ctx%loss%localn, PETSC_DECIDE, phi, ierr)
-
-    ! extract flux and eigenvalue
-    call VecPlaceArray(phi, xptr, ierr)
-    if (rank == n_procs_cmfd - 1) lambda = xptr(size(xptr))
-    call MPI_BCAST(lambda, 1, MPI_REAL8, n_procs_cmfd-1, PETSC_COMM_WORLD, ierr)
-
-    ! compute math (M-lambda*F) M is overwritten here
-    call MatAXPY(ctx%loss%M, -lambda, ctx%prod%F, &
-         DIFFERENT_NONZERO_PATTERN, ierr)
-
-    ! create tmp petsc vector for source
-    call VecCreateMPI(PETSC_COMM_WORLD, ctx%loss%localn, PETSC_DECIDE, &
-         source, ierr)
-
-    ! perform math (-F*phi --> source)
-    call MatMult(ctx%prod%F, phi, source, ierr)
-    call VecScale(source, -ONE, ierr)
-
-    ! get pointer to source
-    call VecGetArrayF90(source, sptr, ierr)
-
-    ! begin loop to insert things into matrix
-    do irow = row_start, row_end - 1
-
-      ! don't do last row
-      if (irow == n) cycle 
-
-      ! get row of matrix
-      call MatGetRow(ctx%loss%M, irow, ncols, cols, vals, ierr)
-
-      ! set that row to Jacobian matrix
-      call MatSetValues(jac_prec, 1, (/irow/), ncols, cols(1:ncols), vals, &
-           INSERT_VALUES, ierr)
-
-      ! restore the row
-      call MatRestoreRow(ctx%loss%M, irow, ncols, cols, vals, ierr)
-
-      ! insert source value
-      call MatSetValue(jac_prec, irow, n, sptr(irow-row_start+1), &
-           INSERT_VALUES, ierr)
-
-    end do
-
-    ! allocate space for flux vector buffer
-    if (rank == n_procs_cmfd - 1) then
-      if (.not. allocated(phi_tmp)) allocate(phi_tmp(0:n-1))
-    end if
-
-    ! get size on each proc
-    if (.not. allocated(dims)) allocate(dims(0:n_procs_cmfd))
-    if (.not. allocated(dims1)) allocate(dims1(0:n_procs_cmfd-1))
-    call VecGetOwnershipRanges(phi, dims, ierr)
-    do k = 0, n_procs_cmfd-1
-      dims1(k) = dims(k+1) - dims(k)
-    end do
-
-    ! gather data on all procs (will truncate xptr if needed for last proc)
-    call MPI_GATHERV(xptr, dims1(rank), MPI_REAL8, phi_tmp, dims1, &
-         dims(0:n_procs_cmfd-1), MPI_REAL8, n_procs_cmfd-1, &
-         PETSC_COMM_WORLD, ierr)
-
-    ! set values in last row of matrix
-    if (rank == n_procs_cmfd - 1) then
-      phi_tmp = -phi_tmp ! negate the transpose
-      call MatSetValues(jac_prec, 1, (/n/), n, (/(k,k=0,n-1)/), phi_tmp, &
-           INSERT_VALUES, ierr)
-      call MatSetValue(jac_prec, n, n, ONE, INSERT_VALUES, ierr)
-    end if
-
-    ! assemble matrix
-    call MatAssemblyBegin(jac_prec, MAT_FINAL_ASSEMBLY, ierr)
-    call MatAssemblyEnd(jac_prec, MAT_FINAL_ASSEMBLY, ierr)
-    call MatAssemblyBegin(jac, MAT_FINAL_ASSEMBLY, ierr)
-    call MatAssemblyEnd(jac, MAT_FINAL_ASSEMBLY, ierr)
-
-    ! reset all vectors
-    call VecResetArray(phi,ierr)
-
-    ! restore all vectors
-    call VecRestoreArrayF90(x, xptr, ierr)
-    call VecRestoreArrayF90(source, sptr, ierr)
-
-    ! destroy all temporary objects
-    call VecDestroy(phi, ierr)
-    call VecDestroy(source, ierr)
-
-    ! deallocate all temporary space
-    if (allocated(cols)) deallocate(cols)
-    if (allocated(vals)) deallocate(vals)
-    if (allocated(phi_tmp)) deallocate(phi_tmp)
-    if (allocated(dims)) deallocate(dims)
-    if (allocated(dims1)) deallocate(dims1)
-
-    ! print jacobian out
-    if (cmfd_write_matrices) call print_J_operator(jac_prec)
-
-  end subroutine build_jacobian_matrix 
-
-!===============================================================================
-! PRINT_J_OPERATOR
-!===============================================================================
-
-  subroutine print_J_operator(jac)
-
-    Mat :: jac
- 
-    PetscViewer :: viewer
-
-    ! write out matrix in binary file (debugging)
-    call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'jacobian.bin', &
-         FILE_MODE_WRITE, viewer, ierr)
-    call MatView(jac, viewer, ierr)
-    call PetscViewerDestroy(viewer, ierr)
-
-  end subroutine print_J_operator
-
-!==============================================================================
-! DESTROY_J_OPERATOR
-!==============================================================================
-
-  subroutine destroy_J_operator(this)
-
-    type(jacobian_operator) :: this
-
-    ! deallocate matrix
-    call MatDestroy(this%J,ierr)
-
-    ! deallocate other parameters
-    if (allocated(this%d_nnz)) deallocate(this%d_nnz)
-    if (allocated(this%o_nnz)) deallocate(this%o_nnz)
-
-  end subroutine destroy_J_operator
-
-# endif
-
-end module cmfd_jacobian_operator
diff --git a/src/cmfd_jfnk_solver.F90 b/src/cmfd_jfnk_solver.F90
new file mode 100644
index 0000000000..7bd21a8758
--- /dev/null
+++ b/src/cmfd_jfnk_solver.F90
@@ -0,0 +1,398 @@
+module cmfd_jfnk_solver
+
+  use cmfd_loss_operator,  only: init_loss_matrix, build_loss_matrix
+  use cmfd_power_solver,   only: cmfd_power_execute
+  use cmfd_prod_operator,  only: init_prod_matrix, build_prod_matrix
+  use matrix_header,       only: Matrix
+  use solver_interface,    only: JFNKSolver, Jfnk_ctx
+  use vector_header,       only: Vector
+
+  implicit none
+  private
+  public :: cmfd_jfnk_execute
+
+  logical          :: adjoint_calc ! true if an adjoint is to be calculated
+  type(Jfnk_ctx)   :: jfnk_data    ! object that holds pointers to user routines
+  type(Matrix)     :: jac_prec     ! Jacobian preconditioner object
+  type(Matrix)     :: loss         ! CMFD loss matrix
+  type(Matrix)     :: prod         ! CMFD production matrix
+  type(JFNKSolver) :: jfnk         ! JFNK solver object
+  type(Vector)     :: resvec       ! JFNK residual vector
+  type(Vector)     :: xvec         ! JFNK solution vector
+
+contains
+
+!===============================================================================
+! CMFD_JFNK_EXECUTE main routine for JFNK solver
+!===============================================================================
+
+  subroutine cmfd_jfnk_execute(adjoint)
+
+    use global,  only: time_cmfdbuild, time_cmfdsolve
+
+    logical, intent(in), optional :: adjoint ! adjoint calculation
+
+    ! Check for adjoint
+    adjoint_calc = .false.
+    if (present(adjoint)) adjoint_calc = adjoint
+
+    ! Seed with power iteration to help converge to fundamental mode
+    call cmfd_power_execute(k_tol=1.E-3_8, s_tol=1.E-3_8, adjoint=adjoint_calc)
+
+    ! Start timer for build
+    call time_cmfdbuild % start()
+
+    ! Initialize data
+    call init_data()
+
+    ! Initialize solver
+#ifdef PETSC
+    call jfnk % create()
+#endif
+
+    ! Set up residual and jacobian routines
+    jfnk_data % res_proc_ptr => compute_nonlinear_residual
+    jfnk_data % jac_proc_ptr => build_jacobian_matrix
+#ifdef PETSC
+    call jfnk % set_functions(jfnk_data, resvec, jac_prec)
+#endif
+
+    ! Stop timer for build
+    call time_cmfdbuild % stop()
+
+    ! Solve the system
+    call time_cmfdsolve % start()
+#ifdef PETSC
+    call jfnk % solve(xvec)
+#endif
+    call time_cmfdsolve % stop()
+
+    ! Extracts results to cmfd object
+    call extract_results()
+
+    ! Deallocate all slepc data
+    call finalize()
+
+  end subroutine cmfd_jfnk_execute
+
+!===============================================================================
+! INIT_DATA allocates matrices vectors for CMFD solution
+!===============================================================================
+
+  subroutine init_data()
+
+    use constants, only: ZERO, ONE
+    use global,    only: cmfd, cmfd_adjoint_type, current_batch
+
+    logical :: physical_adjoint ! physical adjoing calculation logical
+    integer :: n ! size of matrices
+
+    ! Set up all matrices
+    call init_loss_matrix(loss)
+    call init_prod_matrix(prod)
+    call init_jacobian_matrix()
+
+    ! Check for physical adjoint
+    physical_adjoint = .false.
+    if (adjoint_calc .and. trim(cmfd_adjoint_type) == 'physical') &
+         physical_adjoint = .true.
+
+    ! Create matrix operators
+    call build_loss_matrix(loss, adjoint = physical_adjoint)
+    call build_prod_matrix(prod, adjoint = physical_adjoint)
+
+    ! Assemble matrices and use PETSc
+    call loss % assemble()
+    call prod % assemble()
+    call loss % setup_petsc()
+    call prod % setup_petsc()
+
+    ! Check for mathematical adjoint
+    if (adjoint_calc .and. trim(cmfd_adjoint_type) == 'math') &
+         call compute_adjoint()
+
+    ! Get problem size
+    n = loss % n
+
+    ! Create problem vectors
+    call resvec % create(n + 1)
+    call xvec % create(n + 1)
+
+    ! Set flux in guess from rough power iteration
+    if (adjoint_calc) then
+      xvec % val(1:n) = cmfd % adj_phi
+    else
+      xvec % val(1:n) = cmfd % phi
+    end if
+
+    ! Set keff in guess from rough power iteration
+    if (adjoint_calc) then
+      xvec % val(n + 1) = ONE/cmfd % adj_keff
+    else
+      xvec % val(n + 1) = ONE/cmfd % keff
+    end if
+
+    ! Set up vectors for PETSc 
+    call resvec   % setup_petsc()
+    call xvec     % setup_petsc()
+
+    ! Build jacobian from initial guess
+    call build_jacobian_matrix(xvec)
+
+    ! Set up Jacobian for PETSc
+    call jac_prec % setup_petsc()
+
+    ! Set dominance ratio to 0
+    cmfd % dom(current_batch) = ZERO
+
+  end subroutine init_data
+
+!==============================================================================
+! INIT_JACOBIAN_MATRIX preallocates jacobian matrix and initializes it
+!==============================================================================
+
+  subroutine init_jacobian_matrix()
+
+    integer :: nnz ! number of nonzeros in matrix
+    integer :: n   ! dimension of matrix
+
+    ! Get length of matrix and number of nonzeros total in loss matrix
+    nnz = loss % nnz
+    n   = loss % n
+
+    ! There is one more nonzero for each row and and additional row of nonzeros
+    nnz = nnz + 2*n + 1
+
+    ! We need 1 more row for the jacobian
+    n = n + 1
+
+    ! Configure Jacobian matrix
+    call jac_prec % create(n, nnz)
+
+  end subroutine init_jacobian_matrix
+
+!===============================================================================
+! BUILD_JACOBIAN_MATRIX creates the Jacobian of nonlinear eigenvalue problem
+!===============================================================================
+
+  subroutine build_jacobian_matrix(x)
+
+    use constants,  only: ONE
+
+    type(Vector), intent(in) :: x ! solution vector
+
+    integer      :: i      ! loop counter for jacobian rows
+    integer      :: jjac   ! loop counter for jacobian cols
+    integer      :: jloss  ! loop counter for loss matrix cols
+    integer      :: jprod  ! loop counter for prod matrix cols
+    integer      :: n      ! problem size
+    real(8)      :: lambda ! eigenvalue
+    real(8)      :: val    ! temporary real scalar
+    type(Vector) :: flux   ! flux vector
+    type(Vector) :: fsrc   ! fission source vector
+
+    ! Get the problem size
+    n = loss % n
+
+    ! Get flux and eigenvalue
+    flux % val => x % val(1:n)
+    lambda = x % val(n + 1)
+
+    ! Create fission source vector
+    call fsrc % create(n)
+    call prod % vector_multiply(flux, fsrc)
+
+    ! Reset counters in Jacobian matrix
+    jac_prec % n_count  = 1
+    jac_prec % nz_count = 1
+
+    ! Begin loop around rows of Jacobian matrix
+    ROWS: do i = 1, n
+
+      ! Add a new row in Jacobian
+      call jac_prec % new_row()
+
+      ! Begin loop around columns of loss matrix
+      COLS_LOSS: do jloss = loss % get_row(i), loss % get_row(i + 1) - 1
+
+        ! Start with the value in the loss matrix
+        val = loss % val(jloss)
+
+        ! Loop around columns of prod matrix
+        COLS_PROD: do jprod = prod % get_row(i), prod % get_row(i + 1) - 1
+
+          ! See if columns agree with loss matrix
+          if (prod % get_col(jprod) == loss % get_col(jloss)) then
+            val = val - lambda*prod % val(jprod)
+            exit
+          end if
+
+        end do COLS_PROD 
+
+        ! Record value in jacobian
+        call jac_prec % add_value(loss % get_col(jloss), val)
+
+      end do COLS_LOSS
+
+      ! Add fission source value
+      val = -fsrc % val(i)
+      call jac_prec % add_value(n + 1, val)
+
+    end do ROWS
+
+    ! Need to add negative transpose of flux vector on last row
+    call jac_prec % new_row()
+    do jjac = 1, n
+      val = -flux % val(jjac)
+      call jac_prec % add_value(jjac, val)
+    end do
+
+    ! Add unity on bottom right corner of matrix
+    call jac_prec % add_value(n + 1, ONE)
+
+    ! CRS requires a final value in row
+    call jac_prec % new_row()
+
+    ! Free all allocated memory
+    flux % val => null()
+    call fsrc % destroy()
+
+  end subroutine build_jacobian_matrix
+
+!===============================================================================
+! COMPUTE_NONLINEAR_RESIDUAL computes the residual of the nonlinear equations
+!===============================================================================
+
+  subroutine compute_nonlinear_residual(x, res)
+
+    use global,       only: cmfd_write_matrices
+
+    type(Vector), intent(in)    :: x   ! solution vector
+    type(Vector), intent(inout) :: res ! residual vector
+
+    character(len=25) :: filename
+    integer           :: n
+    real(8)           :: lambda
+    type(Vector)      :: res_loss
+    type(Vector)      :: res_prod
+    type(Vector)      :: flux
+
+    ! Get problem size
+    n = loss % n
+
+    ! Set up temporary vectors
+    call res_loss % create(n)
+    call res_prod % create(n)
+
+    ! Extract flux
+    flux % n = n
+    flux % val => x % val(1:n)
+
+    ! Extract eigenvalue
+    lambda = x % val(n + 1)
+
+    ! Calculate M*flux then F*flux
+    call loss % vector_multiply(flux, res_loss)
+    call prod % vector_multiply(flux, res_prod)
+
+    ! Put flux component values in residual vector
+    res % val(1:n) = res_loss % val - lambda*res_prod % val
+
+    ! Put eigenvalue component in residual vector
+    res % val(n+1) = 0.5_8 - 0.5_8*sum(flux % val * flux % val)
+
+    ! Write out data in binary files (debugging)
+    if (cmfd_write_matrices) then
+
+      ! Write out residual vector 
+      if (adjoint_calc) then
+        filename = 'adj_res.bin'
+      else
+        filename = 'res.bin'
+      end if
+      call res % write_petsc_binary(filename)
+
+      ! Write out solution vector
+      if (adjoint_calc) then
+        filename = 'adj_x.bin'
+      else
+        filename = 'x.bin'
+      end if
+      call x % write_petsc_binary(filename)
+
+    end if
+
+  end subroutine compute_nonlinear_residual
+
+!===============================================================================
+! COMPUTE_ADJOINT calculates mathematical adjoint by taking transposes
+!===============================================================================
+
+  subroutine compute_adjoint()
+
+    use global,  only: cmfd_write_matrices
+
+    ! Transpose matrices
+    call loss % transpose()
+    call prod % transpose()
+
+    ! Write out matrix in binary file (debugging)
+    if (cmfd_write_matrices) then
+      call loss % write_petsc_binary('adj_lossmat.bin')
+      call loss % write_petsc_binary('adj_prodmat.bin')
+    end if
+
+  end subroutine compute_adjoint
+
+!===============================================================================
+! EXTRACT_RESULTS gets the results and puts them in global CMFD object
+!===============================================================================
+
+  subroutine extract_results()
+
+    use constants, only: ZERO, ONE
+    use global,    only: cmfd
+
+    integer :: n ! problem size
+
+    ! Get problem size
+    n = loss % n
+
+    ! Also allocate in cmfd object
+    if (adjoint_calc) then
+      if (.not. allocated(cmfd % adj_phi)) allocate(cmfd % adj_phi(n))
+    else
+      if (.not. allocated(cmfd % phi)) allocate(cmfd % phi(n))
+    end if
+
+    ! Get flux and eigenvalue
+    if (adjoint_calc) then
+      cmfd % adj_phi  = xvec % val(1:n)
+      cmfd % adj_keff = ONE / xvec % val(n+1)
+    else
+      cmfd % phi  = xvec % val(1:n)
+      cmfd % keff = ONE / xvec % val(n+1)
+    end if
+
+  end subroutine extract_results
+
+!===============================================================================
+! FINALIZE frees all memory from a JFNK calculation
+!===============================================================================
+
+  subroutine finalize()
+
+    call loss % destroy()
+    call prod % destroy()
+    call jac_prec % destroy()
+    call xvec % destroy()
+    call resvec % destroy()
+#ifdef PETSC
+    call jfnk % destroy()
+#endif
+    nullify(jfnk_data % res_proc_ptr)
+    nullify(jfnk_data % jac_proc_ptr)
+
+  end subroutine finalize
+
+end module cmfd_jfnk_solver
diff --git a/src/cmfd_loss_operator.F90 b/src/cmfd_loss_operator.F90
index 4b5fb1196d..6e9917ac5b 100644
--- a/src/cmfd_loss_operator.F90
+++ b/src/cmfd_loss_operator.F90
@@ -1,103 +1,86 @@
 module cmfd_loss_operator
 
-# ifdef PETSC
+  use constants,      only: CMFD_NOACCEL, ZERO
+  use global,         only: cmfd, cmfd_coremap
+  use matrix_header,  only: Matrix
 
   implicit none
   private
-  public :: init_M_operator, build_loss_matrix, destroy_M_operator
-
-# include <finclude/petsc.h90>
-
-  integer  :: nx   ! maximum number of x cells
-  integer  :: ny   ! maximum number of y cells
-  integer  :: nz   ! maximum number of z cells
-  integer  :: ng   ! maximum number of groups
-  integer  :: ierr ! petsc error code
-
-  type, public :: loss_operator
-    Mat      :: M      ! petsc matrix for neutronic loss operator
-    integer  :: n      ! dimensions of matrix
-    integer  :: nnz    ! max number of nonzeros
-    integer  :: localn ! local size on proc
-    integer, allocatable :: d_nnz(:) ! vector of diagonal preallocation
-    integer, allocatable :: o_nnz(:) ! vector of off-diagonal preallocation
-  end type loss_operator
-
-  logical :: adjoint_calc = .false. ! adjoint calculation
+  public :: init_loss_matrix, build_loss_matrix
 
 contains
 
 !===============================================================================
-! INIT_M_OPERATOR
+! INIT_LOSS_MATRIX preallocates loss matrix and initializes it
 !===============================================================================
 
-  subroutine init_M_operator(this)
+  subroutine init_loss_matrix(loss_matrix)
 
-    type(loss_operator) :: this
+    type(Matrix), intent(inout) :: loss_matrix ! cmfd loss matrix
 
-    ! get indices
-    call get_M_indices(this)
+    integer :: nx   ! maximum number of x cells
+    integer :: ny   ! maximum number of y cells
+    integer :: nz   ! maximum number of z cells
+    integer :: ng   ! maximum number of groups
+    integer :: n    ! total length of matrix
+    integer :: nnz  ! number of nonzeros in matrix
+    integer :: n_i  ! number of interior cells
+    integer :: n_c  ! number of corner cells
+    integer :: n_s  ! number side cells
+    integer :: nz_c ! number of non-zero corner cells
+    integer :: nz_s ! number of non-zero side cells
+    integer :: nz_i ! number of non-zero interior cells
 
-    ! get preallocation
-    call preallocate_loss_matrix(this)
-
-    ! set up M operator
-    call MatCreateAIJ(PETSC_COMM_WORLD, this%localn, this%localn, PETSC_DECIDE,&
-         PETSC_DECIDE, PETSC_NULL_INTEGER, this%d_nnz, PETSC_NULL_INTEGER, &
-         this%o_nnz, this%M,ierr)
-    call MatSetOption(this%M, MAT_NEW_NONZERO_LOCATIONS, PETSC_TRUE, ierr)
-    call MatSetOption(this%M, MAT_IGNORE_ZERO_ENTRIES, PETSC_TRUE, ierr)
-
-  end subroutine init_M_operator
-
-!===============================================================================
-! GET_M_INDICES
-!===============================================================================
-
-  subroutine get_M_indices(this)
-
-    use global,  only: cmfd, cmfd_coremap
-
-    type(loss_operator) :: this
-
-    ! get maximum number of cells in each direction
+    ! Get maximum number of cells in each direction
     nx = cmfd%indices(1)
     ny = cmfd%indices(2)
     nz = cmfd%indices(3)
     ng = cmfd%indices(4)
 
-    ! get number of nonzeros
-    this%nnz = 7 + ng - 1
-
-    ! calculate dimensions of matrix
+    ! Calculate dimensions of matrix
     if (cmfd_coremap) then
-      this%n = cmfd % mat_dim * ng
+      n = cmfd % mat_dim * ng
     else
-      this%n = nx*ny*nz*ng
+      n = nx*ny*nz*ng
     end if
 
-  end subroutine get_M_indices
+    ! Calculate number of nonzeros, if core map -> need to determine manually
+    if (cmfd_coremap) then
+      nnz = preallocate_loss_matrix(nx, ny, nz, ng, n)
+    else ! structured Cartesian grid
+      n_c  = 4                   ! define # of corners
+      n_s  = 2*(nx + ny) - 8     ! define # of sides
+      n_i  = nx*ny - (n_c + n_S) ! define # of interiors  
+      nz_c = ng*n_c*(3 + ng - 1) ! define # nonzero corners
+      nz_s = ng*n_s*(4 + ng - 1) ! define # nonzero sides
+      nz_i = ng*n_i*(5 + ng - 1) ! define # nonzero interiors
+      nnz  = nz_c + nz_s + nz_i
+    end if
+
+    ! Configure loss matrix
+    call loss_matrix % create(n, nnz)
+
+  end subroutine init_loss_matrix
 
 !===============================================================================
-! PREALLOCATE_LOSS_MATRIX
+! PREALLOCATE_LOSS_MATRIX manually preallocates the loss matrix
 !===============================================================================
 
-  subroutine preallocate_loss_matrix(this)
+  function preallocate_loss_matrix(nx, ny, nz, ng, n) result(nnz)
 
-    use constants,  only: CMFD_NOACCEL
-    use global,     only: cmfd, cmfd_coremap
+    integer, intent(in) :: nx  ! maximum number of x cells
+    integer, intent(in) :: ny  ! maximum number of y cells
+    integer, intent(in) :: nz  ! maximum number of z cells
+    integer, intent(in) :: ng  ! maximum number of groups
+    integer, intent(in) :: n   ! total length of matrix
+    integer             :: nnz ! number of nonzeros
 
-    type(loss_operator) :: this
-
-    integer :: rank          ! rank of processor
-    integer :: sizen         ! number of procs
     integer :: i             ! iteration counter for x
     integer :: j             ! iteration counter for y
     integer :: k             ! iteration counter for z
     integer :: g             ! iteration counter for groups
     integer :: l             ! iteration counter for leakages
     integer :: h             ! energy group when doing scattering
-    integer :: n             ! the extent of the matrix
     integer :: irow          ! row counter
     integer :: bound(6)      ! vector for comparing when looking for bound
     integer :: xyz_idx       ! index for determining if x,y or z leakage
@@ -105,113 +88,68 @@ contains
     integer :: neig_idx(3)   ! spatial indices of neighbour
     integer :: nxyz(3,2)     ! single vector containing bound. locations
     integer :: shift_idx     ! parameter to shift index by +1 or -1
-    integer :: row_start     ! index of local starting row
-    integer :: row_end       ! index of local final row
     integer :: neig_mat_idx  ! matrix index of neighbor cell
     integer :: scatt_mat_idx ! matrix index for h-->g scattering terms
 
-    ! initialize size and rank
-    sizen = 0
-    rank = 0
+    ! Reset number of nonzeros to 0
+    nnz = 0
 
-    ! get rank and max rank of procs
-    call MPI_COMM_RANK(PETSC_COMM_WORLD, rank, ierr)
-    call MPI_COMM_SIZE(PETSC_COMM_WORLD, sizen, ierr)
-
-    ! get local problem size
-    n = this%n
-
-    ! determine local size, divide evenly between all other procs
-    this%localn = n/(sizen)
-        
-    ! add 1 more if less proc id is less than mod
-    if (rank < mod(n,sizen)) this%localn = this%localn + 1
-
-    ! determine local starting row
-    row_start = 0
-    if (rank < mod(n,sizen)) then
-      row_start = rank*(n/sizen+1)
-    else
-      row_start = min(mod(n,sizen)*(n/sizen+1) + &
-           (rank - mod(n,sizen))*(n/sizen),n) 
-    end if
-
-    ! determine local final row
-    row_end = row_start + this%localn - 1
-
-    ! allocate counters
-    if (.not. allocated(this%d_nnz)) allocate(this%d_nnz(row_start:row_end))
-    if (.not. allocated(this%o_nnz)) allocate(this%o_nnz(row_start:row_end))
-    this % d_nnz = 0
-    this % o_nnz = 0
-
-    ! create single vector of these indices for boundary calculation
+    ! Create single vector of these indices for boundary calculation
     nxyz(1,:) = (/1,nx/)
     nxyz(2,:) = (/1,ny/)
     nxyz(3,:) = (/1,nz/)
 
-    ! begin loop around local rows
-    ROWS: do irow = row_start,row_end
+    ! Begin loop around local rows
+    ROWS: do irow = 1, n
 
-      ! initialize counters 
-      this%d_nnz(irow) = 1 ! already add in matrix diagonal
-      this%o_nnz(irow) = 0
+      ! Set a nonzero for diagonal
+      nnz = nnz + 1
 
-      ! get location indices
-      call matrix_to_indices(irow, g, i, j, k)
+      ! Get location indices
+      call matrix_to_indices(irow, g, i, j, k, ng, nx, ny, nz)
 
-      ! create boundary vector
+      ! Create boundary vector
       bound = (/i,i,j,j,k,k/)
 
-      ! begin loop over leakages
+      ! Begin loop over leakages
       LEAK: do l = 1,6
 
-        ! define (x,y,z) and (-,+) indices
+        ! Define (x,y,z) and (-,+) indices
         xyz_idx = int(ceiling(real(l)/real(2)))  ! x=1, y=2, z=3
         dir_idx = 2 - mod(l,2) ! -=1, +=2
 
-        ! calculate spatial indices of neighbor
+        ! Calculate spatial indices of neighbor
         neig_idx = (/i,j,k/)                ! begin with i,j,k
         shift_idx = -2*mod(l,2) +1          ! shift neig by -1 or +1
         neig_idx(xyz_idx) = shift_idx + neig_idx(xyz_idx)
 
-        ! check for global boundary
+        ! Check for global boundary
         if (bound(l) /= nxyz(xyz_idx,dir_idx)) then
 
-          ! check for coremap 
+          ! Check for coremap 
           if (cmfd_coremap) then
 
-            ! check for neighbor that is non-acceleartred
+            ! Check for neighbor that is non-acceleartred
             if (cmfd % coremap(neig_idx(1),neig_idx(2),neig_idx(3)) /= &
                  CMFD_NOACCEL) then
 
-              ! get neighbor matrix index
+              ! Get neighbor matrix index
               call indices_to_matrix(g,neig_idx(1), neig_idx(2), & 
-                   neig_idx(3), neig_mat_idx)
+                   neig_idx(3), neig_mat_idx, ng, nx, ny)
 
-              ! record nonzero
-              if (((neig_mat_idx-1) >= row_start) .and. &
-                   ((neig_mat_idx-1) <= row_end)) then
-                this%d_nnz(irow) = this%d_nnz(irow) + 1
-              else
-                this%o_nnz(irow) = this%o_nnz(irow) + 1
-              end if
+              ! Record nonzero
+              nnz = nnz + 1 
 
             end if
 
           else
 
-            ! get neighbor matrix index
+            ! Get neighbor matrix index
             call indices_to_matrix(g, neig_idx(1), neig_idx(2), neig_idx(3), &
-                 neig_mat_idx)
+                 neig_mat_idx, ng, nx, ny)
 
-            ! record nonzero
-            if (((neig_mat_idx-1) >= row_start) .and. &
-                 ((neig_mat_idx-1) <= row_end)) then
-              this%d_nnz(irow) = this%d_nnz(irow) + 1
-            else
-              this%o_nnz(irow) = this%o_nnz(irow) + 1
-            end if
+            ! Record nonzero
+            nnz = nnz + 1
 
           end if
 
@@ -219,214 +157,206 @@ contains
 
       end do LEAK
 
-      ! begin loop over off diagonal in-scattering
+      ! Begin loop over off diagonal in-scattering
       SCATTR: do h = 1, ng
 
-        ! cycle though if h=g
+        ! Cycle though if h=g, it was already banked in removal xs
         if (h == g) cycle
 
-        ! get neighbor matrix index
-        call indices_to_matrix(h, i, j, k, scatt_mat_idx)
+        ! Get neighbor matrix index
+        call indices_to_matrix(h, i, j, k, scatt_mat_idx, ng, nx, ny)
 
-        ! record nonzero
-        if (((scatt_mat_idx-1) >= row_start) .and. &
-             ((scatt_mat_idx-1) <= row_end)) then
-          this%d_nnz(irow) = this%d_nnz(irow) + 1
-        else
-          this%o_nnz(irow) = this%o_nnz(irow) + 1
-        end if
+        ! Record nonzero
+        nnz = nnz + 1
 
       end do SCATTR
 
     end do ROWS
 
-  end subroutine preallocate_loss_matrix
+  end function preallocate_loss_matrix
 
 !===============================================================================
 ! BUILD_LOSS_MATRIX creates the matrix representing loss of neutrons
 !===============================================================================
 
-  subroutine build_loss_matrix(this, adjoint)
+  subroutine build_loss_matrix(loss_matrix, adjoint)
 
-    use constants,  only: CMFD_NOACCEL, ZERO
-    use global,     only: cmfd, cmfd_coremap, cmfd_write_matrices
+    type(Matrix), intent(inout)   :: loss_matrix ! cmfd loss matrix
+    logical, intent(in), optional :: adjoint     ! set up the adjoint
 
-    type(loss_operator) :: this
-    logical, optional :: adjoint    ! set up the adjoint
+    integer :: nxyz(3,2)     ! single vector containing bound. locations
+    integer :: i             ! iteration counter for x
+    integer :: j             ! iteration counter for y
+    integer :: k             ! iteration counter for z
+    integer :: g             ! iteration counter for groups
+    integer :: l             ! iteration counter for leakages
+    integer :: h             ! energy group when doing scattering
+    integer :: nx            ! maximum number of x cells
+    integer :: ny            ! maximum number of y cells
+    integer :: nz            ! maximum number of z cells
+    integer :: ng            ! maximum number of groups
+    integer :: neig_mat_idx  ! matrix index of neighbor cell
+    integer :: scatt_mat_idx ! matrix index for h-->g scattering terms
+    integer :: bound(6)      ! vector for comparing when looking for bound
+    integer :: xyz_idx       ! index for determining if x,y or z leakage
+    integer :: dir_idx       ! index for determining - or + face of cell
+    integer :: neig_idx(3)   ! spatial indices of neighbour
+    integer :: shift_idx     ! parameter to shift index by +1 or -1
+    integer :: irow          ! iteration counter over row
+    logical :: adjoint_calc  ! is this a physical adjoint calculation?
+    real(8) :: totxs         ! total macro cross section
+    real(8) :: scattxsgg     ! scattering macro cross section g-->g
+    real(8) :: scattxshg     ! scattering macro cross section h-->g
+    real(8) :: dtilde(6)     ! finite difference coupling parameter
+    real(8) :: dhat(6)       ! nonlinear coupling parameter
+    real(8) :: hxyz(3)       ! cell lengths in each direction
+    real(8) :: jn            ! direction dependent leakage coeff to neig
+    real(8) :: jo(6)         ! leakage coeff in front of cell flux
+    real(8) :: jnet          ! net leakage from jo
+    real(8) :: val           ! temporary variable before saving to 
 
-    integer :: nxyz(3,2)            ! single vector containing bound. locations
-    integer :: i                    ! iteration counter for x
-    integer :: j                    ! iteration counter for y
-    integer :: k                    ! iteration counter for z
-    integer :: g                    ! iteration counter for groups
-    integer :: l                    ! iteration counter for leakages
-    integer :: h                    ! energy group when doing scattering
-    integer :: neig_mat_idx         ! matrix index of neighbor cell
-    integer :: scatt_mat_idx        ! matrix index for h-->g scattering terms
-    integer :: bound(6)             ! vector for comparing when looking for bound
-    integer :: xyz_idx              ! index for determining if x,y or z leakage
-    integer :: dir_idx              ! index for determining - or + face of cell
-    integer :: neig_idx(3)          ! spatial indices of neighbour
-    integer :: shift_idx            ! parameter to shift index by +1 or -1
-    integer :: row_start            ! the first local row on the processor
-    integer :: row_finish           ! the last local row on the processor
-    integer :: irow                 ! iteration counter over row
-    real(8) :: totxs                ! total macro cross section
-    real(8) :: scattxsgg            ! scattering macro cross section g-->g
-    real(8) :: scattxshg            ! scattering macro cross section h-->g
-    real(8) :: dtilde(6)            ! finite difference coupling parameter
-    real(8) :: dhat(6)              ! nonlinear coupling parameter
-    real(8) :: hxyz(3)              ! cell lengths in each direction
-    real(8) :: jn                   ! direction dependent leakage coeff to neig
-    real(8) :: jo(6)                ! leakage coeff in front of cell flux
-    real(8) :: jnet                 ! net leakage from jo
-    real(8) :: val                  ! temporary variable before saving to 
-
-    ! check for adjoint
+    ! Check for adjoint
+    adjoint_calc = .false.
     if (present(adjoint)) adjoint_calc = adjoint 
 
-    ! create single vector of these indices for boundary calculation
+    ! Get maximum number of cells in each direction
+    nx = cmfd%indices(1)
+    ny = cmfd%indices(2)
+    nz = cmfd%indices(3)
+    ng = cmfd%indices(4)
+
+    ! Create single vector of these indices for boundary calculation
     nxyz(1,:) = (/1,nx/)
     nxyz(2,:) = (/1,ny/)
     nxyz(3,:) = (/1,nz/)
 
-    ! initialize row start and finish
-    row_start = 0
-    row_finish = 0
+    ! Begin iteration loops
+    ROWS: do irow = 1, loss_matrix % n
 
-    ! get row bounds for this processor
-    call MatGetOwnershipRange(this%M, row_start, row_finish, ierr)
+      ! Set up a new row in matrix
+      call loss_matrix % new_row()
 
-    ! begin iteration loops
-    ROWS: do irow = row_start, row_finish-1
+      ! Get indices for that row
+      call matrix_to_indices(irow, g, i, j, k, ng, nx, ny, nz)
 
-      ! get indices for that row
-      call matrix_to_indices(irow, g, i, j, k)
-
-      ! retrieve cell data
+      ! Retrieve cell data
       totxs = cmfd%totalxs(g,i,j,k)
       scattxsgg = cmfd%scattxs(g,g,i,j,k)
       dtilde = cmfd%dtilde(:,g,i,j,k)
       hxyz = cmfd%hxyz(:,i,j,k)
 
-      ! check and get dhat
+      ! Check and get dhat
       if (allocated(cmfd%dhat)) then
         dhat = cmfd%dhat(:,g,i,j,k)
       else
         dhat = ZERO
       end if
 
-      ! create boundary vector 
+      ! Create boundary vector 
       bound = (/i,i,j,j,k,k/)
 
-      ! begin loop over leakages
+      ! Begin loop over leakages
       ! 1=-x, 2=+x, 3=-y, 4=+y, 5=-z, 6=+z 
       LEAK: do l = 1,6
 
-        ! define (x,y,z) and (-,+) indices
+        ! Define (x,y,z) and (-,+) indices
         xyz_idx = int(ceiling(real(l)/real(2)))  ! x=1, y=2, z=3
         dir_idx = 2 - mod(l,2) ! -=1, +=2
 
-        ! calculate spatial indices of neighbor
+        ! Calculate spatial indices of neighbor
         neig_idx = (/i,j,k/)                ! begin with i,j,k
         shift_idx = -2*mod(l,2) +1          ! shift neig by -1 or +1
         neig_idx(xyz_idx) = shift_idx + neig_idx(xyz_idx)
 
-        ! check for global boundary
+        ! Check for global boundary
         if (bound(l) /= nxyz(xyz_idx,dir_idx)) then
 
-          ! check for core map
+          ! Check for core map
           if (cmfd_coremap) then
 
-            ! check that neighbor is not reflector
+            ! Check that neighbor is not reflector
             if (cmfd % coremap(neig_idx(1),neig_idx(2),neig_idx(3)) /= &
                  CMFD_NOACCEL) then
 
-              ! compute leakage coefficient for neighbor
+              ! Compute leakage coefficient for neighbor
               jn = -dtilde(l) + shift_idx*dhat(l)
 
-              ! get neighbor matrix index
+              ! Get neighbor matrix index
               call indices_to_matrix(g, neig_idx(1), neig_idx(2), neig_idx(3), &
-                   neig_mat_idx)
+                   neig_mat_idx, ng, nx, ny)
 
-              ! compute value and record to bank
+              ! Compute value and record to bank
               val = jn/hxyz(xyz_idx)
 
-              ! record value in matrix
-              call MatSetValue(this%M, irow, neig_mat_idx-1, val, &
-                   INSERT_VALUES, ierr)
+              ! Record value in matrix
+              call loss_matrix % add_value(neig_mat_idx, val)
 
             end if
 
           else
 
-            ! compute leakage coefficient for neighbor
+            ! Compute leakage coefficient for neighbor
             jn = -dtilde(l) + shift_idx*dhat(l)
 
-            ! get neighbor matrix index
+            ! Get neighbor matrix index
             call indices_to_matrix(g, neig_idx(1), neig_idx(2), neig_idx(3), &
-                 neig_mat_idx)
+                 neig_mat_idx, ng, nx, ny)
 
-            ! compute value and record to bank
+            ! Compute value and record to bank
             val = jn/hxyz(xyz_idx)
 
-            ! record value in matrix
-            call MatSetValue(this%M, irow, neig_mat_idx-1, val, &
-                 INSERT_VALUES, ierr)
+            ! Record value in matrix
+            call loss_matrix % add_value(neig_mat_idx, val)
 
           end if
 
         end if
 
-        ! compute leakage coefficient for target
+        ! Compute leakage coefficient for target
         jo(l) = shift_idx*dtilde(l) + dhat(l)
 
       end do LEAK
 
-      ! calate net leakage coefficient for target
+      ! Calculate net leakage coefficient for target
       jnet = (jo(2) - jo(1))/hxyz(1) + (jo(4) - jo(3))/hxyz(2) + &
            (jo(6) - jo(5))/hxyz(3)
 
-      ! calculate loss of neutrons
+      ! Calculate loss of neutrons
       val = jnet + totxs - scattxsgg
 
-      ! record diagonal term
-      call MatSetValue(this%M, irow, irow, val, INSERT_VALUES, ierr)
+      ! Record diagonal term
+      call loss_matrix % add_value(irow, val)
 
-      ! begin loop over off diagonal in-scattering
+      ! Begin loop over off diagonal in-scattering
       SCATTR: do h = 1, ng
 
-        ! cycle though if h=g
+        ! Cycle though if h=g, value already banked in removal xs
         if (h == g) cycle
 
-        ! get neighbor matrix index
-        call indices_to_matrix(h, i, j, k, scatt_mat_idx)
+        ! Get neighbor matrix index
+        call indices_to_matrix(h, i, j, k, scatt_mat_idx, ng, nx, ny)
 
-        ! get scattering macro xs
-        scattxshg = cmfd%scattxs(h, g, i, j, k)
+        ! Check for adjoint
+        if (adjoint_calc) then
+          ! Get scattering macro xs, transposed!
+          scattxshg = cmfd%scattxs(g, h, i, j, k)
+        else
+          ! Get scattering macro xs
+          scattxshg = cmfd%scattxs(h, g, i, j, k)
+        end if
 
-        ! record value in matrix (negate it)
+        ! Negate the scattering xs 
         val = -scattxshg
 
-        ! check for adjoint and bank value
-        if (adjoint_calc) then
-          call MatSetValue(this%M, scatt_mat_idx-1, irow, val, &
-               INSERT_VALUES, ierr)
-        else
-          call MatSetValue(this%M, irow, scatt_mat_idx-1, val, &
-               INSERT_VALUES, ierr)
-        end if
+        ! Record value in matrix
+        call loss_matrix % add_value(scatt_mat_idx, val)
 
       end do SCATTR
 
     end do ROWS 
 
-    ! assemble matrix
-    call MatAssemblyBegin(this%M, MAT_FLUSH_ASSEMBLY, ierr)
-    call MatAssemblyEnd(this%M, MAT_FINAL_ASSEMBLY, ierr)
-
-    ! print out operator to file
-    if (cmfd_write_matrices) call print_M_operator(this)
+    ! CSR requires n+1 row
+    call loss_matrix % new_row()
 
   end subroutine build_loss_matrix
 
@@ -434,25 +364,26 @@ contains
 ! INDICES_TO_MATRIX takes (x,y,z,g) indices and computes location in matrix
 !===============================================================================
 
-  subroutine indices_to_matrix(g, i, j, k, matidx)
+  subroutine indices_to_matrix(g, i, j, k, matidx, ng, nx, ny)
 
-    use global,  only: cmfd, cmfd_coremap
+    integer, intent(out) :: matidx ! the index location in matrix
+    integer, intent(in)  :: i      ! current x index
+    integer, intent(in)  :: j      ! current y index
+    integer, intent(in)  :: k      ! current z index
+    integer, intent(in)  :: g      ! current group index
+    integer, intent(in)  :: nx     ! maximum number of x cells
+    integer, intent(in)  :: ny     ! maximum number of y cells
+    integer, intent(in)  :: ng     ! maximum number of groups
 
-    integer :: matidx ! the index location in matrix
-    integer :: i      ! current x index
-    integer :: j      ! current y index
-    integer :: k      ! current z index
-    integer :: g      ! current group index
-
-    ! check if coremap is used
+    ! Check if coremap is used
     if (cmfd_coremap) then
 
-      ! get idx from core map
+      ! Get idx from core map
       matidx = ng*(cmfd % coremap(i,j,k)) - (ng - g)
 
     else
 
-      ! compute index
+      ! Compute index
       matidx = g + ng*(i - 1) + ng*nx*(j - 1) + ng*nx*ny*(k - 1)
 
     end if
@@ -460,80 +391,40 @@ contains
   end subroutine indices_to_matrix
 
 !===============================================================================
-! MATRIX_TO_INDICES
+! MATRIX_TO_INDICES converts a matrix index to spatial and group indicies
 !===============================================================================
 
-  subroutine matrix_to_indices(irow, g, i, j, k)
+  subroutine matrix_to_indices(irow, g, i, j, k, ng, nx, ny, nz)
 
-    use global,  only: cmfd, cmfd_coremap
+    integer, intent(out) :: i     ! iteration counter for x
+    integer, intent(out) :: j     ! iteration counter for y
+    integer, intent(out) :: k     ! iteration counter for z
+    integer, intent(out) :: g     ! iteration counter for groups
+    integer, intent(in)  :: irow  ! iteration counter over row (0 reference)
+    integer, intent(in)  :: nx    ! maximum number of x cells
+    integer, intent(in)  :: ny    ! maximum number of y cells
+    integer, intent(in)  :: nz    ! maximum number of z cells
+    integer, intent(in)  :: ng    ! maximum number of groups
 
-    integer :: i     ! iteration counter for x
-    integer :: j     ! iteration counter for y
-    integer :: k     ! iteration counter for z
-    integer :: g     ! iteration counter for groups
-    integer :: irow  ! iteration counter over row (0 reference)
-
-    ! check for core map
+    ! Check for core map
     if (cmfd_coremap) then
 
-      ! get indices from indexmap
-      g = mod(irow, ng) + 1
-      i = cmfd % indexmap(irow/ng+1,1)
-      j = cmfd % indexmap(irow/ng+1,2)
-      k = cmfd % indexmap(irow/ng+1,3)
+      ! Get indices from indexmap
+      g = mod(irow-1, ng) + 1
+      i = cmfd % indexmap((irow-1)/ng+1,1)
+      j = cmfd % indexmap((irow-1)/ng+1,2)
+      k = cmfd % indexmap((irow-1)/ng+1,3)
 
     else
 
-      ! compute indices
-      g = mod(irow, ng) + 1
-      i = mod(irow, ng*nx)/ng + 1
-      j = mod(irow, ng*nx*ny)/(ng*nx)+ 1
-      k = mod(irow, ng*nx*ny*nz)/(ng*nx*ny) + 1
+      ! Compute indices
+      g = mod(irow-1, ng) + 1
+      i = mod(irow-1, ng*nx)/ng + 1
+      j = mod(irow-1, ng*nx*ny)/(ng*nx)+ 1
+      k = mod(irow-1, ng*nx*ny*nz)/(ng*nx*ny) + 1
 
     end if
 
   end subroutine matrix_to_indices
 
-!===============================================================================
-! PRINT_M_OPERATOR
-!===============================================================================
-
-  subroutine print_M_operator(this)
-
-    type(loss_operator) :: this
-
-    PetscViewer :: viewer
-
-    ! write out matrix in binary file (debugging)
-    if (adjoint_calc) then
-      call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_lossmat.bin', &
-           FILE_MODE_WRITE, viewer, ierr)
-    else
-      call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'lossmat.bin', &
-           FILE_MODE_WRITE, viewer, ierr)
-    end if
-    call MatView(this%M, viewer, ierr)
-    call PetscViewerDestroy(viewer, ierr)
-
-  end subroutine print_M_operator
-
-!==============================================================================
-! DESTROY_M_OPERATOR
-!==============================================================================
-
-  subroutine destroy_M_operator(this)
-
-    type(loss_operator) :: this
-
-    ! deallocate matrix
-    call MatDestroy(this%M, ierr)
-
-    ! deallocate other parameters
-    if (allocated(this%d_nnz)) deallocate(this%d_nnz)
-    if (allocated(this%o_nnz)) deallocate(this%o_nnz)
-
-  end subroutine destroy_M_operator
-
-# endif
-
 end module cmfd_loss_operator
diff --git a/src/cmfd_message_passing.F90 b/src/cmfd_message_passing.F90
deleted file mode 100644
index f5269bc2a7..0000000000
--- a/src/cmfd_message_passing.F90
+++ /dev/null
@@ -1,98 +0,0 @@
-module cmfd_message_passing
-
-# ifdef PETSC
-
-! This module contains MPI routines related to the CMFD calculation
-
-  implicit none
-  private
-  public :: petsc_init_mpi, cmfd_bcast 
-
-# include <finclude/petsc.h90>
-
-contains
-
-!===============================================================================
-! PETSC_INIT_MPI
-!===============================================================================
-
-  subroutine petsc_init_mpi()
-
-    use global,       only: n_procs_cmfd, time_cmfd, rank, mpi_err
-
-    integer               :: new_comm   ! new communicator
-    integer               :: color      ! communicator color
-
-    ! assign color
-    if (rank < n_procs_cmfd) then
-      color = 1
-    else
-      color = 2
-    end if
-
-    ! split up procs
-    call MPI_COMM_SPLIT(MPI_COMM_WORLD,color,0,new_comm,mpi_err)
-
-    ! assign to PETSc
-    PETSC_COMM_WORLD = new_comm
-
-    ! initialize petsc on all procs
-    call PetscInitialize(PETSC_NULL_CHARACTER,mpi_err)
-
-    ! initialize timer
-    call time_cmfd % reset()
-
-  end subroutine petsc_init_mpi
-
-!===============================================================================
-! CMFD_BCAST
-!===============================================================================
-
-  subroutine cmfd_bcast()
-
-    use global,       only: cmfd_coremap, cmfd, mpi_err
-    use cmfd_header,  only: allocate_cmfd
-
-    integer :: nx  ! number of mesh cells in x direction
-    integer :: ny  ! number of mesh cells in y direction
-    integer :: nz  ! number of mesh cells in z direction
-    integer :: ng  ! number of energy groups
-
-    ! extract spatial and energy indices from object
-    nx = cmfd % indices(1)
-    ny = cmfd % indices(2)
-    nz = cmfd % indices(3)
-    ng = cmfd % indices(4)
-
-    ! initialize data
-    if (.not. allocated(cmfd%flux)) call allocate_cmfd(cmfd)
-
-    ! sync up procs
-    call MPI_Barrier(PETSC_COMM_WORLD, mpi_err)
-
-    ! broadcast all data
-    call MPI_BCAST(cmfd%flux, ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%totalxs, ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%p1scattxs, ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%scattxs, ng*ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%nfissxs, ng*ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%diffcof, ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%dtilde, 6*ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%dhat, 6*ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%hxyz, 3*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-    call MPI_BCAST(cmfd%current, 12*ng*nx*ny*nz, MPI_REAL8, 0, PETSC_COMM_WORLD, mpi_err)
-
-    ! broadcast coremap info
-    if (cmfd_coremap) then
-      call MPI_BCAST(cmfd%coremap, nx*ny*nz, MPI_INTEGER8, 0, PETSC_COMM_WORLD, mpi_err)
-      call MPI_BCAST(cmfd%mat_dim, 1, MPI_INTEGER8, 0, PETSC_COMM_WORLD, mpi_err)
-      if (.not. allocated(cmfd % indexmap)) &
-           allocate(cmfd % indexmap(cmfd % mat_dim,3))
-      call MPI_BCAST(cmfd%indexmap, cmfd%mat_dim*3, MPI_INTEGER8, 0, PETSC_COMM_WORLD, mpi_err)
-    end if
-
-  end subroutine cmfd_bcast
-
-# endif
-
-end module cmfd_message_passing
diff --git a/src/cmfd_output.F90 b/src/cmfd_output.F90
deleted file mode 100644
index ebfec812b0..0000000000
--- a/src/cmfd_output.F90
+++ /dev/null
@@ -1,61 +0,0 @@
-module cmfd_output
-
-#ifdef PETSC
-
-! This modules cleans up cmfd objects and echos the results
-
-  implicit none
-  private
-  public :: finalize_cmfd
-
-contains
-
-!===============================================================================
-! FINALIZE_CMFD
-!===============================================================================
-
-  subroutine finalize_cmfd() 
-
-    use global,      only: cmfd, cmfd_write_balance, cmfd_write_hdf5, &
-                           master, mpi_err
-    use cmfd_header, only: deallocate_cmfd
-
-    ! finalize petsc
-    call PetscFinalize(mpi_err)
-
-    ! write out final neutron balance
-    if (master .and. cmfd_write_balance) call write_neutron_balance()
-
-    ! deallocate cmfd object
-    call deallocate_cmfd(cmfd)
-
-  end subroutine finalize_cmfd 
-
-!===============================================================================
-! WRITE_NEUTRON_BALANCE 
-!===============================================================================
-
-  subroutine write_neutron_balance()
-
-    use cmfd_data,    only: neutron_balance
-    use constants,    only: CMFD_BALANCE, MAX_FILE_LEN
-    use global,       only: path_output
-
-    character(MAX_FILE_LEN) :: filename
-
-    filename = trim(path_output) // 'neutron_balance.out'
-
-    ! open file for output
-    open(UNIT=CMFD_BALANCE, FILE=filename, ACTION='write')
-
-    ! write out the tally
-    call neutron_balance(CMFD_BALANCE) 
-
-    ! close file
-    close(CMFD_BALANCE)
-
-  end subroutine write_neutron_balance
-
-#endif
-
-end module cmfd_output
diff --git a/src/cmfd_power_solver.F90 b/src/cmfd_power_solver.F90
index 8376b964d0..c1309b72cd 100644
--- a/src/cmfd_power_solver.F90
+++ b/src/cmfd_power_solver.F90
@@ -1,445 +1,346 @@
 module cmfd_power_solver
 
-# ifdef PETSC
-
 ! This module contains routines to execute the power iteration solver
 
-  use cmfd_loss_operator, only: loss_operator,init_M_operator, &
-                                build_loss_matrix,destroy_M_operator
-  use cmfd_prod_operator, only: prod_operator,init_F_operator, &
-                                build_prod_matrix,destroy_F_operator
+  use cmfd_loss_operator, only: init_loss_matrix, build_loss_matrix
+  use cmfd_prod_operator, only: init_prod_matrix, build_prod_matrix
+  use matrix_header,      only: Matrix
+  use solver_interface,   only: GMRESSolver
+  use vector_header,      only: Vector
 
   implicit none
   private
   public :: cmfd_power_execute 
 
-# include <finclude/petsc.h90>
-
-  type(loss_operator) :: loss   ! M loss matrix
-  type(prod_operator) :: prod   ! F production matrix
-
-  Vec     :: phi_n                ! new flux eigenvector
-  Vec     :: phi_o                ! old flux eigenvector
-  Vec     :: S_n                  ! new source vector
-  Vec     :: S_o                  ! old source vector
-  KSP     :: krylov               ! krylov solver
-  KSP     :: sub_krylov           ! sub-ksp for bjacobi
-  PC      :: prec                 ! preconditioner for krylov
-  PC      :: sub_prec             ! sub-prec for bjacobi
-  integer :: ierr                 ! error flag
-  integer :: nlocal
-  integer :: first
+  logical :: iconv                ! did the problem converged
   real(8) :: k_n                  ! new k-eigenvalue
   real(8) :: k_o                  ! old k-eigenvalue
   real(8) :: ktol = 1.e-8_8       ! tolerance on keff
   real(8) :: stol = 1.e-8_8       ! tolerance on source
-  logical :: iconv                ! did the problem converged
-
-  logical :: adjoint_calc = .false. ! run an adjoint calculation
+  real(8) :: norm_n               ! current norm of source vector
+  real(8) :: norm_o               ! old norm of source vector
+  real(8) :: kerr                 ! error in keff
+  real(8) :: serr                 ! error in source
+  logical :: adjoint_calc         ! run an adjoint calculation
+  type(Matrix) :: loss            ! cmfd loss matrix
+  type(Matrix) :: prod            ! cmfd prod matrix
+  type(Vector) :: phi_n           ! new flux vector
+  type(Vector) :: phi_o           ! old flux vector
+  type(Vector) :: s_n             ! new source vector
+  type(Vector) :: s_o             ! old flux vector
+  type(Vector) :: serr_v          ! error in source
+  type(GMRESSolver) :: gmres      ! gmres solver
 
 contains
 
 !===============================================================================
-! CMFD_POWER_EXECUTE
+! CMFD_POWER_EXECUTE sets up and runs power iteration solver for CMFD
 !===============================================================================
 
   subroutine cmfd_power_execute(k_tol, s_tol, adjoint)
 
-    use global,  only: cmfd_adjoint_type, n_procs_cmfd
+    use global,  only: cmfd_adjoint_type, time_cmfdbuild, time_cmfdsolve
 
-    real(8), optional :: k_tol    ! tolerance on keff
-    real(8), optional :: s_tol    ! tolerance on source
-    logical, optional :: adjoint  ! adjoint calc
+    real(8), intent(in), optional :: k_tol    ! tolerance on keff
+    real(8), intent(in), optional :: s_tol    ! tolerance on source
+    logical, intent(in), optional :: adjoint  ! adjoint calc
 
-    logical :: physical_adjoint = .false.
+    logical :: physical_adjoint = .false. ! physical adjoint default false
 
-    ! set tolerances if present
+    ! Set tolerances if present
     if (present(k_tol)) ktol = k_tol
     if (present(s_tol)) stol = s_tol
 
-    ! check for adjoint execution
+    ! Check for adjoint execution
+    adjoint_calc = .false.
     if (present(adjoint)) adjoint_calc = adjoint
 
-    ! check for physical adjoint
+    ! Check for physical adjoint
     if (adjoint_calc .and. trim(cmfd_adjoint_type) == 'physical') &
         physical_adjoint = .true.
 
-    ! initialize solver
-    call init_solver()
+    ! Start timer for build
+    call time_cmfdbuild % start()
 
-    ! initialize matrices and vectors
-    call init_data()
+    ! Initialize solver
+#ifdef PETSC
+    call gmres % create()
+#endif
 
-    ! set up M loss matrix
-    call build_loss_matrix(loss, adjoint=physical_adjoint) 
+    ! Initialize matrices and vectors
+    call init_data(physical_adjoint)
 
-    ! set up F production matrix
-    call build_prod_matrix(prod, adjoint=physical_adjoint)
-
-    ! check for adjoint calculation
+    ! Check for mathematical adjoint calculation
     if (adjoint_calc .and. trim(cmfd_adjoint_type) == 'math') &
         call compute_adjoint()
 
-    ! set up krylov info
-    call KSPSetOperators(krylov, loss%M, loss%M, SAME_NONZERO_PATTERN, ierr)
+    ! Set up krylov info
+#ifdef PETSC
+    call gmres % set_oper(loss, loss)
+#endif
 
-    ! precondition matrix
-    call precondition_matrix()
+    ! Stop timer for build
+    call time_cmfdbuild % stop()
 
-    ! begin power iteration 
+    ! Begin power iteration 
+    call time_cmfdsolve % start()
     call execute_power_iter()
+    call time_cmfdsolve % stop()
 
-    ! extract results
+    ! Extract results
     call extract_results()
 
-    ! deallocate petsc objects
+    ! Deallocate data 
     call finalize()
 
   end subroutine cmfd_power_execute
 
 !===============================================================================
-! INIT_DATA allocates matrices vectors for CMFD solution
+! INIT_DATA allocates matrices and vectors for CMFD solution
 !===============================================================================
 
-  subroutine init_data()
+  subroutine init_data(adjoint)
 
-    use constants, only: ONE
+    use constants, only: ONE, ZERO
+    use global,    only: cmfd_write_matrices
+
+    logical, intent(in) :: adjoint ! adjoint calcualtion
 
     integer :: n      ! problem size
     real(8) :: guess  ! initial guess
 
-    ! set up matrices
-    call init_M_operator(loss)
-    call init_F_operator(prod)
+    ! Set up matrices
+    call init_loss_matrix(loss)
+    call init_prod_matrix(prod)
 
-    ! get problem size
-    n = loss%localn
+    ! Get problem size
+    n = loss % n
 
-    ! set up flux vectors
-    call VecCreateMPI(PETSC_COMM_WORLD, n, PETSC_DECIDE, phi_n, ierr)
-    call VecCreateMPI(PETSC_COMM_WORLD, n, PETSC_DECIDE, phi_o, ierr)
+    ! Set up flux vectors
+    call phi_n % create(n)
+    call phi_o % create(n)
 
-    ! set up source vectors
-    call VecCreateMPI(PETSC_COMM_WORLD, n, PETSC_DECIDE, S_n, ierr)
-    call VecCreateMPI(PETSC_COMM_WORLD, n, PETSC_DECIDE, S_o, ierr)
+    ! Set up source vectors
+    call s_n % create(n)
+    call s_o % create(n)
+    call serr_v % create(n)
 
-    ! set initial guess
+    ! Set initial guess
     guess = ONE
-    call VecSet(phi_n, guess, ierr)
-    call VecSet(phi_o, guess, ierr)
+    phi_n % val = guess
+    phi_o % val = guess
     k_n = guess
     k_o = guess
 
+    ! Fill in loss matrix
+    call build_loss_matrix(loss, adjoint=adjoint) 
+
+    ! Fill in production matrix
+    call build_prod_matrix(prod, adjoint=adjoint)
+
+    ! Setup petsc for everything
+    call loss % assemble()
+    call prod % assemble()
+#ifdef PETSC
+    call loss % setup_petsc()
+    call prod % setup_petsc()
+    call phi_n % setup_petsc()
+    call phi_o % setup_petsc()
+    call s_o % setup_petsc()
+    call s_n % setup_petsc()
+    if (cmfd_write_matrices) call loss % write_petsc_binary('loss.bin')
+    if (cmfd_write_matrices) call prod % write_petsc_binary('prod.bin')
+#endif
+
+    ! Set norms to 0
+    norm_n = ZERO
+    norm_o = ZERO
+
   end subroutine init_data
 
 !===============================================================================
-! INIT_SOLVER
-!===============================================================================
-
-  subroutine init_solver()
-
-    real(8)  :: rtol ! relative tolerance
-    real(8)  :: atol ! absolute tolerance  
-
-    ! set tolerance
-    rtol = 1.0e-10_8
-    atol = 1.0e-10_8
-
-    ! set up krylov solver
-    call KSPCreate(PETSC_COMM_WORLD, krylov, ierr)
-    call KSPSetTolerances(krylov, rtol, atol, PETSC_DEFAULT_DOUBLE_PRECISION, &
-         PETSC_DEFAULT_INTEGER, ierr)
-    call KSPSetType(krylov, KSPGMRES, ierr)
-    call KSPSetInitialGuessNonzero(krylov, PETSC_TRUE, ierr)
-
-  end subroutine init_solver
-
-!===============================================================================
-! PRECONDITION_MATRIX 
-!===============================================================================
-
-  subroutine precondition_matrix()
-
-    use global, only: cmfd_power_monitor, cmfd_solver_type, &
-                      n_procs_cmfd, cmfd_ilu_levels, master
-    use string, only: to_str
-    use, intrinsic :: ISO_FORTRAN_ENV
-
-    character(len=20) :: ksptype,pctype
- 
-    ! set up preconditioner
-    call KSPGetPC(krylov, prec, ierr)
-    if (n_procs_cmfd == 1) then
-      call PCSetType(prec, PCILU, ierr)
-      call PCFactorSetLevels(prec, cmfd_ilu_levels, ierr)
-      call KSPSetUp(krylov, ierr)
-      call PCFactorGetMatrix(prec, loss%M, ierr)
-    else
-      call PetscOptionsSetValue("-pc_type", "bjacobi", ierr)
-      call PetscOptionsSetValue("-sub_pc_type", "ilu", ierr)  
-      call PetscOptionsSetValue("-sub_pc_factor_levels", &
-           trim(to_str(cmfd_ilu_levels)), ierr)
-      call PCSetFromOptions(prec, ierr)
-      call KSPSetUp(krylov, ierr)
-    end if
-
-    ! get options
-    if (trim(cmfd_solver_type) == 'power') call KSPSetFromOptions(krylov, ierr)
-
-    ! get all types and print
-    call KSPGetType(krylov, ksptype, ierr)
-    call PCGetType(prec, pctype, ierr)
-
-    ! print heading information
-    if (cmfd_power_monitor .and. master) then
-      write(OUTPUT_UNIT,'(A)') ''
-      write(OUTPUT_UNIT,'(A)') '########################################################'
-      write(OUTPUT_UNIT,'(A)') '################ Power Iteration Solver ################'
-      write(OUTPUT_UNIT,'(A)') '########################################################'
-      write(OUTPUT_UNIT,'(A)')
-      write(OUTPUT_UNIT,100)   'Eigenvalue Tolerance:', ktol
-      write(OUTPUT_UNIT,100)   'Source Tolerance:    ', stol
-      write(OUTPUT_UNIT,'(A)') ''
-      write(OUTPUT_UNIT,102)   'Linear Solver Type:  ', ksptype
-      write(OUTPUT_UNIT,102)   'Preconditioner Type: ', pctype
-      write(OUTPUT_UNIT,101)   'ILU Fill Levels:', cmfd_ilu_levels
-      write(OUTPUT_UNIT,'(A)') ''
-      write(OUTPUT_UNIT,'(A)') '---------------------------------------------'
-      write(OUTPUT_UNIT,'(A)') ''
-    end if
-
- 100 FORMAT(A,1X,1PE11.4)
- 101 FORMAT(A,1X,I0)
- 102 FORMAT(A,1X,A)
-
-  end subroutine precondition_matrix
-
-!===============================================================================
-! COMPUTE_ADJOINT 
+! COMPUTE_ADJOINT computes a mathematical adjoint of CMFD problem 
 !===============================================================================
 
   subroutine compute_adjoint()
 
     use global,  only: cmfd_write_matrices
 
-    PetscViewer :: viewer
+    ! Transpose matrices
+    call loss % transpose()
+    call prod % transpose()
 
-    ! transpose matrices
-    call MatTranspose(loss%M, MAT_REUSE_MATRIX, loss%M, ierr)
-    call MatTranspose(prod%F, MAT_REUSE_MATRIX, prod%F, ierr)
-
-    ! write out matrix in binary file (debugging)
+    ! Write out matrix in binary file (debugging)
     if (cmfd_write_matrices) then
-      call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_lossmat.bin', &
-           FILE_MODE_WRITE, viewer, ierr)
-      call MatView(loss%M, viewer, ierr)
-      call PetscViewerDestroy(viewer, ierr)
-
-      call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_prodmat.bin', &
-           FILE_MODE_WRITE, viewer, ierr)
-      call MatView(prod%F, viewer, ierr)
-      call PetscViewerDestroy(viewer, ierr)
+      call loss % write_petsc_binary('adj_lossmat.bin')
+      call prod % write_petsc_binary('adj_prodmat.bin')
     end if
 
   end subroutine compute_adjoint
 
 !===============================================================================
-! EXECUTE_POWER_ITER  in the main power iteration routine
+! EXECUTE_POWER_ITER  is the main power iteration routine
 !                     for the cmfd calculation
 !===============================================================================
 
   subroutine execute_power_iter()
 
-    use constants, only: ONE
+    integer :: i ! iteration counter
 
-    real(8)     :: num       ! numerator for eigenvalue update
-    real(8)     :: den       ! denominator for eigenvalue update
-    integer     :: i         ! iteration counter
-
-    ! reset convergence flag
+    ! Reset convergence flag
     iconv = .false.
 
-    ! begin power iteration
+    ! Begin power iteration
     do i = 1, 10000
 
-      ! compute source vector
-      call MatMult(prod%F, phi_o, S_o, ierr)
+      ! Compute source vector
+      call prod % vector_multiply(phi_o, s_o)
 
-      ! normalize source vector
-      call VecScale(S_o, ONE/k_o, ierr)
+      ! Normalize source vector
+      s_o % val = s_o % val / k_o
 
-      ! compute new flux vector
-      call KSPSolve(krylov, S_o, phi_n, ierr)
+      ! Compute new flux vector
+#ifdef PETSC
+      call gmres % solve(s_o, phi_n)
+#endif
 
-      ! compute new source vector
-      call MatMult(prod%F, phi_n, S_n, ierr)
+      ! Compute new source vector
+      call prod % vector_multiply(phi_n, s_n)
 
-      ! compute new k-eigenvalue
-      call VecSum(S_n, num, ierr)
-      call VecSum(S_o, den, ierr)
-      k_n = num/den
+      ! Compute new k-eigenvalue
+      k_n = sum(s_n % val) / sum(s_o % val)
 
-      ! renormalize the old source
-      call VecScale(S_o, k_o, ierr)
+      ! Renormalize the old source
+      s_o % val = s_o % val * k_o
 
-      ! check convergence
+      ! Check convergence
       call convergence(i)
 
-      ! to break or not to break
+      ! Break loop if converged
       if (iconv) exit
 
-      ! record old values
-      call VecCopy(phi_n, phi_o, ierr)
+      ! Record old values
+      phi_o % val = phi_n % val
       k_o = k_n
+      norm_o = norm_n
 
     end do
 
   end subroutine execute_power_iter 
 
 !===============================================================================
-! CONVERGENCE
+! CONVERGENCE checks the convergence of the CMFD problem
 !===============================================================================
 
   subroutine convergence(iter)
 
-    use global,  only: cmfd_power_monitor, master
+    use constants,  only: ONE, TINY_BIT
+    use global,     only: cmfd_power_monitor, master
     use, intrinsic :: ISO_FORTRAN_ENV
 
-    integer     :: iter           ! iteration number
+    integer, intent(in) :: iter ! iteration number
 
-    real(8)     :: kerr           ! error in keff
-    real(8)     :: serr           ! error in source
-    real(8)     :: norm_n         ! L2 norm of new source
-    real(8)     :: norm_o         ! L2 norm of old source
-    integer     :: ierr           ! petsc error code
-
-    ! reset convergence flag
+    ! Reset convergence flag
     iconv = .false.
 
-    ! calculate error in keff
+    ! Calculate error in keff
     kerr = abs(k_o - k_n)/k_n
 
-    ! calculate max error in source
-    call VecNorm(S_n, NORM_2, norm_n, ierr)
-    call VecNorm(S_o, NORM_2, norm_o, ierr)
-    serr = abs(norm_n-norm_o)/norm_n
+    ! Calculate max error in source
+    where (s_n % val > TINY_BIT)
+      serr_v % val = ((s_n % val - s_o % val)/s_n % val)**2
+    end where
+    serr = sqrt(ONE/dble(s_n % n) * sum(serr_v % val))
 
-    ! check for convergence
+    ! Check for convergence
     if(kerr < ktol .and. serr < stol) iconv = .true.
 
-    ! print out to user
+    ! Save the L2 norm of the source
+    norm_n = serr
+
+    ! Print out to user
     if (cmfd_power_monitor .and. master) then
-      write(OUTPUT_UNIT,FMT='(I0,":",T10,"k-eff: ",F0.8,T30,"k-error: ",1PE12.5,T55, &
-           "src-error: ",1PE12.5)') iter, k_n, kerr, serr
+      write(OUTPUT_UNIT,FMT='(I0,":",T10,"k-eff: ",F0.8,T30,"k-error: ", &
+           &1PE12.5,T55, "src-error: ",1PE12.5)') iter, k_n, kerr, serr
     end if
 
   end subroutine convergence
 
 !===============================================================================
-! EXTRACT_RESULTS 
+! EXTRACT_RESULTS takes results and puts them in CMFD global data object
 !===============================================================================
 
   subroutine extract_results()
 
-    use constants,        only: ZERO
-    use global,           only: cmfd, n_procs_cmfd, cmfd_write_matrices
+    use global, only: cmfd, cmfd_write_matrices, current_batch
 
+    character(len=25)    :: filename  ! name of file to write data 
     integer              :: n         ! problem size
-    integer              :: row_start ! local row start
-    integer              :: row_end   ! local row end
-    real(8),allocatable  :: mybuf(:)  ! temp buffer
-    PetscScalar, pointer :: phi_v(:)  ! pointer to eigenvector info
-    PetscViewer          :: viewer    ! petsc viewer for binary write
 
-    ! get problem size
-    n = loss%n
+    ! Get problem size
+    n = loss % n
 
-    ! also allocate in cmfd object
+    ! Allocate in cmfd object if not already allocated
     if (adjoint_calc) then
       if (.not. allocated(cmfd%adj_phi)) allocate(cmfd%adj_phi(n))
     else
       if (.not. allocated(cmfd%phi)) allocate(cmfd%phi(n))
     end if
-    if (.not. allocated(mybuf)) allocate(mybuf(n))
 
-    ! get ownership range
-    call VecGetOwnershipRange(phi_n, row_start, row_end, ierr)
-
-    ! convert petsc phi_object to cmfd_obj
-    call VecGetArrayF90(phi_n, phi_v, ierr)
+    ! Save values 
     if (adjoint_calc) then
-      cmfd%adj_phi(row_start+1:row_end) = phi_v
+      cmfd % adj_phi = phi_n % val
     else
-      cmfd%phi(row_start+1:row_end) = phi_v 
+      cmfd % phi = phi_n % val 
     end if
-    call VecRestoreArrayF90(phi_n, phi_v, ierr)
 
-    ! save eigenvalue
+    ! Save eigenvalue
     if(adjoint_calc) then
       cmfd%adj_keff = k_n
     else
       cmfd%keff = k_n
     end if
 
-    ! reduce result to all 
-    mybuf = ZERO
-    if (adjoint_calc) then
-      call MPI_ALLREDUCE(cmfd%adj_phi, mybuf, n, MPI_REAL8, MPI_SUM, &
-           PETSC_COMM_WORLD, ierr)
-    else
-      call MPI_ALLREDUCE(cmfd%phi, mybuf, n, MPI_REAL8, MPI_SUM, &
-           PETSC_COMM_WORLD, ierr)
-    end if
-
-    ! normalize phi to 1
+    ! Normalize phi to 1
     if (adjoint_calc) then
       cmfd%adj_phi = cmfd%adj_phi/sqrt(sum(cmfd%adj_phi*cmfd%adj_phi))
     else
       cmfd%phi = cmfd%phi/sqrt(sum(cmfd%phi*cmfd%phi))
     end if
 
-    ! write out results
+    ! Save dominance ratio
+    cmfd % dom(current_batch) = norm_n/norm_o
+
+    ! Write out results
     if (cmfd_write_matrices) then
       if (adjoint_calc) then
-        call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_fluxvec.bin', &
-             FILE_MODE_WRITE, viewer, ierr)
+        filename = 'adj_fluxvec.bin'
       else
-        call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'fluxvec.bin', &
-             FILE_MODE_WRITE, viewer, ierr)
+        filename = 'fluxvec.bin'
       end if
-      call VecView(phi_n, viewer, ierr)
-      call PetscViewerDestroy(viewer, ierr)
+      call phi_n % write_petsc_binary(filename)
     end if
 
-    ! nullify pointer and deallocate local vars
-    if (associated(phi_v)) nullify(phi_v)
-    if (allocated(mybuf)) deallocate(mybuf)
-
   end subroutine extract_results
 
 !===============================================================================
-! FINALIZE 
+! FINALIZE frees all memory associated with power iteration
 !===============================================================================
 
   subroutine finalize()
 
-    use global,  only: n_procs_cmfd
-
-    ! finalize solver objects
-    call KSPDestroy(krylov, ierr)
-    call KSPDestroy(sub_krylov, ierr)
-
-    ! finalize data objects
-    if (n_procs_cmfd > 1) call destroy_M_operator(loss) ! only destroy for jacobi
-    call destroy_F_operator(prod)
-
-    call VecDestroy(phi_n, ierr)
-    call VecDestroy(phi_o, ierr)
-    call VecDestroy(S_n, ierr)
-    call VecDestroy(S_o, ierr)
+    ! Destroy all objects 
+#ifdef PETSC
+    call gmres  % destroy()
+#endif
+    call loss   % destroy() 
+    call prod   % destroy()
+    call phi_n  % destroy()
+    call phi_o  % destroy()
+    call s_n    % destroy()
+    call s_o    % destroy()
+    call serr_v % destroy
 
   end subroutine finalize
 
-# endif
-
 end module cmfd_power_solver
diff --git a/src/cmfd_prod_operator.F90 b/src/cmfd_prod_operator.F90
index a5eac6df74..e6d055ced8 100644
--- a/src/cmfd_prod_operator.F90
+++ b/src/cmfd_prod_operator.F90
@@ -1,66 +1,72 @@
 module cmfd_prod_operator
 
-# ifdef PETSC
+  use constants,      only: CMFD_NOACCEL
+  use global,         only: cmfd, cmfd_coremap
+  use matrix_header,  only: Matrix
 
   implicit none
   private
-  public :: init_F_operator,build_prod_matrix,destroy_F_operator
-
-# include <finclude/petsc.h90>
-
-  integer  :: nx   ! maximum number of x cells
-  integer  :: ny   ! maximum number of y cells
-  integer  :: nz   ! maximum number of z cells
-  integer  :: ng   ! maximum number of groups
-  integer  :: ierr ! petsc error code
-
-  type, public :: prod_operator
-
-    Mat      :: F    ! petsc matrix for neutronic prod operator
-    integer  :: n    ! dimensions of matrix
-    integer  :: nnz  ! max number of nonzeros
-    integer  :: localn ! local size on proc
-    integer, allocatable :: d_nnz(:) ! vector of diagonal preallocation
-    integer, allocatable :: o_nnz(:) ! vector of off-diagonal preallocation
-
-  end type prod_operator
-
-  logical :: adjoint_calc = .false. ! adjoint calculation 
+  public :: init_prod_matrix, build_prod_matrix
 
 contains
 
 !==============================================================================
-! INIT_F_OPERATOR
+! INIT_PROD_MATRIX preallocates prod matrix and initializes it
 !==============================================================================
 
-  subroutine init_F_operator(this)
+  subroutine init_prod_matrix(prod_matrix)
 
-    type(prod_operator) :: this
+    type(Matrix), intent(inout) :: prod_matrix ! production matrix
 
-    ! get indices
-    call get_F_indices(this)
+    integer :: nx   ! maximum number of x cells
+    integer :: ny   ! maximum number of y cells
+    integer :: nz   ! maximum number of z cells
+    integer :: ng   ! maximum number of groups
+    integer :: n    ! total length of matrix
+    integer :: nnz  ! number of nonzeros in matrix
 
-    ! get preallocation
-    call preallocate_prod_matrix(this)
+    ! Get maximum number of cells in each direction
+    nx = cmfd%indices(1)
+    ny = cmfd%indices(2)
+    nz = cmfd%indices(3)
+    ng = cmfd%indices(4)
 
-    ! set up M operator
-    call MatCreateAIJ(PETSC_COMM_WORLD, this%localn, this%localn, PETSC_DECIDE, &
-         PETSC_DECIDE, PETSC_NULL_INTEGER, this%d_nnz, PETSC_NULL_INTEGER, &
-         this%o_nnz, this%F, ierr)
-    call MatSetOption(this%F, MAT_NEW_NONZERO_LOCATIONS, PETSC_TRUE, ierr)
-    call MatSetOption(this%F, MAT_IGNORE_ZERO_ENTRIES, PETSC_TRUE, ierr)
+    ! Calculate dimensions and number of nonzeros in matrix
+    if (cmfd_coremap) then
+      n = cmfd % mat_dim * ng
+    else
+      n = nx*ny*nz*ng
+    end if
+    nnz = n * ng
 
-  end subroutine init_F_operator
+    ! Configure prod matrix
+    call prod_matrix % create(n, nnz)
 
-!==============================================================================
-! GET_F_INDICES
-!==============================================================================
+  end subroutine init_prod_matrix
 
-  subroutine get_F_indices(this)
+!===============================================================================
+! BUILD_PROD_MATRIX creates the matrix representing production of neutrons
+!===============================================================================
 
-    use global,  only: cmfd, cmfd_coremap
+  subroutine build_prod_matrix(prod_matrix, adjoint)
 
-    type(prod_operator) :: this
+    type(Matrix), intent(inout)   :: prod_matrix ! production matrix
+    logical, intent(in), optional :: adjoint     ! adjoint calculation logical
+
+    integer :: i                  ! iteration counter for x
+    integer :: j                  ! iteration counter for y
+    integer :: k                  ! iteration counter for z
+    integer :: g                  ! iteration counter for groups
+    integer :: h                  ! energy group when doing scattering
+    integer :: nx                 ! maximum number of x cells
+    integer :: ny                 ! maximum number of y cells
+    integer :: nz                 ! maximum number of z cells
+    integer :: ng                 ! maximum number of groups
+    integer :: hmat_idx           ! index in matrix for energy group h
+    integer :: irow               ! iteration counter over row
+    logical :: adjoint_calc       ! is this a physical adjoint?
+    real(8) :: nfissxs            ! nufission cross section h-->g
+    real(8) :: val                ! temporary variable for nfissxs
 
     ! get maximum number of cells in each direction
     nx = cmfd%indices(1)
@@ -68,162 +74,18 @@ contains
     nz = cmfd%indices(3)
     ng = cmfd%indices(4)
 
-    ! get number of nonzeros
-    this%nnz = 7 + ng - 1
-
-    ! calculate dimensions of matrix
-    if (cmfd_coremap) then
-      this%n = cmfd % mat_dim * ng
-    else
-      this%n = nx*ny*nz*ng
-    end if
-
-  end subroutine get_F_indices
-
-!===============================================================================
-! PREALLOCATE_PROD_MATRIX
-!===============================================================================
-
-  subroutine preallocate_prod_matrix(this)
-
-    use constants,  only: CMFD_NOACCEL
-    use global,     only: cmfd, cmfd_coremap
-
-    type(prod_operator) :: this
-
-    integer :: rank          ! rank of processor
-    integer :: sizen         ! number of procs
-    integer :: i             ! iteration counter for x
-    integer :: j             ! iteration counter for y
-    integer :: k             ! iteration counter for z
-    integer :: g             ! iteration counter for groups
-    integer :: h             ! energy group when doing scattering
-    integer :: n             ! the extent of the matrix
-    integer :: irow          ! row counter
-    integer :: row_start     ! index of local starting row
-    integer :: row_end       ! index of local final row
-    integer :: hmat_idx      ! index in matrix for energy group h
-
-    ! initialize size and rank
-    rank = 0
-    sizen = 0
-
-    ! get rank and max rank of procs
-    call MPI_COMM_RANK(PETSC_COMM_WORLD, rank, ierr)
-    call MPI_COMM_SIZE(PETSC_COMM_WORLD, sizen, ierr)
-
-    ! get local problem size
-    n = this%n
-
-    ! determine local size, divide evenly between all other procs
-    this%localn = n/(sizen)
-
-    ! add 1 more if less proc id is less than mod
-    if (rank < mod(n, sizen)) this%localn = this%localn + 1
-
-    ! determine local starting row
-    row_start = 0
-    if (rank < mod(n, sizen)) then
-      row_start = rank*(n/sizen+1)
-    else
-      row_start = min(mod(n, sizen)*(n/sizen+1) + &
-           (rank - mod(n, sizen))*(n/sizen), n)
-    end if
-
-    ! determine local final row
-    row_end = row_start + this%localn - 1
-
-    ! allocate counters
-    if (.not. allocated(this%d_nnz)) allocate(this%d_nnz(row_start:row_end))
-    if (.not. allocated(this%o_nnz)) allocate(this%o_nnz(row_start:row_end))
-    this % d_nnz = 0
-    this % o_nnz = 0
-
-    ! begin loop around local rows
-    ROWS: do irow = row_start, row_end
-
-      ! initialize counters 
-      this%d_nnz(irow) = 1 ! already add in matrix diagonal
-      this%o_nnz(irow) = 0
-
-      ! get location indices
-      call matrix_to_indices(irow, g, i, j, k)
-
-      ! check if not including reflector
-      if (cmfd_coremap) then
-
-        ! check if at a reflector
-        if (cmfd % coremap(i,j,k) == CMFD_NOACCEL) then
-          cycle
-        end if
-
-      end if
-
-      ! begin loop over off diagonal in-scattering
-      NFISS: do h = 1, ng
-
-        ! cycle though if h=g
-        if (h == g) then
-          cycle
-        end if
-
-        ! get neighbor matrix index
-        call indices_to_matrix(h, i, j, k, hmat_idx)
-
-        ! record nonzero
-        if (((hmat_idx-1) >= row_start) .and. &
-             ((hmat_idx-1) <= row_end)) then
-          this%d_nnz(irow) = this%d_nnz(irow) + 1
-        else
-          this%o_nnz(irow) = this%o_nnz(irow) + 1
-        end if
-
-      end do NFISS
-
-    end do ROWS
-
-  end subroutine preallocate_prod_matrix
-
-!===============================================================================
-! BUILD_PROD_MATRIX creates the matrix representing loss of neutrons
-!===============================================================================
-
-  subroutine build_prod_matrix(this, adjoint)
-
-    use constants,  only: CMFD_NOACCEL
-    use global,     only: cmfd, cmfd_coremap, cmfd_write_matrices
-
-    type(prod_operator) :: this
-    logical, optional :: adjoint
-
-    integer :: i                  ! iteration counter for x
-    integer :: j                  ! iteration counter for y
-    integer :: k                  ! iteration counter for z
-    integer :: g                  ! iteration counter for groups
-    integer :: h                  ! energy group when doing scattering
-    integer :: hmat_idx           ! index in matrix for energy group h
-    integer :: ierr               ! Petsc error code
-    integer :: row_start          ! the first local row on the processor
-    integer :: row_finish         ! the last local row on the processor
-    integer :: irow               ! iteration counter over row
-    real(8) :: nfissxs            ! nufission cross section h-->g
-    real(8) :: val                ! temporary variable for nfissxs
-
     ! check for adjoint
+    adjoint_calc = .false.
     if (present(adjoint)) adjoint_calc = adjoint
 
-    ! initialize row start and finish
-    row_start = 0
-    row_finish = 0
-
-    ! get row bounds for this processor
-    call MatGetOwnershipRange(this%F, row_start, row_finish, ierr)
-
     ! begin iteration loops
-    ROWS: do irow = row_start, row_finish-1
+    ROWS: do irow = 1, prod_matrix % n
+
+      ! add a new row to matrix
+      call prod_matrix % new_row()
 
       ! get indices for that row
-      call matrix_to_indices(irow, g, i, j, k)
+      call matrix_to_indices(irow, g, i, j, k, ng, nx, ny, nz)
 
       ! check if not including reflector
       if (cmfd_coremap) then
@@ -236,34 +98,34 @@ contains
       end if
 
       ! loop around all other groups 
+
       NFISS: do h = 1, ng
 
-        ! get cell data
-        nfissxs = cmfd%nfissxs(h,g,i,j,k)
-
         ! get matrix column location
-        call indices_to_matrix(h, i, j, k, hmat_idx)
-
-        ! reocrd value in matrix
-        val = nfissxs
+        call indices_to_matrix(h, i, j, k, hmat_idx, ng, nx, ny)
 
         ! check for adjoint and bank val
         if (adjoint_calc) then
-          call MatSetValue(this%F, hmat_idx-1, irow, val, INSERT_VALUES, ierr)
+          ! get nu-fission cross section from cell
+          nfissxs = cmfd%nfissxs(g,h,i,j,k)
         else
-          call MatSetValue(this%F, irow, hmat_idx-1, val, INSERT_VALUES, ierr)
+          ! get nu-fission cross section from cell
+          nfissxs = cmfd%nfissxs(h,g,i,j,k)
         end if
 
+        ! set as value to be recorded
+        val = nfissxs
+
+        ! record value in matrix
+
+        call prod_matrix % add_value(hmat_idx, val)
+
       end do NFISS
 
     end do ROWS 
 
-    ! assemble matrix 
-    call MatAssemblyBegin(this%F, MAT_FLUSH_ASSEMBLY, ierr)
-    call MatAssemblyEnd(this%F, MAT_FINAL_ASSEMBLY, ierr)
-
-    ! print out operator to file
-    if (cmfd_write_matrices) call print_F_operator(this)
+    ! CSR requires n+1 row
+    call prod_matrix % new_row()
 
   end subroutine build_prod_matrix
 
@@ -271,19 +133,20 @@ contains
 ! INDICES_TO_MATRIX takes (x,y,z,g) indices and computes location in matrix
 !===============================================================================
 
-  subroutine indices_to_matrix(g, i, j, k, matidx)
+  subroutine indices_to_matrix(g, i, j, k, matidx, ng, nx, ny)
 
-    use global,  only: cmfd, cmfd_coremap
+    integer, intent(out) :: matidx ! the index location in matrix
+    integer, intent(in)  :: i      ! current x index
+    integer, intent(in)  :: j      ! current y index
+    integer, intent(in)  :: k      ! current z index
+    integer, intent(in)  :: g      ! current group index
+    integer, intent(in)  :: nx     ! maximum number of x cells
+    integer, intent(in)  :: ny     ! maximum number of y cells
+    integer, intent(in)  :: ng     ! maximum number of groups
 
-    integer :: matidx         ! the index location in matrix
-    integer :: i              ! current x index
-    integer :: j              ! current y index
-    integer :: k              ! current z index
-    integer :: g              ! current group index
-    
     ! check if coremap is used
     if (cmfd_coremap) then
-    
+
       ! get idx from core map
       matidx = ng*(cmfd % coremap(i,j,k)) - (ng - g)
 
@@ -297,80 +160,40 @@ contains
   end subroutine indices_to_matrix
 
 !===============================================================================
-! MATRIX_TO_INDICES
+! MATRIX_TO_INDICES converts matrix index to spatial and group indices
 !===============================================================================
 
-  subroutine matrix_to_indices(irow, g, i, j, k)
+  subroutine matrix_to_indices(irow, g, i, j, k, ng, nx, ny, nz)
 
-    use global,  only: cmfd, cmfd_coremap
-
-    integer :: i                    ! iteration counter for x
-    integer :: j                    ! iteration counter for y
-    integer :: k                    ! iteration counter for z
-    integer :: g                    ! iteration counter for groups
-    integer :: irow                 ! iteration counter over row (0 reference)
+    integer, intent(out) :: i    ! iteration counter for x
+    integer, intent(out) :: j    ! iteration counter for y
+    integer, intent(out) :: k    ! iteration counter for z
+    integer, intent(out) :: g    ! iteration counter for groups
+    integer, intent(in)  :: irow ! iteration counter over row (0 reference)
+    integer, intent(in)  :: nx   ! maximum number of x cells
+    integer, intent(in)  :: ny   ! maximum number of y cells
+    integer, intent(in)  :: nz   ! maximum number of z cells
+    integer, intent(in)  :: ng   ! maximum number of groups
 
     ! check for core map
     if (cmfd_coremap) then
 
       ! get indices from indexmap
-      g = mod(irow,ng) + 1
-      i = cmfd % indexmap(irow/ng+1,1)
-      j = cmfd % indexmap(irow/ng+1,2)
-      k = cmfd % indexmap(irow/ng+1,3)
+      g = mod(irow-1, ng) + 1
+      i = cmfd % indexmap((irow-1)/ng+1,1)
+      j = cmfd % indexmap((irow-1)/ng+1,2)
+      k = cmfd % indexmap((irow-1)/ng+1,3)
 
     else
 
       ! compute indices
-      g = mod(irow, ng) + 1
-      i = mod(irow, ng*nx)/ng + 1
-      j = mod(irow, ng*nx*ny)/(ng*nx)+ 1
-      k = mod(irow, ng*nx*ny*nz)/(ng*nx*ny) + 1
+      g = mod(irow-1, ng) + 1
+      i = mod(irow-1, ng*nx)/ng + 1
+      j = mod(irow-1, ng*nx*ny)/(ng*nx)+ 1
+      k = mod(irow-1, ng*nx*ny*nz)/(ng*nx*ny) + 1
 
     end if
 
   end subroutine matrix_to_indices
 
-!===============================================================================
-! PRINT_F_OPERATOR
-!===============================================================================
-
-  subroutine print_F_operator(this)
-
-    type(prod_operator) :: this
-
-    PetscViewer :: viewer
-
-    ! write out matrix in binary file (debugging)
-    if (adjoint_calc) then
-      call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_prodmat.bin', &   
-           FILE_MODE_WRITE, viewer, ierr)
-    else
-      call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'prodmat.bin', &
-           FILE_MODE_WRITE, viewer, ierr)
-    end if
-    call MatView(this%F, viewer, ierr)
-    call PetscViewerDestroy(viewer, ierr)
-
-  end subroutine print_F_operator
-
-!==============================================================================
-! DESTROY_F_OPERATOR
-!==============================================================================
-
-  subroutine destroy_F_operator(this)
-
-    type(prod_operator) :: this
-
-    ! deallocate matrix
-    call MatDestroy(this%F, ierr)
-
-    ! deallocate other parameters
-    if (allocated(this%d_nnz)) deallocate(this%d_nnz)
-    if (allocated(this%o_nnz)) deallocate(this%o_nnz)
-
-  end subroutine destroy_F_operator
-
-# endif 
-
 end module cmfd_prod_operator
diff --git a/src/cmfd_snes_solver.F90 b/src/cmfd_snes_solver.F90
deleted file mode 100644
index 6293720207..0000000000
--- a/src/cmfd_snes_solver.F90
+++ /dev/null
@@ -1,473 +0,0 @@
-module cmfd_snes_solver
-
-# ifdef PETSC
-
-  use cmfd_loss_operator,     only: loss_operator,init_M_operator,           &
-                                    build_loss_matrix,destroy_M_operator
-  use cmfd_prod_operator,     only: prod_operator,init_F_operator,           &
-                                    build_prod_matrix,destroy_F_operator
-  use cmfd_jacobian_operator, only: jacobian_operator,init_J_operator,       &
-                                    build_jacobian_matrix,destroy_J_operator,&
-                                    operators
-  use cmfd_power_solver,      only: cmfd_power_execute
-
-  implicit none
-  private
-  public :: cmfd_snes_execute
-
-# include <finclude/petsc.h90>
-
-  type(jacobian_operator) :: jac_prec
-  type(operators) :: ctx
-
-  Mat         :: jac                    ! jacobian matrix
-  Vec         :: resvec                 ! residual vector
-  Vec         :: xvec                   ! results
-  KSP         :: ksp                    ! linear solver context
-  PC          :: pc                     ! preconditioner
-  SNES        :: snes                   ! nonlinear solver context
-  integer     :: ierr                   ! error flag
-
-  logical     :: adjoint_calc = .false. ! adjoint calculation
-
-contains
-
-!===============================================================================
-! CMFD_SNES_EXECUTE
-!===============================================================================
-
-  subroutine cmfd_snes_execute(adjoint)
-
-    logical, optional :: adjoint ! adjoint calculation
-
-    ! check for adjoint
-    if (present(adjoint)) adjoint_calc = adjoint
-
-    ! seed with power iteration
-    call cmfd_power_execute(k_tol=1.E-3_8, s_tol=1.E-3_8, adjoint=adjoint_calc)
-
-    ! initialize data
-    call init_data()
-
-    ! initialize solver
-    call init_solver()
-
-    ! precondition
-    call precondition_matrix()
-
-    ! solve the system
-    call SNESSolve(snes, PETSC_NULL_DOUBLE, xvec, ierr)
-
-    ! extracts results to cmfd object
-    call extract_results()
-
-    ! deallocate all slepc data
-    call finalize()
-
-  end subroutine cmfd_snes_execute
-
-!===============================================================================
-! INIT_DATA allocates matrices vectors for CMFD solution
-!===============================================================================
-
-  subroutine init_data()
-
-    use constants, only: ONE
-    use global,    only: cmfd, n_procs_cmfd, rank
-
-    integer              :: k         ! implied do counter
-    integer              :: n         ! problem size
-    integer              :: row_start ! local row start
-    integer              :: row_end   ! local row end
-    real(8), pointer     :: xptr(:)   ! solution pointer
-
-    ! set up operator matrices
-    call init_M_operator(ctx%loss)
-    call init_F_operator(ctx%prod)
-    call init_J_operator(jac_prec, ctx)
-
-    ! get problem size
-    n = jac_prec%n
-
-    ! create PETSc vectors
-    call VecCreateMPI(PETSC_COMM_WORLD, jac_prec%localn, PETSC_DECIDE, &
-         resvec, ierr)
-    call VecCreateMPI(PETSC_COMM_WORLD, jac_prec%localn, PETSC_DECIDE, &
-         xvec, ierr)
-
-    ! get the local dimensions for each process
-    call VecGetOwnershipRange(xvec, row_start, row_end, ierr)
-
-    if (rank == n_procs_cmfd - 1) row_end = n
-
-    ! set flux in guess
-    if (adjoint_calc) then
-      call VecSetValues(xvec, row_end-row_start, (/(k,k=row_start,row_end-1)/), &
-           cmfd%adj_phi(row_start+1:row_end), INSERT_VALUES, ierr)
-    else
-      call VecSetValues(xvec, row_end-row_start, (/(k,k=row_start,row_end-1)/), &
-           cmfd%phi(row_start+1:row_end), INSERT_VALUES, ierr)
-    end if
-    call VecAssemblyBegin(xvec, ierr)
-    call VecAssemblyEnd(xvec, ierr)
-
-    ! set keff in guess
-    if (rank == n_procs_cmfd - 1) then
-      call VecGetArrayF90(xvec, xptr, ierr)
-      if (adjoint_calc) then
-        xptr(size(xptr)) = ONE/cmfd%adj_keff
-      else
-        xptr(size(xptr)) = ONE/cmfd%keff
-      end if
-      call VecRestoreArrayF90(xvec, xptr, ierr)
-    end if
-
-    ! nullify pointers
-    if (associated(xptr)) nullify(xptr)
-
-  end subroutine init_data
-
-!===============================================================================
-! INIT_SOLVER
-!===============================================================================
-
-  subroutine init_solver()
-
-    use global,  only: cmfd_snes_monitor, n_procs_cmfd
-
-    ! turn on mf_operator option
-    call PetscOptionsSetValue("-snes_mf_operator", "TRUE", ierr)
-
-    ! create SNES context
-    call SNESCreate(PETSC_COMM_WORLD, snes, ierr)
-
-    ! set the residual function
-    call SNESSetFunction(snes, resvec, compute_nonlinear_residual, &
-         PETSC_NULL_DOUBLE, ierr)
-
-    ! set GMRES solver
-    call SNESGetKSP(snes, ksp, ierr)
-    call KSPSetType(ksp, KSPGMRES, ierr)
-
-    ! create matrix free jacobian
-    call MatCreateSNESMF(snes, jac, ierr)
-
-    ! set matrix free finite difference
-    call SNESSetJacobian(snes, jac, jac_prec, build_jacobian_matrix, ctx, ierr)
-
-    ! set lags
-    call SNESSetLagJacobian(snes, -2, ierr)
-    call SNESSetLagPreconditioner(snes, -1, ierr)
-
-    ! set convergence
-!   call SNESSetTolerances(snes, 1.0e-8_8, 1.0e-8_8, 1.0e-8_8, 50, 10000, ierr)
-
-    ! set SNES options
-    call SNESSetFromOptions(snes, ierr)
-
-    ! turn off line searching
-!   call SNESLineSearchSet(snes, SNESLineSearchNo, PETSC_NULL_OBJECT, ierr)
-
-  end subroutine init_solver
-
-!===============================================================================
-! PRECONDITION_MATRIX 
-!===============================================================================
-
-  subroutine precondition_matrix()
-
-    use global,                 only: cmfd_snes_monitor, cmfd_solver_type, &
-                                      n_procs_cmfd, cmfd_ilu_levels, master
-    use string,                 only: to_str
-    use, intrinsic :: ISO_FORTRAN_ENV
-
-    character(LEN=20) :: snestype, ksptype, pctype
-
-    ! set up preconditioner
-    call KSPGetPC(ksp, pc, ierr)
-    if (n_procs_cmfd == 1) then
-      call PCSetType(pc, PCILU, ierr)
-      call PCFactorSetLevels(pc, cmfd_ilu_levels, ierr)
-    else
-      call PetscOptionsSetValue("-pc_type", "bjacobi", ierr)
-      call PetscOptionsSetValue("-sub_pc_type", "ilu", ierr)
-      call PetscOptionsSetValue("-sub_pc_factor_levels", &
-           trim(to_str(cmfd_ilu_levels)), ierr)
-    end if
-
-    ! get options
-    call PCSetFromOptions(pc, ierr)
-    call KSPSetFromOptions(ksp, ierr)
-
-    ! finalize ksp setup
-!   call KSPSetUp(ksp, ierr)
-
-    ! get all types and print
-    call SNESGetType(snes, snestype, ierr)
-    call KSPGetType(ksp, ksptype, ierr)
-    call PCGetType(pc, pctype, ierr)
-
-    ! display solver info to user
-    if (cmfd_snes_monitor .and. master) then
-      write(OUTPUT_UNIT,'(A)') ''
-      write(OUTPUT_UNIT,'(A)') '########################################################'
-      write(OUTPUT_UNIT,'(A)') '################ JFNK Nonlinear Solver  ################'
-      write(OUTPUT_UNIT,'(A)') '########################################################'
-      write(OUTPUT_UNIT,'(A)') ''
-      write(OUTPUT_UNIT,101) 'NONLINEAR SOLVER: ', snestype
-      write(OUTPUT_UNIT,101) 'LINEAR SOLVER:    ', ksptype
-      write(OUTPUT_UNIT,101) 'PRECONDITIONER:   ', pctype
-      write(OUTPUT_UNIT,100) 'ILU levels: ', cmfd_ilu_levels
-      write(OUTPUT_UNIT,'(A)') ''
-      write(OUTPUT_UNIT,'(A)') '---------------------------------------------'
-      write(OUTPUT_UNIT,'(A)') ''
-    end if
-
- 100 FORMAT(A,1X,I0)
- 101 FORMAT(A,1X,A)
- 
-  end subroutine precondition_matrix
-
-!===============================================================================
-! COMPUTE_NONLINEAR_RESIDUAL
-!===============================================================================
-
-  subroutine compute_nonlinear_residual(snes, x, res, ierr)
-
-    use global,       only: n_procs_cmfd, cmfd_write_matrices, &
-                            cmfd_adjoint_type, rank
-
-    SNES        :: snes          ! nonlinear solver context
-    Vec         :: x             ! independent vector
-    Vec         :: res           ! residual vector
-    integer     :: ierr          ! error flag
-
-    Vec         :: phi           ! flux vector
-    Vec         :: rphi          ! flux part of residual
-    Vec         :: phiM          ! M part of residual flux calc
-    integer     :: n             ! problem size
-    real(8)     :: lambda        ! eigenvalue
-    real(8)     :: reslamb       ! residual for lambda
-
-    real(8), pointer :: xptr(:)  ! pointer to solution vector
-    real(8), pointer :: rptr(:)  ! pointer to residual vector
-    PetscViewer :: viewer
-
-    logical :: physical_adjoint = .false.
-
-    ! check for physical adjoint
-    if (adjoint_calc .and. trim(cmfd_adjoint_type) == 'physical') &
-         physical_adjoint = .true.
-
-    ! create operators
-    call build_loss_matrix(ctx%loss, adjoint = physical_adjoint)
-    call build_prod_matrix(ctx%prod, adjoint = physical_adjoint)
-
-    ! check for adjoint
-    if (adjoint_calc .and. trim(cmfd_adjoint_type) == 'math') &
-         call compute_adjoint()
-
-    ! get problem size
-    n = ctx%loss%n 
-
-    ! get pointers to vectors
-    call VecGetArrayF90(x, xptr, ierr)
-    call VecGetArrayF90(res, rptr, ierr)
-
-    ! create petsc vector for flux
-    call VecCreateMPI(PETSC_COMM_WORLD, ctx%loss%localn, &
-         PETSC_DECIDE, phi, ierr)
-    call VecCreateMPI(PETSC_COMM_WORLD, ctx%loss%localn, &
-         PETSC_DECIDE, rphi, ierr)
-
-    ! extract flux and place in petsc vector 
-    call VecPlaceArray(phi, xptr, ierr)
-    call VecPlaceArray(rphi, rptr, ierr)
-
-    ! extract eigenvalue and broadcast (going to want to make this more general in future)
-    if (rank == n_procs_cmfd - 1) lambda = xptr(size(xptr)) 
-    call MPI_BCAST(lambda, 1, MPI_REAL8, n_procs_cmfd-1, PETSC_COMM_WORLD, ierr)
-
-    ! create new petsc vectors to perform math
-    call VecCreateMPI(PETSC_COMM_WORLD, ctx%loss%localn, PETSC_DECIDE, &
-         phiM, ierr)
-
-    ! calculate flux part of residual vector
-    call MatMult(ctx%loss%M, phi, phiM, ierr)
-    call MatMult(ctx%prod%F, phi, rphi, ierr)
-    call VecAYPX(rphi, -lambda, phiM, ierr)
-
-    ! set eigenvalue part of residual vector
-    call VecDot(phi, phi, reslamb, ierr)
-
-    ! map to ptr
-    if (rank == n_procs_cmfd - 1) rptr(size(rptr)) = 0.5_8 - 0.5_8*reslamb
-
-    ! reset arrays that are not used
-    call VecResetArray(phi, ierr)
-    call VecResetArray(rphi, ierr)
-
-    ! restore arrays for residual and solution
-    call VecRestoreArrayF90(x, xptr, ierr)
-    call VecRestoreArrayF90(res, rptr, ierr)
-
-    ! destroy all temp vectors
-    call VecDestroy(phi, ierr)
-    call VecDestroy(phiM, ierr)
-    call VecDestroy(rphi, ierr)
-
-    ! nullify all pointers
-    if (associated(xptr)) nullify(xptr)
-    if (associated(rptr)) nullify(rptr)
-
-    ! write out matrix in binary file (debugging)
-    if (cmfd_write_matrices) then
-      if (adjoint_calc) then
-        call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_res.bin', &
-             FILE_MODE_WRITE, viewer, ierr)
-      else
-        call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'res.bin', &
-             FILE_MODE_WRITE, viewer, ierr)
-      end if
-      call VecView(res, viewer, ierr)
-      call PetscViewerDestroy(viewer, ierr)
-
-      if (adjoint_calc) then
-        call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_x.bin', &
-             FILE_MODE_WRITE, viewer, ierr)
-      else
-        call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'x.bin', &
-             FILE_MODE_WRITE, viewer, ierr)
-      end if
-       call VecView(x, viewer, ierr)
-       call PetscViewerDestroy(viewer, ierr)
-
-    end if
-
-  end subroutine compute_nonlinear_residual
-
-!===============================================================================
-! COMPUTE_ADJOINT
-!===============================================================================
-
-  subroutine compute_adjoint()
-
-    use global,  only: cmfd_write_matrices
-
-    PetscViewer :: viewer
-
-    ! transpose matrices
-    call MatTranspose(ctx%loss%M, MAT_REUSE_MATRIX, ctx%loss%M, ierr)
-    call MatTranspose(ctx%prod%F, MAT_REUSE_MATRIX, ctx%prod%F, ierr)
-
-    ! write out matrix in binary file (debugging)
-    if (cmfd_write_matrices) then
-      call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_lossmat.bin', &
-           FILE_MODE_WRITE, viewer, ierr)
-      call MatView(ctx%loss%M, viewer, ierr)
-      call PetscViewerDestroy(viewer, ierr)
-
-      call PetscViewerBinaryOpen(PETSC_COMM_WORLD, 'adj_prodmat.bin', &
-           FILE_MODE_WRITE, viewer, ierr)
-      call MatView(ctx%prod%F, viewer, ierr)
-      call PetscViewerDestroy(viewer, ierr)
-    end if
-
-  end subroutine compute_adjoint
-
-!===============================================================================
-! EXTRACT_RESULTS 
-!===============================================================================
-
-  subroutine extract_results()
-
-    use constants, only: ZERO, ONE
-    use global,    only: cmfd, n_procs_cmfd, rank
-
-    integer              :: n         ! problem size
-    integer              :: row_start ! local row start
-    integer              :: row_end   ! local row end
-    real(8)              :: keff      ! keff of problem
-    real(8),allocatable  :: mybuf(:)  ! temp buffer
-    PetscScalar, pointer :: xptr(:)   ! pointer to eigenvector info
-
-    ! get problem size
-    n = ctx%loss%n
-
-    ! also allocate in cmfd object
-    if (adjoint_calc) then
-      if (.not. allocated(cmfd%adj_phi)) allocate(cmfd%adj_phi(n))
-    else
-      if (.not. allocated(cmfd%phi)) allocate(cmfd%phi(n))
-    end if
-    if (.not. allocated(mybuf)) allocate(mybuf(n))
-
-    ! get ownership range
-    call VecGetOwnershipRange(xvec, row_start, row_end, ierr)
-
-    ! resize the last proc
-    if (rank == n_procs_cmfd - 1) row_end = row_end - 1
-
-    ! convert petsc phi_object to cmfd_obj
-    call VecGetArrayF90(xvec, xptr, ierr)
-    if (adjoint_calc) then
-      cmfd%adj_phi(row_start+1:row_end) = xptr(1:row_end - row_start)
-    else
-      cmfd%phi(row_start+1:row_end) = xptr(1:row_end - row_start) 
-    end if
-
-    ! reduce result to all 
-    mybuf = ZERO
-    if (adjoint_calc) then
-      call MPI_ALLREDUCE(cmfd%adj_phi, mybuf, n, MPI_REAL8, MPI_SUM, &
-           PETSC_COMM_WORLD, ierr)
-    else
-      call MPI_ALLREDUCE(cmfd%phi, mybuf, n, MPI_REAL8, MPI_SUM, &
-           PETSC_COMM_WORLD, ierr)
-    end if
-
-    ! move buffer to object and deallocate
-    cmfd%phi = mybuf
-    if(allocated(mybuf)) deallocate(mybuf)
-
-    ! save eigenvalue
-    if(rank == n_procs_cmfd - 1) keff = ONE / xptr(size(xptr))
-    if (adjoint_calc) then
-      cmfd%adj_keff = keff
-      call MPI_BCAST(cmfd%adj_keff, 1, MPI_REAL8, n_procs_cmfd-1, &
-           PETSC_COMM_WORLD, ierr)
-    else
-      cmfd%keff = keff
-      call MPI_BCAST(cmfd%keff, 1, MPI_REAL8, n_procs_cmfd-1, &
-           PETSC_COMM_WORLD, ierr)
-    end if
-    call VecRestoreArrayF90(xvec, xptr, ierr)
-
-    ! nullify pointers and deallocate local variables
-    if (associated(xptr)) nullify(xptr)
-    if (allocated(mybuf)) deallocate(mybuf)
-
-  end subroutine extract_results
-
-!===============================================================================
-! FINALIZE 
-!===============================================================================
-
-  subroutine finalize()
-
-    ! finalize data objects
-    call destroy_M_operator(ctx%loss)
-    call destroy_F_operator(ctx%prod)
-    call destroy_J_operator(jac_prec)
-    call VecDestroy(xvec, ierr)
-    call VecDestroy(resvec, ierr)
-    call MatDestroy(jac, ierr)
-
-    ! finalize solver objects
-    call SNESDestroy(snes, ierr)
-
-  end subroutine finalize
-
-# endif
-
-end module cmfd_snes_solver
diff --git a/src/constants.F90 b/src/constants.F90
index 46ae8e9c1a..5587722869 100644
--- a/src/constants.F90
+++ b/src/constants.F90
@@ -11,7 +11,7 @@ module constants
   integer, parameter :: VERSION_RELEASE = 2
 
   ! Revision numbers for binary files
-  integer, parameter :: REVISION_STATEPOINT       = 9
+  integer, parameter :: REVISION_STATEPOINT       = 10
   integer, parameter :: REVISION_PARTICLE_RESTART = 1
 
   ! Binary file types
@@ -362,9 +362,8 @@ module constants
   integer, parameter :: UNIT_TALLY    = 12 ! unit # for writing tally file
   integer, parameter :: UNIT_PLOT     = 13 ! unit # for writing plot file
   integer, parameter :: UNIT_XS       = 14 ! unit # for writing xs summary file
-  integer, parameter :: CMFD_BALANCE  = 15 ! unit # for writing cmfd balance file
-  integer, parameter :: UNIT_PARTICLE = 16 ! unit # for writing particle restart
-  integer, parameter :: UNIT_OUTPUT   = 17 ! unit # for writing output
+  integer, parameter :: UNIT_PARTICLE = 15 ! unit # for writing particle restart
+  integer, parameter :: UNIT_OUTPUT   = 16 ! unit # for writing output
 
   !=============================================================================
   ! CMFD CONSTANTS
diff --git a/src/eigenvalue.F90 b/src/eigenvalue.F90
index 4bf7112988..f5375bf3a9 100644
--- a/src/eigenvalue.F90
+++ b/src/eigenvalue.F90
@@ -44,7 +44,7 @@ contains
     if (master) call header("K EIGENVALUE SIMULATION", level=1)
 
     ! Display column titles
-    call print_columns()
+    if(master) call print_columns()
 
     ! Turn on inactive timer
     call time_inactive % start()
diff --git a/src/finalize.F90 b/src/finalize.F90
index 1006403fb3..9d5f5c633b 100644
--- a/src/finalize.F90
+++ b/src/finalize.F90
@@ -1,8 +1,5 @@
 module finalize
 
-# ifdef PETSC
-  use cmfd_output,    only: finalize_cmfd
-# endif
   use global
   use output,         only: print_runtime, print_results, &
                             print_overlap_check, write_tallies
@@ -42,11 +39,11 @@ contains
     end if
 
 #ifdef PETSC
-    ! finalize cmfd
-    if (cmfd_run) call finalize_cmfd()
+    ! Finalize PETSc
+    if (cmfd_run) call PetscFinalize(mpi_err)
 #endif
 
-    ! stop timers and show timing statistics
+    ! Stop timers and show timing statistics
     call time_finalize % stop()
     call time_total % stop()
     if (master .and. (run_mode /= MODE_PLOTTING .and. &
@@ -56,7 +53,7 @@ contains
       if (check_overlaps) call print_overlap_check()
     end if
 
-    ! deallocate arrays
+    ! Deallocate arrays
     call free_memory()
 
 #ifdef HDF5
diff --git a/src/global.F90 b/src/global.F90
index e234b9825e..f12171b257 100644
--- a/src/global.F90
+++ b/src/global.F90
@@ -15,10 +15,6 @@ module global
   use tally_header,     only: TallyObject, TallyMap, TallyResult
   use timer_header,     only: Timer
 
-#ifdef MPI
-  use mpi
-#endif
-
 #ifdef HDF5
   use hdf5_interface,  only: HID_T
 #endif
@@ -300,80 +296,64 @@ module global
   logical :: cmfd_run = .false.
  
   ! Timing objects
-  type(Timer) :: time_cmfd   ! timer for whole cmfd calculation
-  type(Timer) :: time_solver ! timer for solver 
+  type(Timer) :: time_cmfd      ! timer for whole cmfd calculation
+  type(Timer) :: time_cmfdbuild ! timer for matrix build
+  type(Timer) :: time_cmfdsolve ! timer for solver 
 
-  ! Flag for CMFD only
-  logical :: cmfd_only = .false.
-
-  ! Flag for coremap accelerator
+  ! Flag for active core map
   logical :: cmfd_coremap = .false.
 
-  ! number of processors for cmfd
-  integer :: n_procs_cmfd
-
-  ! reset dhats to zero
+  ! Flag to reset dhats to zero
   logical :: dhat_reset = .false.
 
-  ! activate neutronic feedback
+  ! Flag to activate neutronic feedback via source weights
   logical :: cmfd_feedback = .false.
 
-  ! activate auto-balance of tallies (2grp only)
-! logical :: cmfd_balance = .false.
+  ! User-defined tally information
+  integer :: n_cmfd_meshes  = 1 ! # of structured meshes
+  integer :: n_cmfd_tallies = 3 ! # of user-defined tallies
 
-  ! calculate effective downscatter
-! logical :: cmfd_downscatter = .false.
-
-  ! user-defined tally information
-  integer :: n_cmfd_meshes              = 1 ! # of structured meshes
-  integer :: n_cmfd_tallies             = 3 ! # of user-defined tallies
-
-  ! overwrite with 2grp xs
-  logical :: cmfd_run_2grp = .false.
-
-  ! hold cmfd weight adjustment factors
+  ! Flag to hold cmfd weight adjustment factors
   logical :: cmfd_hold_weights = .false.
 
-  ! eigenvalue solver type
+  ! Eigenvalue solver type
   character(len=10) :: cmfd_solver_type = 'power'
 
-  ! adjoint method type
+  ! Adjoint method type
   character(len=10) :: cmfd_adjoint_type = 'physical'
 
-  ! number of incomplete ilu factorization levels
+  ! Number of incomplete ilu factorization levels
   integer :: cmfd_ilu_levels = 1
 
-  ! batch to begin cmfd
+  ! Batch to begin cmfd
   integer :: cmfd_begin = 1
 
-  ! when and how long to flush cmfd tallies during inactive batches
+  ! When and how long to flush cmfd tallies during inactive batches
   integer :: cmfd_inact_flush(2) = (/9999,1/)
 
-  ! batch to last flush before active batches
+  ! Batch to last flush before active batches
   integer :: cmfd_act_flush = 0
 
-  ! compute effective downscatter cross section
+  ! Compute effective downscatter cross section
   logical :: cmfd_downscatter = .false.
 
-  ! convergence monitoring
+  ! Convergence monitoring
   logical :: cmfd_snes_monitor  = .false.
   logical :: cmfd_ksp_monitor   = .false.
   logical :: cmfd_power_monitor = .false.
 
-  ! cmfd output
-  logical :: cmfd_write_balance  = .false.
+  ! Cmfd output
   logical :: cmfd_write_matrices = .false.
-  logical :: cmfd_write_hdf5     = .false.
 
-  ! run an adjoint calculation (last batch only)
+  ! Run an adjoint calculation (last batch only)
   logical :: cmfd_run_adjoint = .false.
 
-  ! cmfd run logicals
+  ! CMFD run logicals
   logical :: cmfd_on             = .false.
   logical :: cmfd_tally_on       = .true. 
 
-  ! tolerance on keff to run cmfd
-  real(8) :: cmfd_keff_tol = 0.005_8
+  ! CMFD display info
+  character(len=25) :: cmfd_display
 
   ! Information about state points to be written
   integer :: n_state_points = 0
@@ -463,7 +443,7 @@ contains
     ! Deallocate array of work indices
     if (allocated(work_index)) deallocate(work_index)
 
-    ! Deallocate cmfd
+    ! Deallocate CMFD
     call deallocate_cmfd(cmfd)
 
     ! Deallocate tally node lists
diff --git a/src/input_xml.F90 b/src/input_xml.F90
index c2d1a86d16..f50a4557c5 100644
--- a/src/input_xml.F90
+++ b/src/input_xml.F90
@@ -575,15 +575,13 @@ contains
       end do
     end if
 
-    ! check for cmfd run
+    ! Check for cmfd run
     call lower_case(run_cmfd_)
     if (run_cmfd_ == 'true' .or. run_cmfd_ == '1') then
       cmfd_run = .true.
 #ifndef PETSC
-      if (master) then
-        message = 'CMFD is not available, compile OpenMC with PETSc'
-        call fatal_error()
-      end if
+      message = 'CMFD is not available, recompile OpenMC with PETSc'
+      call fatal_error()
 #endif
     end if
 
diff --git a/src/matrix_header.F90 b/src/matrix_header.F90
new file mode 100644
index 0000000000..591eeaaed1
--- /dev/null
+++ b/src/matrix_header.F90
@@ -0,0 +1,358 @@
+module matrix_header
+
+#ifdef PETSC
+  use petscmat
+#endif
+
+  implicit none
+  private
+
+  type, public :: Matrix
+    integer :: n        ! number of rows/cols in matrix
+    integer :: nnz      ! number of nonzeros in matrix
+    integer :: n_count  ! counter for length of matrix
+    integer :: nz_count ! counter for number of non zeros
+    integer, allocatable :: row(:) ! csr row vector
+    integer, allocatable :: col(:) ! column vector
+    real(8), allocatable :: val(:) ! matrix value vector
+#  ifdef PETSC
+    type(mat) :: petsc_mat
+#  endif
+    logical :: petsc_active
+   contains
+     procedure :: create       => matrix_create
+     procedure :: destroy      => matrix_destroy
+     procedure :: add_value    => matrix_add_value
+     procedure :: new_row      => matrix_new_row
+     procedure :: assemble     => matrix_assemble
+     procedure :: get_row      => matrix_get_row
+     procedure :: get_col      => matrix_get_col
+     procedure :: setup_petsc        => matrix_setup_petsc
+     procedure :: write_petsc_binary => matrix_write_petsc_binary
+     procedure :: transpose          => matrix_transpose
+     procedure :: vector_multiply    => matrix_vector_multiply
+  end type matrix
+
+  integer :: petsc_err
+
+contains
+
+!===============================================================================
+! MATRIX_CREATE allocates CSR vectors
+!===============================================================================
+
+  subroutine matrix_create(self, n, nnz)
+
+    integer, intent(in) :: n             ! dimension of matrix
+    integer, intent(in) :: nnz           ! number of nonzeros
+    class(Matrix), intent(inout) :: self ! matrix instance
+
+    ! Preallocate vectors
+    if (.not.allocated(self % row)) allocate(self % row(n+1))
+    if (.not.allocated(self % col)) allocate(self % col(nnz))
+    if (.not.allocated(self % val)) allocate(self % val(nnz))
+
+    ! Set counters to 1
+    self % n_count  = 1
+    self % nz_count = 1
+
+    ! Set n and nnz
+    self % n = n
+    self % nnz = nnz
+
+    ! Set PETSc active by default to false
+    self % petsc_active = .false.
+
+  end subroutine matrix_create
+
+!===============================================================================
+! MATRIX_DESTROY deallocates all space associated with a matrix
+!===============================================================================
+
+  subroutine matrix_destroy(self)
+
+    class(Matrix), intent(inout) :: self ! matrix instance
+
+#ifdef PETSC
+    if (self % petsc_active) call MatDestroy(self % petsc_mat, petsc_err)
+#endif
+
+    if (allocated(self % row)) deallocate(self % row)
+    if (allocated(self % col)) deallocate(self % col)
+    if (allocated(self % val)) deallocate(self % val)
+
+  end subroutine matrix_destroy
+
+!===============================================================================
+! MATRIX_ADD_VALUE adds a value to the matrix
+!===============================================================================
+
+  subroutine matrix_add_value(self, col, val)
+
+    integer, intent(in) :: col           ! col location in matrix
+    real(8), intent(in) :: val           ! value to store in matrix
+    class(Matrix), intent(inout) :: self ! matrix instance
+
+    ! Record the data
+    self % col(self % nz_count) = col
+    self % val(self % nz_count) = val
+
+    ! Need to adjust column indices if PETSc is active
+    if (self % petsc_active) self % col(self % nz_count) = &
+                             self % col(self % nz_count) - 1
+
+    ! Increment the number of nonzeros currently stored
+    self % nz_count = self % nz_count + 1
+
+  end subroutine matrix_add_value
+
+!===============================================================================
+! MATRIX_NEW_ROW adds a new row by saving the column position of the new row
+!===============================================================================
+
+  subroutine matrix_new_row(self)
+
+    class(Matrix), intent(inout) :: self ! matrix instance
+
+    ! Record the current number of nonzeros
+    self % row(self % n_count) = self % nz_count
+
+    ! If PETSc is active, we have to reference indices off 0
+    if (self % petsc_active) self % row(self % n_count) = &
+                             self % row(self % n_count) - 1
+
+    ! Increment the current row that we are on
+    self % n_count = self % n_count + 1
+
+  end subroutine matrix_new_row
+
+!===============================================================================
+! MATRIX_ASSEMBLE main rountine to sort all the columns in CSR matrix
+!===============================================================================
+
+  subroutine matrix_assemble(self)
+
+    class(Matrix), intent(inout) :: self ! matrix instance
+
+    integer :: i
+    integer :: first
+    integer :: last
+
+    ! Loop around row vector
+    do i = 1, self % n
+
+      ! Get bounds
+      first = self % row(i)
+      last  = self % row(i+1) - 1
+
+      ! Sort a row
+      call sort_csr(self % col, self % val, first, last)
+
+    end do
+
+  end subroutine matrix_assemble
+
+!===============================================================================
+! SORT_CSR main routine that performs a sort on a CSR col/val subvector
+!===============================================================================
+
+  recursive subroutine sort_csr(col, val, first, last)
+
+    integer :: col(:) ! column vector to sort
+    integer :: first  ! first value in sort
+    integer :: last   ! last value in sort
+    real(8) :: val(:) ! value vector to be sorted like columns
+
+    integer :: mid ! midpoint value
+
+    if (first < last) then
+      call split(col, val, first, last, mid)
+      call sort_csr(col, val, first, mid-1)
+      call sort_csr(col, val, mid+1, last)
+    end if
+
+  end subroutine sort_csr
+
+!===============================================================================
+! SPLIT bisects the search space for the sorting routine
+!===============================================================================
+
+  subroutine split(col, val, low, high, mid)
+
+    integer :: col(:) ! column vector to sort
+    integer :: low    ! low index to sort
+    integer :: high   ! high index to sort
+    integer :: mid    ! middle of sort
+    real(8) :: val(:) ! value vector to be sorted like columns
+
+    integer :: left  ! contains left value in sort
+    integer :: right ! contains right value in sort
+    integer :: iswap ! temporary interger swap
+    integer :: pivot ! pivotting variable for columns
+    real(8) :: rswap ! temporary real swap
+    real(8) :: val0  ! pivot for value vector
+
+    left = low
+    right = high
+    pivot = col(low)
+    val0 = val(low)
+
+    ! Repeat the following while left and right havent met
+    do while (left < right)
+
+      ! Scan right to left to find element < pivot
+      do while (left < right .and. col(right) >= pivot)
+        right = right - 1
+      end do
+
+      ! Scan left to right to find element > pivot
+      do while (left < right .and. col(left) <= pivot)
+        left = left + 1
+      end do
+
+      ! If left and right havent met, exchange the items
+      if (left < right) then
+        iswap = col(left)
+        col(left) = col(right)
+        col(right) = iswap
+        rswap = val(left)
+        val(left) = val(right)
+        val(right) = rswap
+      end if
+
+    end do
+
+    ! Switch the element in split position with pivot
+    col(low) = col(right)
+    col(right) = pivot
+    mid = right
+    val(low) = val(right)
+    val(right) = val0
+
+  end subroutine split
+
+!===============================================================================
+! MATRIX_GET_ROW is a method to get row and checks for PETSc C indexing use
+!===============================================================================
+
+  function matrix_get_row(self, i) result(row)
+
+    class(Matrix), intent(in) :: self ! matrix instance
+    integer, intent(in)       :: i    ! row to get
+    integer                   :: row  ! index in col where row begins
+
+    row = self % row(i)
+
+    if (self % petsc_active) row = row + 1
+
+  end function matrix_get_row
+
+!===============================================================================
+! MATRIX_GET_COL is a method to get column and checks for PETSc C index use
+!===============================================================================
+
+  function matrix_get_col(self, i) result(col)
+
+    class(Matrix), intent(in) :: self ! matrix instance
+    integer, intent(in)       :: i    ! index from row vector
+    integer                   :: col  ! the actual column
+
+    col = self % col(i)
+
+    if (self % petsc_active) col = col + 1
+
+  end function matrix_get_col
+
+!===============================================================================
+! MATRIX_SETUP_PETSC configures the row/col vectors and links to a PETSc object
+!===============================================================================
+
+  subroutine matrix_setup_petsc(self)
+
+    class(Matrix), intent(inout) :: self ! matrix instance
+
+    ! change indices to c notation
+    self % row = self % row - 1
+    self % col = self % col - 1
+
+    ! Link to petsc
+#ifdef PETSC
+    call MatCreateSeqAIJWithArrays(PETSC_COMM_WORLD, self % n, self % n, &
+            self % row, self % col, self % val, self % petsc_mat, petsc_err)
+#endif
+
+    ! Petsc is now active
+    self % petsc_active = .true.
+
+  end subroutine matrix_setup_petsc
+
+!===============================================================================
+! MATRIX_WRITE_PETSC_BINARY writes a PETSc matrix binary file
+!===============================================================================
+
+  subroutine matrix_write_petsc_binary(self, filename)
+
+    character(*), intent(in)  :: filename ! file name to write to
+    class(Matrix), intent(in) :: self     ! matrix instance
+
+#ifdef PETSC
+    type(PetscViewer) :: viewer ! a petsc viewer instance
+
+    call PetscViewerBinaryOpen(PETSC_COMM_WORLD, trim(filename), &
+         FILE_MODE_WRITE, viewer, petsc_err)
+    call MatView(self % petsc_mat, viewer, petsc_err)
+    call PetscViewerDestroy(viewer, petsc_err)
+#endif
+
+  end subroutine matrix_write_petsc_binary
+
+!===============================================================================
+! MATRIX_TRANSPOSE uses PETSc to transpose a matrix
+!===============================================================================
+
+  subroutine matrix_transpose(self)
+
+    class(Matrix), intent(inout) :: self ! matrix instance
+
+#ifdef PETSC
+    call MatTranspose(self % petsc_mat, MAT_REUSE_MATRIX, self % petsc_mat, &
+         petsc_err)
+#endif
+
+  end subroutine matrix_transpose
+
+!===============================================================================
+! MATRIX_VECTOR_MULTIPLY allow a vector to multiply the matrix
+!===============================================================================
+
+  subroutine matrix_vector_multiply(self, vec_in, vec_out)
+
+    use constants,      only: ZERO
+    use vector_header,  only: Vector
+
+    class(Matrix), intent(in)   :: self    ! matrix instance
+    type(Vector), intent(in)    :: vec_in  ! vector to multiply matrix against
+    type(Vector), intent(inout) :: vec_out ! resulting vector
+
+    integer :: i ! row iteration counter
+    integer :: j ! column iteration counter
+
+    ! Begin loop around rows
+    ROWS: do i = 1, self % n
+
+      ! Initialize target location in vector
+      vec_out % val(i) = ZERO
+
+      ! Begin loop around columns
+      COLS: do j = self % get_row(i), self % get_row(i + 1) - 1
+
+        vec_out % val(i) = vec_out % val(i) + self % val(j) * &
+                           vec_in % val(self % get_col(j))
+
+      end do COLS
+
+    end do ROWS
+
+  end subroutine matrix_vector_multiply
+
+end module matrix_header
diff --git a/src/output.F90 b/src/output.F90
index 8b54dc54a9..b1290f057e 100644
--- a/src/output.F90
+++ b/src/output.F90
@@ -1223,37 +1223,40 @@ contains
 
   subroutine print_columns()
 
-    if (entropy_on) then
-      if (cmfd_run) then
-        message = " Bat./Gen.      k       Entropy         Average k          CMFD k    CMFD Ent"
-        call write_message(1)
-        message = " =========   ========   ========   ====================   ========   ========"
-        call write_message(1)
-      else
-        message = " Bat./Gen.      k       Entropy         Average k"
-        call write_message(1)
-        message = " =========   ========   ========   ===================="
-        call write_message(1)
-      end if
-    else
-      if (cmfd_run) then
-        message = " Bat./Gen.      k            Average k          CMFD k"
-        call write_message(1)
-        message = " =========   ========   ====================   ========"
-        call write_message(1)
-      else
-        message = " Bat./Gen.      k            Average k"
-        call write_message(1)
-        message = " =========   ========   ===================="
-        call write_message(1)
-      end if
+    write(UNIT=ou, FMT='(2X,A9,3X)', ADVANCE='NO') "Bat./Gen."
+    write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "   k    "
+    if (entropy_on) write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "Entropy "
+    write(UNIT=ou, FMT='(A20,3X)', ADVANCE='NO') "     Average k      "
+    if (cmfd_run) then
+      write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') " CMFD k "
+      select case(trim(cmfd_display))
+        case('entropy')
+          write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "CMFD Ent"
+        case('balance')
+          write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "RMS Bal "
+        case('source')
+          write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "RMS Src "
+        case('dominance')
+          write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "Dom Rat "
+      end select
     end if
+    write(UNIT=ou, FMT=*)
+    
+    write(UNIT=ou, FMT='(2X,A9,3X)', ADVANCE='NO') "========="
+    write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "========"
+    if (entropy_on) write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "========"
+    write(UNIT=ou, FMT='(A20,3X)', ADVANCE='NO') "===================="
+    if (cmfd_run) then
+      write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "========"
+      if (cmfd_display /= '') &
+        write(UNIT=ou, FMT='(A8,3X)', ADVANCE='NO') "========"
+    end if
+    write(UNIT=ou, FMT=*)
 
   end subroutine print_columns
 
 !===============================================================================
-! PRINT_GENERATION displays information for a generation of neutrons. For now,
-! if the user has entropy on, it will print out the entropy
+! PRINT_GENERATION displays information for a generation of neutrons.
 !===============================================================================
 
   subroutine print_generation()
@@ -1303,14 +1306,25 @@ contains
       write(UNIT=OUTPUT_UNIT, FMT='(23X)', ADVANCE='NO')
     end if
 
-
-    ! write out cmfd keff if it is active
-    if (cmfd_on) write(UNIT=OUTPUT_UNIT, FMT='(3X, F8.5)', ADVANCE='NO') &
-         cmfd % keff 
-
-    ! write out cmfd entopy
-    if (cmfd_on .and. entropy_on) write(UNIT=OUTPUT_UNIT, &
-         FMT='(3X, F8.5)', ADVANCE='NO') cmfd % entropy
+    ! write out cmfd keff if it is active and other display info
+    if (cmfd_on) then
+      write(UNIT=OUTPUT_UNIT, FMT='(3X, F8.5)', ADVANCE='NO') &
+         cmfd % k_cmfd(current_batch) 
+      select case(trim(cmfd_display))
+        case('entropy')
+          write(UNIT=OUTPUT_UNIT, FMT='(3X, F8.5)', ADVANCE='NO') &
+            cmfd % entropy(current_batch)
+        case('balance')
+          write(UNIT=OUTPUT_UNIT, FMT='(3X, F8.5)', ADVANCE='NO') &
+            cmfd % balance(current_batch)
+        case('source')
+          write(UNIT=OUTPUT_UNIT, FMT='(3X, F8.5)', ADVANCE='NO') &
+            cmfd % src_cmp(current_batch)
+        case('dominance')
+          write(UNIT=OUTPUT_UNIT, FMT='(3X, F8.5)', ADVANCE='NO') &
+            cmfd % dom(current_batch)
+      end select
+    end if
 
     ! next line
     write(UNIT=OUTPUT_UNIT, FMT=*)
@@ -1405,13 +1419,17 @@ contains
     write(ou,100) "Total time in simulation", time_inactive % elapsed + &
          time_active % elapsed
     write(ou,100) "  Time in transport only", time_transport % elapsed
-    if(cmfd_run) write(ou,100) "Total CMFD time", time_cmfd % elapsed
     write(ou,100) "  Time in inactive batches", time_inactive % elapsed
     write(ou,100) "  Time in active batches", time_active % elapsed
     write(ou,100) "  Time synchronizing fission bank", time_bank % elapsed
     write(ou,100) "    Sampling source sites", time_bank_sample % elapsed
     write(ou,100) "    SEND/RECV source sites", time_bank_sendrecv % elapsed
     write(ou,100) "  Time accumulating tallies", time_tallies % elapsed
+    if (cmfd_run) write(ou,100) "  Time in CMFD", time_cmfd % elapsed
+    if (cmfd_run) write(ou,100) "    Building matrices", &
+                  time_cmfdbuild % elapsed
+    if (cmfd_run) write(ou,100) "    Solving matrices", &
+                  time_cmfdsolve % elapsed
     write(ou,100) "Total time for finalization", time_finalize % elapsed
     write(ou,100) "Total time elapsed", time_total % elapsed
 
diff --git a/src/solver_interface.F90 b/src/solver_interface.F90
new file mode 100644
index 0000000000..b629a0789f
--- /dev/null
+++ b/src/solver_interface.F90
@@ -0,0 +1,294 @@
+module solver_interface
+
+  use error,          only: fatal_error
+  use global,         only: message
+  use matrix_header,  only: Matrix
+  use vector_header,  only: Vector
+
+#ifdef PETSC
+  use petscksp
+  use petscsnes
+#endif
+
+  implicit none
+  private
+
+  ! GMRES solver type 
+  type, public :: GMRESSolver 
+#ifdef PETSC
+    type(ksp) :: ksp_ ! Krylov linear solver instance
+    type(pc)  :: pc_  ! Preconditioner instance
+#endif
+   contains
+#ifdef PETSC
+     procedure :: create       => petsc_gmres_create
+     procedure :: set_oper     => petsc_gmres_set_oper
+     procedure :: destroy      => petsc_gmres_destroy
+     procedure :: solve        => petsc_gmres_solve
+#endif
+  end type GMRESSolver
+
+  ! Derived type to contain list of data needed to be passed to procedures
+  type, public :: Jfnk_ctx
+    procedure (res_interface), pointer, nopass :: res_proc_ptr
+    procedure (jac_interface), pointer, nopass :: jac_proc_ptr
+  end type Jfnk_ctx
+
+  ! JFNK solver type 
+  type, public :: JFNKSolver 
+#ifdef PETSC
+    type(ksp)  :: ksp_   ! Krylov linear solver instance
+    type(pc)   :: pc_    ! Preconditioner instance
+    type(snes) :: snes_  ! Nonlinear solver instance
+    type(mat)  :: jac_mf ! Matrix free jacobian instance
+    integer    :: ls     ! Line search instance
+#endif
+   contains
+#ifdef PETSC
+     procedure :: create        => petsc_jfnk_create
+     procedure :: destroy       => petsc_jfnk_destroy
+     procedure :: set_functions => petsc_jfnk_set_functions
+     procedure :: solve         => petsc_jfnk_solve
+#endif
+  end type JFNKSolver
+
+  ! Abstract interface stating how jacobian and residual routines look
+  abstract interface
+    subroutine res_interface(x, res)
+      import :: Vector
+      type(Vector), intent(in)    :: x   ! solution vector
+      type(Vector), intent(inout) :: res ! residual vector
+    end subroutine res_interface
+
+    subroutine jac_interface(x)
+      import :: Vector
+      type(Vector), intent(in) :: x ! solution vector
+    end subroutine jac_interface
+  end interface
+
+#ifdef PETSC
+  integer :: petsc_err ! petsc error code
+#endif
+
+contains
+
+#ifdef PETSC
+!===============================================================================
+! PETSC_GMRES_CREATE sets up a PETSc GMRES solver
+!===============================================================================
+
+  subroutine petsc_gmres_create(self)
+
+    class(GMRESSolver), intent(inout) :: self ! GMRES solver instance
+
+    integer :: ilu_levels = 5
+    real(8) :: rtol = 1.0e-10_8
+    real(8) :: atol = 1.0e-10_8
+
+    call KSPCreate(PETSC_COMM_WORLD, self % ksp_, petsc_err)
+    call KSPSetTolerances(self % ksp_, rtol, atol, &
+         PETSC_DEFAULT_DOUBLE_PRECISION, PETSC_DEFAULT_INTEGER, petsc_err)
+    call KSPSetType(self % ksp_, 'gmres', petsc_err)
+    call KSPSetInitialGuessNonzero(self % ksp_, PETSC_TRUE, petsc_err)
+    call KSPGetPC(self % ksp_, self % pc_, petsc_err)
+    call PCFactorSetLevels(self % pc_, ilu_levels, petsc_err)
+    call KSPSetFromOptions(self % ksp_, petsc_err)
+
+  end subroutine petsc_gmres_create
+
+!===============================================================================
+! PETSC_GMRES_SET_OPER sets the matrix opetors for the GMRES solver
+!===============================================================================
+
+  subroutine petsc_gmres_set_oper(self, prec_mat, mat_in)
+
+    class(GMRESSolver), intent(inout) :: self     ! GMRES solver instanace
+    type(Matrix), intent(inout)       :: prec_mat ! preconditioner matrix
+    type(Matrix), intent(inout)       :: mat_in   ! coefficient matrix
+
+    call KSPSetOperators(self % ksp_, mat_in % petsc_mat, prec_mat % petsc_mat, &
+         SAME_NONZERO_PATTERN, petsc_err)
+    call KSPSetUp(self % ksp_, petsc_err) 
+
+  end subroutine petsc_gmres_set_oper
+
+!===============================================================================
+! PETSC_GMRES_DESTROY frees all memory associated with the GMRES solver
+!===============================================================================
+
+  subroutine petsc_gmres_destroy(self)
+
+   class(GMRESSolver), intent(inout) :: self ! GMRES solver instance
+
+    call KSPDestroy(self % ksp_, petsc_err)
+
+  end subroutine petsc_gmres_destroy
+
+!===============================================================================
+! PETSC_GMRES_SOLVE solves the linear system
+!===============================================================================
+
+  subroutine petsc_gmres_solve(self, b, x)
+
+    class(GMRESSolver), intent(inout) :: self ! GMRES solver instance
+    type(Vector), intent(inout)       :: b    ! right hand side vector
+    type(Vector), intent(inout)       :: x    ! solution vector
+
+    call KSPSolve(self % ksp_, b % petsc_vec, x % petsc_vec, petsc_err)
+
+  end subroutine petsc_gmres_solve
+
+!===============================================================================
+! PETSC_JFNK_CREATE sets up a JFNK solver using PETSc SNES
+!===============================================================================
+
+  subroutine petsc_jfnk_create(self)
+
+    class(JFNKSolver), intent(inout) :: self ! JFNK solver instance
+
+    ! Turn on mf_operator option for matrix free jacobian
+    call PetscOptionsSetValue("-snes_mf_operator", "TRUE", petsc_err)
+
+    ! Create the SNES context
+    call SNESCreate(PETSC_COMM_WORLD, self % snes_, petsc_err)
+
+    ! Set up the GMRES solver
+    call SNESGetKSP(self % snes_, self % ksp_, petsc_err)
+    call KSPSetType(self % ksp_, 'gmres', petsc_err)    
+
+    ! Apply options
+    call SNESGetLineSearch(self % snes_, self % ls, petsc_err)
+    call SNESLineSearchSetType(self % ls, 'basic', petsc_err)
+    call SNESSetFromOptions(self % snes_, petsc_err)
+
+    ! Set up preconditioner
+    call KSPGetPC(self % ksp_, self % pc_, petsc_err)
+    call PCSetType(self % pc_, 'ilu', petsc_err)
+    call PCFactorSetLevels(self % pc_, 5, petsc_err)
+    call PCSetFromOptions(self % pc_, petsc_err)
+    call KSPSetFromOptions(self % ksp_, petsc_err)
+
+  end subroutine petsc_jfnk_create
+
+!===============================================================================
+! PETSC_JFNK_DESTROY frees all memory associated with JFNK solver
+!===============================================================================
+
+  subroutine petsc_jfnk_destroy(self)
+
+    class(JFNKSolver), intent(inout) :: self ! JFNK solver instance
+
+    call MatDestroy(self % jac_mf, petsc_err)
+    call SNESDestroy(self % snes_, petsc_err)
+
+  end subroutine petsc_jfnk_destroy
+
+!===============================================================================
+! PETSC_JFNK_SET_FUNCTIONS sets user functions and matrix free objects
+!===============================================================================
+
+  subroutine petsc_jfnk_set_functions(self, ctx, res, jac_prec)
+
+    class(JFNKSolver), intent(inout) :: self     ! JFNK solver instance
+    type(Jfnk_ctx),    intent(inout) :: ctx      ! JFNK context instance
+    type(Vector),      intent(inout) :: res      ! residual vector
+    type(Matrix),      intent(inout) :: jac_prec ! preconditioner matrix
+
+    ! Set residual procedure
+    call SNESSetFunction(self % snes_, res % petsc_vec, &
+         petsc_jfnk_compute_residual, ctx, petsc_err)
+
+    ! Create the matrix free jacobian
+    call MatCreateSNESMF(self % snes_, self % jac_mf, petsc_err)
+
+    ! Set Jacobian procedure
+    call SNESSetJacobian(self % snes_, self % jac_mf, jac_prec % petsc_mat, &
+         petsc_jfnk_compute_jacobian, ctx, petsc_err)
+
+    ! Set up Jacobian Lags
+    call SNESSetLagJacobian(self % snes_, -2, petsc_err)
+    call SNESSetLagPreconditioner(self % snes_, -1, petsc_err)
+
+  end subroutine petsc_jfnk_set_functions
+
+!===============================================================================
+! PETSC_JFNK_SOLVE solves the nonlinear system
+!===============================================================================
+
+  subroutine petsc_jfnk_solve(self, xvec)
+
+    class(JFNKSolver), intent(inout) :: self ! JFNK instance
+    type(Vector), intent(inout)      :: xvec ! solution vector
+
+    call SNESSolve(self % snes_, PETSC_NULL_DOUBLE, xvec % petsc_vec, petsc_err)
+
+  end subroutine petsc_jfnk_solve
+
+!===============================================================================
+! PETSC_JFNK_COMPUTE_RESIDUAL buffer routine to user specifed residual routine
+!===============================================================================
+
+  subroutine petsc_jfnk_compute_residual(snes_, x, res, ctx, ierr)
+
+    type(snes), intent(inout)     :: snes_ ! PETSc SNES object
+    type(vec), intent(inout)      :: x     ! PETSc solution vector
+    type(vec), intent(inout)      :: res   ! PETSc residual vector
+    integer, intent(inout)        :: ierr  ! error code
+    type(Jfnk_ctx), intent(inout) :: ctx ! JFNK context instance
+
+    type(Vector) :: xvec   ! solution vector
+    type(Vector) :: resvec ! residual vector
+
+    ! We need to use the x and res that come from PETSc because the pointer
+    ! location changes and therefore we can not use module variables
+    ! for the residual and x vectors here
+
+    ! Need to point an OpenMC vector to PETSc vector
+    call VecGetArrayF90(x, xvec % val, ierr)
+    call VecGetArrayF90(res, resvec % val, ierr)
+
+    ! Call user residual routine
+    call ctx % res_proc_ptr(xvec, resvec)
+
+    ! Need to restore the PETSc vector
+    call VecRestoreArrayF90(x, xvec % val, ierr)
+    call VecRestoreArrayF90(res, resvec % val, ierr)
+
+  end subroutine petsc_jfnk_compute_residual
+
+!===============================================================================
+! PETSC_JFNK_COMPUTE_JACOBIAN buffer routine to user specified jacobian routine
+!===============================================================================
+
+  subroutine petsc_jfnk_compute_jacobian(snes_, x, jac_mf, jac_prec, flag, &
+             ctx, ierr)
+
+    type(snes), intent(inout)     :: snes_     ! PETSc snes instance
+    type(vec), intent(inout)      :: x         ! PETSc solution vector
+    type(mat), intent(inout)      :: jac_mf    ! PETSc matrix free jacobian
+    type(mat), intent(inout)      :: jac_prec  ! PETSc matrix jacobian precond.
+    integer, intent(inout)        :: flag      ! unused madatory flag
+    type(Jfnk_ctx), intent(inout) :: ctx       ! JFNK context instance
+    integer, intent(inout)        :: ierr      ! error code
+
+    type(Vector) :: xvec ! solution vector
+
+    ! Again, we use the vector that comes from Petsc to build the Jacobian
+    ! matrix
+
+    ! Need to point OpenMC vector to PETSc Vector
+    call VecGetArrayF90(x, xvec % val, ierr)
+
+    ! Evaluate user jacobian routine
+    call ctx % jac_proc_ptr(xvec)
+
+    ! Restore the PETSc vector
+    call VecRestoreArrayF90(x, xvec % val, ierr)
+
+    call MatAssemblyBegin(jac_mf, MAT_FINAL_ASSEMBLY, petsc_err)
+    call MatAssemblyEnd(jac_mf, MAT_FINAL_ASSEMBLY, petsc_err)
+
+  end subroutine petsc_jfnk_compute_jacobian
+#endif
+
+end module solver_interface
diff --git a/src/state_point.F90 b/src/state_point.F90
index dfa88674cc..9ad9b8d9f2 100644
--- a/src/state_point.F90
+++ b/src/state_point.F90
@@ -20,6 +20,10 @@ module state_point
   use output_interface
   use tally_header,       only: TallyObject
 
+#ifdef MPI
+  use mpi
+#endif
+
   implicit none
 
   type(BinaryOutput) :: sp ! statepoint/source output file
@@ -96,6 +100,28 @@ contains
         call sp % write_data(k_col_tra, "k_col_tra")
         call sp % write_data(k_abs_tra, "k_abs_tra")
         call sp % write_data(k_combined, "k_combined", length=2)
+
+        ! Write out CMFD info
+        if (cmfd_on) then
+          call sp % write_data(1, "cmfd_on")
+          call sp % write_data(cmfd % indices, "indicies", length=4, group="cmfd")
+          call sp % write_data(cmfd % k_cmfd, "k_cmfd", length=current_batch, &
+               group="cmfd")
+          call sp % write_data(cmfd % cmfd_src, "cmfd_src", &
+               length=(/cmfd % indices(4), cmfd % indices(1), &
+               cmfd % indices(2), cmfd % indices(3)/), &
+               group="cmfd")
+          call sp % write_data(cmfd % entropy, "cmfd_entropy", &
+                          length=current_batch, group="cmfd")
+          call sp % write_data(cmfd % balance, "cmfd_balance", &
+               length=current_batch, group="cmfd")
+          call sp % write_data(cmfd % dom, "cmfd_dominance", &
+               length = current_batch, group="cmfd")
+          call sp % write_data(cmfd % src_cmp, "cmfd_srccmp", &
+               length = current_batch, group="cmfd")
+        else
+          call sp % write_data(0, "cmfd_on")
+        end if
       end if
 
       ! Write number of meshes
@@ -507,6 +533,28 @@ contains
 
       ! Take maximum of statepoint n_inactive and input n_inactive
       n_inactive = max(n_inactive, int_array(1))
+
+      ! Read in to see if CMFD was on
+      call sp % read_data(int_array(1), "cmfd_on")
+
+      ! Write out CMFD info
+      if (int_array(1) == 1) then
+        call sp % read_data(cmfd % indices, "indicies", length=4, group="cmfd")
+        call sp % read_data(cmfd % k_cmfd, "k_cmfd", length=restart_batch, &
+             group="cmfd")
+        call sp % read_data(cmfd % cmfd_src, "cmfd_src", &
+             length=(/cmfd % indices(4), cmfd % indices(1), &
+             cmfd % indices(2), cmfd % indices(3)/), &
+             group="cmfd")
+        call sp % read_data(cmfd % entropy, "cmfd_entropy", &
+                       length=restart_batch, group="cmfd")
+        call sp % read_data(cmfd % balance, "cmfd_balance", &
+             length=restart_batch, group="cmfd")
+        call sp % read_data(cmfd % dom, "cmfd_dominance", &
+             length = restart_batch, group="cmfd")
+        call sp % read_data(cmfd % src_cmp, "cmfd_srccmp", &
+             length = restart_batch, group="cmfd")
+      end if
     end if
 
     ! Read number of meshes
diff --git a/src/templates/cmfd_t.xml b/src/templates/cmfd_t.xml
index c9ae3c1fdc..0e49e0e5ab 100644
--- a/src/templates/cmfd_t.xml
+++ b/src/templates/cmfd_t.xml
@@ -13,27 +13,22 @@
     <component name="energy"      type="double-array"  />
   </typedef>
 
-  <variable name="mesh_"           tag="mesh"           type="mesh_xml"                                     />
-  <variable name="norm_"           tag="norm"           type="double"        default="1.0"                  />
-  <variable name="feedback_"       tag="feedback"       type="word"          length="5"                     />
-  <variable name="n_cmfd_procs_"   tag="n_procs"        type="integer"       default="1"                    />
-  <variable name="reset_"          tag="reset"          type="word"          length="5"                     />
-  <variable name="balance_"        tag="balance"        type="word"          length="5"                     />
-  <variable name="downscatter_"    tag="downscatter"    type="word"          length="5"                     />
-  <variable name="run_2grp_"       tag="run_2grp"       type="word"          length="5"                     />
-  <variable name="solver_"         tag="solver"         type="word"          default="'power'" length="250" />
-  <variable name="snes_monitor_"   tag="snes_monitor"   type="word"          length="5"                     />
-  <variable name="ksp_monitor_"    tag="ksp_monitor"    type="word"          length="5"                     />
-  <variable name="power_monitor_"  tag="power_monitor"  type="word"          length="5"                     />
-  <variable name="write_balance_"  tag="write_balance"  type="word"          length="5"                     />
-  <variable name="write_matrices_" tag="write_matrices" type="word"          length="5"                     />
-  <variable name="run_adjoint_"    tag="run_adjoint"    type="word"          length="5"                     />
-  <variable name="write_hdf5_"     tag="write_hdf5"     type="word"          length="5"                     />
-  <variable name="begin_"          tag="begin"          type="integer"       default="1"                    />
-  <variable name="inactive_"       tag="inactive"       type="word"          length="5"                     />
-  <variable name="active_flush_"   tag="active_flush"   type="integer"       default="0"                    />
-  <variable name="keff_tol_"       tag="keff_tol"       type="double"        default="0.005"                />
-  <variable name="inactive_flush_" tag="inactive_flush" type="integer"       default="9999"                 />
-  <variable name="num_flushes_"    tag="num_flushes"    type="integer"       default="1"                    />
+  <variable name="mesh_"           tag="mesh"           type="mesh_xml"                                    />
+  <variable name="norm_"           tag="norm"           type="double"        default="1.0"                 />
+  <variable name="feedback_"       tag="feedback"       type="word"          length="5"                    />
+  <variable name="reset_"          tag="reset"          type="word"          length="5"                    />
+  <variable name="downscatter_"    tag="downscatter"    type="word"          length="5"                    />
+  <variable name="solver_"         tag="solver"         type="word"          default="'power'" length="25" />
+  <variable name="snes_monitor_"   tag="snes_monitor"   type="word"          length="5"                    />
+  <variable name="ksp_monitor_"    tag="ksp_monitor"    type="word"          length="5"                    />
+  <variable name="power_monitor_"  tag="power_monitor"  type="word"          length="5"                    />
+  <variable name="write_matrices_" tag="write_matrices" type="word"          length="5"                    />
+  <variable name="run_adjoint_"    tag="run_adjoint"    type="word"          length="5"                    />
+  <variable name="begin_"          tag="begin"          type="integer"       default="1"                   />
+  <variable name="inactive_"       tag="inactive"       type="word"          length="5"                    />
+  <variable name="active_flush_"   tag="active_flush"   type="integer"       default="0"                   />
+  <variable name="inactive_flush_" tag="inactive_flush" type="integer"       default="9999"                />
+  <variable name="num_flushes_"    tag="num_flushes"    type="integer"       default="1"                   />
+  <variable name="display_"        tag="display"        type="word"          default="''" length="25"      />
 
 </template>
diff --git a/src/utils/build_dependencies.py b/src/utils/build_dependencies.py
index a2fd261f26..19ac978857 100755
--- a/src/utils/build_dependencies.py
+++ b/src/utils/build_dependencies.py
@@ -11,7 +11,8 @@ for src in glob.iglob('*.F90'):
     d = re.findall(r'\n\s*use\s+(\w+)',
                    open(src, 'r').read())
     for name in d:
-        if name in ['mpi', 'hdf5', 'h5lt']:
+        if name in ['mpi', 'hdf5', 'h5lt', 'petscsys', 'petscmat', 'petscksp',
+                    'petscsnes', 'petscvec']:
             continue
         if name.startswith('xml_data_'):
             name = name.replace('xml_data_', 'templates/')
@@ -20,7 +21,7 @@ for src in glob.iglob('*.F90'):
         dependencies[module] = sorted(list(deps))
 
 
-for module in dependencies.keys():
+for module in sorted(dependencies.keys()):
     for dep in dependencies[module]:
         print("{0}.o: {1}.o".format(module, dep))
     print('')
diff --git a/src/utils/statepoint.py b/src/utils/statepoint.py
index df0629a2b2..ec9d9edf78 100644
--- a/src/utils/statepoint.py
+++ b/src/utils/statepoint.py
@@ -152,6 +152,8 @@ class StatePoint(object):
 
         # Read statepoint revision
         self.revision = self._get_int(path='revision')[0]
+        if self.revision != 10:
+          raise Exception('Statepoint Revision is not consistent.')
 
         # Read OpenMC version
         if self._hdf5:
@@ -186,11 +188,29 @@ class StatePoint(object):
                 self.current_batch*self.gen_per_batch, path='k_generation')
             self.entropy = self._get_double(
                 self.current_batch*self.gen_per_batch, path='entropy')
-            if self.revision >= 8:
-                self.k_col_abs = self._get_double(path='k_col_abs')[0]
-                self.k_col_tra = self._get_double(path='k_col_tra')[0]
-                self.k_abs_tra = self._get_double(path='k_abs_tra')[0]
-                self.k_combined = self._get_double(2, path='k_combined')
+            self.k_col_abs = self._get_double(path='k_col_abs')[0]
+            self.k_col_tra = self._get_double(path='k_col_tra')[0]
+            self.k_abs_tra = self._get_double(path='k_abs_tra')[0]
+            self.k_combined = self._get_double(2, path='k_combined')
+
+            # Read CMFD information
+            cmfd_present = self._get_int(path='cmfd_on')[0]
+            if cmfd_present == 1:
+                self.cmfd_indices = self._get_int(4, path='cmfd/indicies')
+                self.k_cmfd = self._get_double(self.current_batch,
+                              path='cmfd/k_cmfd')
+                self.cmfd_src = self._get_double_array(np.product(self.cmfd_indices),
+                                path='cmfd/cmfd_src')
+                self.cmfd_src = np.reshape(self.cmfd_src,
+                                tuple(self.cmfd_indices), order='F')
+                self.cmfd_entropy = self._get_double(self.current_batch,
+                                    path='cmfd/cmfd_entropy')
+                self.cmfd_balance = self._get_double(self.current_batch,
+                                    path='cmfd/cmfd_balance')
+                self.cmfd_dominance = self._get_double(self.current_batch,
+                                      path='cmfd/cmfd_dominance')
+                self.cmfd_srccmp = self._get_double(self.current_batch,
+                                   path='cmfd/cmfd_srccmp')
 
         # Read number of meshes
         n_meshes = self._get_int(path='tallies/n_meshes')[0]
@@ -581,6 +601,12 @@ class StatePoint(object):
         else:
             return [float(v) for v in self._get_data(n, 'd', 8)]
 
+    def _get_double_array(self, n=1, path=None):
+        if self._hdf5:
+            return self._f[path].value
+        else:
+            return self._get_data(n, 'd', 8)
+
     def _get_string(self, n=1, path=None):
         if self._hdf5:
             return str(self._f[path].value)
diff --git a/src/vector_header.F90 b/src/vector_header.F90
new file mode 100644
index 0000000000..3d1bd9afae
--- /dev/null
+++ b/src/vector_header.F90
@@ -0,0 +1,126 @@
+module vector_header
+
+  use constants,  only: ZERO
+
+#ifdef PETSC
+  use petscvec
+#endif
+
+  implicit none
+  private
+
+  type, public :: Vector 
+    integer :: n        ! number of rows/cols in matrix
+    real(8), allocatable :: data(:) ! where vector data is stored
+    real(8), pointer :: val(:) ! pointer to vector data
+#  ifdef PETSC
+    type(vec) :: petsc_vec ! PETSc vector
+#  endif
+    logical :: petsc_active ! Logical if PETSc is being used
+   contains
+     procedure :: create       => vector_create
+     procedure :: destroy      => vector_destroy
+     procedure :: add_value    => vector_add_value
+     procedure :: setup_petsc  => vector_setup_petsc
+     procedure :: write_petsc_binary => vector_write_petsc_binary
+  end type Vector
+
+  integer :: petsc_err ! petsc error code
+
+contains
+
+!===============================================================================
+! VECTOR_CREATE allocates and initializes a vector 
+!===============================================================================
+
+  subroutine vector_create(self, n)
+
+    integer, intent(in)                    :: n    ! size of vector
+    class(Vector), intent(inout), target   :: self ! vector instance
+
+    ! Preallocate vector
+    if (.not.allocated(self % data)) allocate(self % data(n))
+    self % val => self % data(1:n)
+
+    ! Set n
+    self % n = n
+
+    ! Initialize to zero
+    self % val = ZERO
+
+    ! Petsc is default not active
+    self % petsc_active = .false.
+
+  end subroutine vector_create
+
+!===============================================================================
+! VECTOR_DESTROY deallocates all space associated with a vector
+!===============================================================================
+
+  subroutine vector_destroy(self)
+
+    class(Vector), intent(inout) :: self ! vector instance
+
+#ifdef PETSC
+    if (self % petsc_active) call VecDestroy(self % petsc_vec, petsc_err)
+#endif
+
+    if (associated(self % val)) nullify(self % val)
+    if (allocated(self % data)) deallocate(self % data)
+
+  end subroutine vector_destroy
+
+!===============================================================================
+! VECTOR_ADD_VALUE adds a value to the vector
+!===============================================================================
+
+  subroutine vector_add_value(self, idx, val)
+
+    integer, intent(in)          :: idx  ! index location in vector
+    real(8), intent(in)          :: val  ! value to add
+    class(Vector), intent(inout) :: self ! vector instance
+
+    self % val(idx) = val
+
+  end subroutine vector_add_value
+
+!===============================================================================
+! VECTOR_SETUP_PETSC links the data to a PETSc vector
+!===============================================================================
+
+  subroutine vector_setup_petsc(self)
+
+    class(Vector), intent(inout) :: self ! vector instance
+
+    ! Link to PETSc 
+#ifdef PETSC
+    call VecCreateSeqWithArray(PETSC_COMM_WORLD, 1, self % n, self % val, &
+         self % petsc_vec, petsc_err) 
+#endif
+
+    ! Set that PETSc is now active
+    self % petsc_active = .true.
+
+  end subroutine vector_setup_petsc
+
+!===============================================================================
+! VECTOR_WRITE_PETSC_BINARY writes the PETSc vector to a binary file
+!===============================================================================
+
+  subroutine vector_write_petsc_binary(self, filename)
+
+    character(*), intent(in)  :: filename ! name of file to write to
+    class(Vector), intent(in) :: self     ! vector instance
+
+#ifdef PETSC
+    type(PetscViewer) :: viewer ! PETSc viewer instance
+
+    call PetscViewerBinaryOpen(PETSC_COMM_WORLD, trim(filename), &
+         FILE_MODE_WRITE, viewer, petsc_err)
+    call VecView(self % petsc_vec, viewer, petsc_err)
+    call PetscViewerDestroy(viewer, petsc_err)
+#endif
+
+  end subroutine vector_write_petsc_binary
+
+end module vector_header
diff --git a/tests/run_tests.py b/tests/run_tests.py
index 1a1a05e839..49975c218a 100755
--- a/tests/run_tests.py
+++ b/tests/run_tests.py
@@ -22,7 +22,7 @@ def run_compile():
         os.remove(exe)
 
     # run compile test
-    result = nose.run(argv=['run_tests.py', 'test_compile'] + flags)
+    result = nose.run(argv=['nosetests', 'test_compile'] + flags)
     if not result:
         print('Did not pass compile tests.')
     results.append(('compile', result))
@@ -35,7 +35,7 @@ def run_suite(name=None, mpi=False):
 
     # Set arguments list. Note that the first argument is a dummy argument (the
     # script name). It's not actually recursively calling run_tests.py
-    argv = ['run_tests.py', '--exclude', 'test_compile'] + flags
+    argv = ['nosetests', '--exclude', 'test_compile'] + flags
 
     # Add MPI plugin if set
     if mpi:
@@ -48,11 +48,13 @@ def run_suite(name=None, mpi=False):
         os.chdir(pwd)
         os.rename(pwd + '/../src/openmc-' + name, pwd + '/../src/openmc')
         result = nose.run(argv=argv, addplugins=plugins)
-    finally:
         os.rename(pwd + '/../src/openmc', pwd + '/../src/openmc-' + name)
-        if not result:
-            print('Did not pass ' + name + ' tests')
-        results.append((name, result))
+    except OSError:
+        result = False
+        print('No OpenMC executable found for ' + name + ' tests')
+    if not result:
+        print('Did not pass ' + name + ' tests')
+    results.append((name, result))
 
 # set mpiexec path
 if 'COMPILER' in os.environ:
@@ -67,23 +69,77 @@ sys.path.append(pwd)
 
 # Set list of tests, either default or from command line
 flags = []
-tests = ['compile', 'normal', 'debug', 'optimize', 'mpi', 'omp', 'hdf5',
-         'petsc', 'mpi-omp', 'omp-hdf5', 'phdf5', 'phdf5-petsc',
-         'phdf5-petsc-optimize', 'omp-phdf5-petsc-optimize']
+tests = ['compile', 'normal', 'debug', 'optimize',
+         'omp', 'omp-debug', 'omp-optimize',
+         'hdf5', 'hdf5-debug', 'hdf5-optimize',
+         'omp-hdf5', 'omp-hdf5-debug', 'omp-hdf5-optimize',
+         'mpi', 'mpi-debug', 'mpi-optimize',
+         'mpi-omp', 'mpi-omp-debug', 'mpi-omp-optimize',
+         'phdf5', 'phdf5-debug', 'phdf5-optimize',
+         'phdf5-omp', 'phdf5-omp-debug', 'phdf5-omp-optimize',
+         'petsc', 'petsc-debug', 'petsc-optimize',
+         'phdf5-petsc', 'phdf5-petsc-debug', 'phdf5-petsc-optimize',
+         'omp-phdf5-petsc', 'omp-phdf5-petsc-debug',
+         'omp-phdf5-petsc-optimize']
 if len(sys.argv) > 1:
     flags = [i for i in sys.argv[1:] if i.startswith('-')]
     tests_ = [i for i in sys.argv[1:] if not i.startswith('-')]
-    tests = tests_ if tests_ else tests
+
+    # Check for special subsets of tests
+    tests__ = []
+    for i in tests_:
+
+        # All tests will run all the tests except for compile unless
+        # it is also specified on the command line. Note that specifying
+        # compile all-tests is the same as not specifying any args
+        if i == 'all-tests':
+            tests__ = tests
+            try:
+                idx = tests_.index('compile')  # check for compile test
+            except ValueError:
+                del tests__[0]
+            finally:
+                break  # don't need to check for anything else
+
+        # This checks for any subsets of tests. The string after
+        # all-XXXX will be used to search through all tests.
+        # Specifying XXXX=normal will run tests that don't contain
+        # debug or optimize substring.
+        if i.startswith('all-'):
+            suffix = i.split('all-')[1]
+            if suffix == 'normal':
+                for j in tests:
+                    if j.rfind('debug') == -1 and \
+                       j.rfind('optimize') == -1:
+                        tests__.append(j)
+            else:
+                for j in tests:
+                    if j.rfind(suffix) != -1:
+                        if suffix == 'omp' and j == 'compile':
+                            continue
+                        tests__.append(j)
+        else:
+            tests__.append(i)  # append specific test (e.g., mpi-debug)
+    tests = tests__ if tests__ else tests
 
 # Run tests
 results = []
 for name in tests:
     if name == 'compile':
         run_compile()
-    elif name in ['normal', 'debug', 'optimize', 'omp', 'hdf5', 'omp-hdf5']:
+    elif name in ['normal', 'debug', 'optimize',
+                  'hdf5', 'hdf5-debug', 'hdf5-optimize',
+                  'omp', 'omp-debug', 'omp-optimize',
+                  'omp-hdf5', 'omp-hdf5-debug', 'omp-hdf5-optimize']:
         run_suite(name=name)
-    elif name in ['mpi', 'mpi-omp' 'phdf5', 'petsc', 'phdf5-petsc',
-                  'phdf5-petsc-optimize', 'omp-phdf5-petsc-optimize']:
+    elif name in ['mpi', 'mpi-debug', 'mpi-optimize',
+                  'mpi-omp', 'mpi-omp-debug', 'mpi-omp-optimize',
+                  'phdf5', 'phdf5-debug', 'phdf5-optimize',
+                  'phdf5-omp', 'phdf5-omp-debug', 'phdf5-omp-optimize',
+                  'petsc', 'petsc-debug', 'petsc-optimize',
+                  'phdf5-petsc', 'phdf5-petsc-debug', 'phdf5-petsc-optimize',
+                  'omp-phdf5-petsc', 'omp-phdf5-petsc-debug',
+                  'omp-phdf5-petsc-optimize']:
         run_suite(name=name, mpi=True)
 
 # print out summary of results
diff --git a/tests/test_basic/test_basic.py b/tests/test_basic/test_basic.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_basic/test_basic.py
+++ b/tests/test_basic/test_basic.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_cmfd_feed/results.py b/tests/test_cmfd_feed/results.py
index 533a624558..26bcb36d9f 100644
--- a/tests/test_cmfd_feed/results.py
+++ b/tests/test_cmfd_feed/results.py
@@ -34,7 +34,7 @@ results4 = np.reshape(results4, size4)
 
 # set up output string
 outstr = ''
- 
+
 # write out k-combined
 outstr += 'k-combined:\n'
 outstr += "{0:12.6E} {1:12.6E}\n".format(sp.k_combined[0], sp.k_combined[1])
@@ -42,17 +42,41 @@ outstr += "{0:12.6E} {1:12.6E}\n".format(sp.k_combined[0], sp.k_combined[1])
 # write out tally results
 outstr += 'tally 1:\n'
 for item in results1:
-  outstr += "{0:12.6E}\n".format(item)
+    outstr += "{0:12.6E}\n".format(item)
 outstr += 'tally 2:\n'
 for item in results2:
-  outstr += "{0:12.6E}\n".format(item)
+    outstr += "{0:12.6E}\n".format(item)
 outstr += 'tally 3:\n'
 for item in results3:
-  outstr += "{0:12.6E}\n".format(item)
+    outstr += "{0:12.6E}\n".format(item)
 outstr += 'tally 4:\n'
 for item in results4:
-  outstr += "{0:12.6E}\n".format(item)
+    outstr += "{0:12.6E}\n".format(item)
+
+# write out cmfd answers
+outstr += 'cmfd indices\n'
+for item in sp.cmfd_indices:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'k cmfd\n'
+for item in sp.k_cmfd:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd entropy\n'
+for item in sp.cmfd_entropy:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd balance\n'
+for item in sp.cmfd_balance:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd dominance ratio\n'
+for item in sp.cmfd_dominance:
+    outstr += "{0:10.3E}\n".format(item)
+outstr += 'cmfd openmc source comparison\n'
+for item in sp.cmfd_srccmp:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd source\n'
+cmfdsrc = np.reshape(sp.cmfd_src, np.product(sp.cmfd_indices), order='F')
+for item in cmfdsrc:
+    outstr += "{0:12.6E}\n".format(item)
 
 # write results to file
-with open('results_test.dat','w') as fh:
+with open('results_test.dat', 'w') as fh:
     fh.write(outstr)
diff --git a/tests/test_cmfd_feed/results_true.dat b/tests/test_cmfd_feed/results_true.dat
index ad4c85c761..0e46fd1371 100644
--- a/tests/test_cmfd_feed/results_true.dat
+++ b/tests/test_cmfd_feed/results_true.dat
@@ -1,14 +1,14 @@
 k-combined:
-1.167124E+00 1.217343E-02
+1.167124E+00 1.217344E-02
 tally 1:
 1.126891E+01
-1.275755E+01
-2.086556E+01
-4.369501E+01
+1.275756E+01
+2.086557E+01
+4.369504E+01
 2.853629E+01
-8.169559E+01
-3.404506E+01
-1.165019E+02
+8.169563E+01
+3.404507E+01
+1.165020E+02
 3.723908E+01
 1.389881E+02
 3.760048E+01
@@ -16,86 +16,86 @@ tally 1:
 3.455469E+01
 1.197186E+02
 2.837503E+01
-8.086291E+01
-2.151789E+01
-4.648315E+01
+8.086287E+01
+2.151788E+01
+4.648312E+01
 1.194541E+01
-1.432331E+01
+1.432330E+01
 tally 2:
-2.292173E+01
-2.650792E+01
-1.598183E+01
-1.288692E+01
-2.194143E+00
-2.510557E-01
-4.217816E+01
-8.938583E+01
+2.292174E+01
+2.650794E+01
+1.598184E+01
+1.288693E+01
+2.194144E+00
+2.510560E-01
+4.217817E+01
+8.938588E+01
 2.974466E+01
-4.447038E+01
-3.939269E+00
-7.882413E-01
-5.742289E+01
+4.447040E+01
+3.939271E+00
+7.882419E-01
+5.742290E+01
 1.654599E+02
-4.067581E+01
-8.308454E+01
-5.566955E+00
-1.570072E+00
-6.810586E+01
+4.067582E+01
+8.308457E+01
+5.566956E+00
+1.570073E+00
+6.810587E+01
 2.331019E+02
-4.831543E+01
+4.831544E+01
 1.174673E+02
-6.244999E+00
-1.967938E+00
+6.245000E+00
+1.967939E+00
 7.258131E+01
 2.646756E+02
 5.185146E+01
-1.351425E+02
-6.622488E+00
+1.351426E+02
+6.622489E+00
 2.231695E+00
 7.246115E+01
 2.641295E+02
-5.159542E+01
+5.159541E+01
 1.339441E+02
 6.702239E+00
 2.277825E+00
-6.716397E+01
+6.716396E+01
 2.266815E+02
-4.765063E+01
-1.141627E+02
-6.384098E+00
+4.765062E+01
+1.141626E+02
+6.384097E+00
 2.063760E+00
-5.549542E+01
-1.545361E+02
+5.549541E+01
+1.545360E+02
 3.940819E+01
-7.794880E+01
-5.183375E+00
-1.369483E+00
-4.155685E+01
-8.679439E+01
-2.926744E+01
-4.307145E+01
-3.938852E+00
-7.877949E-01
-2.335180E+01
-2.758241E+01
-1.622683E+01
-1.331996E+01
-2.248921E+00
-2.594934E-01
+7.794877E+01
+5.183373E+00
+1.369482E+00
+4.155684E+01
+8.679434E+01
+2.926743E+01
+4.307143E+01
+3.938851E+00
+7.877944E-01
+2.335179E+01
+2.758239E+01
+1.622682E+01
+1.331995E+01
+2.248920E+00
+2.594933E-01
 tally 3:
 1.537807E+01
 1.193746E+01
 1.044189E+00
 5.627797E-02
 2.861614E+01
-4.118322E+01
+4.118324E+01
 1.846208E+00
 1.729007E-01
-3.921462E+01
-7.723576E+01
+3.921463E+01
+7.723579E+01
 2.427723E+00
 2.983153E-01
-4.653207E+01
+4.653208E+01
 1.089818E+02
 3.128394E+00
 4.945409E-01
@@ -111,15 +111,15 @@ tally 3:
 1.061019E+02
 2.936424E+00
 4.356307E-01
-3.783939E+01
-7.188489E+01
+3.783938E+01
+7.188486E+01
 2.485680E+00
 3.117618E-01
 2.819358E+01
-3.997588E+01
+3.997585E+01
 1.875908E+00
 1.772000E-01
-1.562421E+01
+1.562420E+01
 1.236020E+01
 1.013140E+00
 5.260420E-02
@@ -160,8 +160,8 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-3.126365E+00
-4.930682E-01
+3.126366E+00
+4.930686E-01
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -208,10 +208,10 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-5.456410E+00
-1.499416E+00
-2.727755E+00
-3.788867E-01
+5.456412E+00
+1.499417E+00
+2.727756E+00
+3.788869E-01
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -256,10 +256,10 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-7.453480E+00
-2.790159E+00
-5.240103E+00
-1.386895E+00
+7.453481E+00
+2.790160E+00
+5.240104E+00
+1.386896E+00
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -304,10 +304,10 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-8.741563E+00
-3.830225E+00
-7.137968E+00
-2.555042E+00
+8.741565E+00
+3.830226E+00
+7.137969E+00
+2.555043E+00
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -354,8 +354,8 @@ tally 4:
 0.000000E+00
 9.221731E+00
 4.266573E+00
-8.474116E+00
-3.604647E+00
+8.474118E+00
+3.604648E+00
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -401,9 +401,9 @@ tally 4:
 0.000000E+00
 0.000000E+00
 9.054937E+00
-4.124537E+00
-9.066119E+00
-4.133833E+00
+4.124536E+00
+9.066120E+00
+4.133834E+00
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -448,8 +448,8 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-8.362253E+00
-3.517245E+00
+8.362251E+00
+3.517244E+00
 9.086548E+00
 4.156118E+00
 0.000000E+00
@@ -496,10 +496,10 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-6.871573E+00
-2.369190E+00
-8.399584E+00
-3.549784E+00
+6.871571E+00
+2.369189E+00
+8.399583E+00
+3.549783E+00
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -544,10 +544,10 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-4.998103E+00
-1.258194E+00
-7.145296E+00
-2.568115E+00
+4.998101E+00
+1.258193E+00
+7.145294E+00
+2.568114E+00
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -592,10 +592,10 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-2.716076E+00
-3.753568E-01
-5.427593E+00
-1.484154E+00
+2.716075E+00
+3.753566E-01
+5.427591E+00
+1.484153E+00
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -642,8 +642,8 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
-3.056645E+00
-4.702349E-01
+3.056644E+00
+4.702345E-01
 0.000000E+00
 0.000000E+00
 0.000000E+00
@@ -652,3 +652,124 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
+cmfd indices
+1.000000E+01
+1.000000E+00
+1.000000E+00
+1.000000E+00
+k cmfd
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+1.158665E+00
+1.173348E+00
+1.180514E+00
+1.168646E+00
+1.152485E+00
+1.154989E+00
+1.152285E+00
+1.152275E+00
+1.149476E+00
+1.154912E+00
+1.164393E+00
+1.170145E+00
+1.169161E+00
+1.167966E+00
+1.170153E+00
+1.170989E+00
+cmfd entropy
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+3.229801E+00
+3.226763E+00
+3.225011E+00
+3.230316E+00
+3.232043E+00
+3.228854E+00
+3.227439E+00
+3.228206E+00
+3.226749E+00
+3.222635E+00
+3.218882E+00
+3.219196E+00
+3.218057E+00
+3.220324E+00
+3.219099E+00
+3.219847E+00
+cmfd balance
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+6.244292E-03
+5.210436E-03
+4.212224E-03
+3.321578E-03
+3.067254E-03
+3.076634E-03
+2.601637E-03
+2.666441E-03
+2.490817E-03
+2.571411E-03
+3.013819E-03
+2.802836E-03
+2.725356E-03
+2.475281E-03
+2.279123E-03
+2.110816E-03
+cmfd dominance ratio
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 5.509E-01
+ 5.473E-01
+ 5.461E-01
+ 5.519E-01
+ 5.540E-01
+ 5.514E-01
+ 5.505E-01
+ 5.499E-01
+ 5.486E-01
+ 5.450E-01
+ 5.438E-01
+ 5.422E-01
+ 5.417E-01
+ 5.437E-01
+ 5.430E-01
+ 5.428E-01
+cmfd openmc source comparison
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+1.136193E-02
+1.059214E-02
+9.675100E-03
+7.296572E-03
+5.812484E-03
+6.138879E-03
+6.070297E-03
+5.791570E-03
+6.020032E-03
+7.287183E-03
+7.068423E-03
+6.380521E-03
+5.828010E-03
+4.741191E-03
+4.848114E-03
+4.686772E-03
+cmfd source
+4.367444E-02
+7.866919E-02
+1.050799E-01
+1.366581E-01
+1.421824E-01
+1.393453E-01
+1.269612E-01
+1.060419E-01
+7.920472E-02
+4.218277E-02
diff --git a/tests/test_cmfd_feed/test_cmfd_feed.py b/tests/test_cmfd_feed/test_cmfd_feed.py
index db36784888..1388ca85a5 100644
--- a/tests/test_cmfd_feed/test_cmfd_feed.py
+++ b/tests/test_cmfd_feed/test_cmfd_feed.py
@@ -14,28 +14,31 @@ pwd = os.path.dirname(__file__)
 
 skipAll = False
 
+
 def setup():
     os.putenv('PWD', pwd)
     os.chdir(pwd)
     global skipAll
     skipAll = False
 
+
 def test_run():
     openmc_path = pwd + '/../../src/openmc'
     if int(NoseMPI.mpi_np) > 0:
         proc = Popen([NoseMPI.mpi_exec, '-np', NoseMPI.mpi_np, openmc_path],
-               stderr=STDOUT, stdout=PIPE)
+                     stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     output = proc.communicate()[0]
     print(output)
+    returncode = proc.returncode
     if 'CMFD is not available' in output:
         global skipAll
         skipAll = True
         raise SkipTest
     assert returncode == 0
 
+
 def test_created_statepoint():
     if skipAll:
         raise SkipTest
@@ -43,11 +46,13 @@ def test_created_statepoint():
     assert len(statepoint) == 1
     assert statepoint[0].endswith('binary') or statepoint[0].endswith('h5')
 
+
 def test_output_exists():
     if skipAll:
         raise SkipTest
     assert os.path.exists(pwd + '/tallies.out')
 
+
 def test_results():
     if skipAll:
         raise SkipTest
@@ -55,9 +60,10 @@ def test_results():
     call(['python', 'results.py', statepoint[0]])
     compare = filecmp.cmp('results_test.dat', 'results_true.dat')
     if not compare:
-      os.rename('results_test.dat', 'results_error.dat')
+        os.rename('results_test.dat', 'results_error.dat')
     assert compare
 
+
 def teardown():
     output = glob.glob(pwd + '/statepoint.20.*')
     output.append(pwd + '/tallies.out')
@@ -65,4 +71,3 @@ def teardown():
     for f in output:
         if os.path.exists(f):
             os.remove(f)
-
diff --git a/tests/test_cmfd_jfnk/cmfd.xml b/tests/test_cmfd_jfnk/cmfd.xml
new file mode 100644
index 0000000000..d97b098fd2
--- /dev/null
+++ b/tests/test_cmfd_jfnk/cmfd.xml
@@ -0,0 +1,16 @@
+<?xml version="1.0"?>
+<cmfd>
+
+  <mesh>
+    <lower_left>  -10 -1 -1 </lower_left>
+    <upper_right>  10  1  1 </upper_right>
+    <dimension> 10 1 1 </dimension>
+    <albedo> 0.0 0.0 1.0 1.0 1.0 1.0 </albedo>
+  </mesh>
+
+  <begin> 5 </begin>
+  <display> balance </display>
+  <solver> jfnk </solver>
+  <feedback> true </feedback>
+
+</cmfd>
diff --git a/tests/test_cmfd_jfnk/geometry.xml b/tests/test_cmfd_jfnk/geometry.xml
new file mode 100644
index 0000000000..57c4aa2285
--- /dev/null
+++ b/tests/test_cmfd_jfnk/geometry.xml
@@ -0,0 +1,43 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<geometry>
+
+<!-- Definition of Cells -->
+  <cell id="1">
+    <universe>0</universe>
+    <surfaces>-1 2 -3 4 -5 6</surfaces>
+    <material>1</material>
+  </cell>
+
+<!-- Defition of Surfaces -->
+  <surface id="1">
+    <type>x-plane</type>
+    <coeffs>10</coeffs>
+    <boundary> vacuum </boundary>
+  </surface>
+  <surface id="2">
+    <type>x-plane</type>
+    <coeffs>-10</coeffs>
+    <boundary> vacuum </boundary>
+  </surface>
+  <surface id="3">
+    <type>y-plane</type>
+    <coeffs>1</coeffs>
+    <boundary>reflective</boundary>
+  </surface>
+  <surface id="4">
+    <type>y-plane</type>
+    <coeffs>-1</coeffs>
+    <boundary>reflective</boundary>
+  </surface>
+  <surface id="5">
+    <type>z-plane</type>
+    <coeffs>1</coeffs>
+    <boundary>reflective</boundary>
+  </surface>
+  <surface id="6">
+    <type>z-plane</type>
+    <coeffs>-1</coeffs>
+    <boundary>reflective</boundary>
+  </surface>
+   
+</geometry>
diff --git a/tests/test_cmfd_jfnk/materials.xml b/tests/test_cmfd_jfnk/materials.xml
new file mode 100644
index 0000000000..496774661e
--- /dev/null
+++ b/tests/test_cmfd_jfnk/materials.xml
@@ -0,0 +1,12 @@
+<?xml version="1.0"?>
+<materials>
+
+<!-- Definition of materials -->
+  <material id="1">
+    <density value="19" units="g/cc" />
+    <nuclide name="U-235" xs="70c" wo="0.21" />
+    <nuclide name="U-238" xs="70c" wo="0.68" />
+    <nuclide name="O-16" xs="70c" wo="0.11" />
+  </material>
+
+</materials>
diff --git a/tests/test_cmfd_jfnk/results.py b/tests/test_cmfd_jfnk/results.py
new file mode 100644
index 0000000000..26bcb36d9f
--- /dev/null
+++ b/tests/test_cmfd_jfnk/results.py
@@ -0,0 +1,82 @@
+#!/usr/bin/env python
+
+import sys
+import numpy as np
+
+# import statepoint
+sys.path.append('../../src/utils')
+import statepoint
+
+# read in statepoint file
+if len(sys.argv) > 1:
+    sp = statepoint.StatePoint(sys.argv[1])
+else:
+    sp = statepoint.StatePoint('statepoint.20.binary')
+sp.read_results()
+
+# extract tally results and convert to vector
+results1 = sp.tallies[0].results
+shape1 = results1.shape
+size1 = (np.product(shape1))
+results1 = np.reshape(results1, size1)
+results2 = sp.tallies[1].results
+shape2 = results2.shape
+size2 = (np.product(shape2))
+results2 = np.reshape(results2, size2)
+results3 = sp.tallies[2].results
+shape3 = results3.shape
+size3 = (np.product(shape3))
+results3 = np.reshape(results3, size3)
+results4 = sp.tallies[3].results
+shape4 = results4.shape
+size4 = (np.product(shape4))
+results4 = np.reshape(results4, size4)
+
+# set up output string
+outstr = ''
+
+# write out k-combined
+outstr += 'k-combined:\n'
+outstr += "{0:12.6E} {1:12.6E}\n".format(sp.k_combined[0], sp.k_combined[1])
+
+# write out tally results
+outstr += 'tally 1:\n'
+for item in results1:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'tally 2:\n'
+for item in results2:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'tally 3:\n'
+for item in results3:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'tally 4:\n'
+for item in results4:
+    outstr += "{0:12.6E}\n".format(item)
+
+# write out cmfd answers
+outstr += 'cmfd indices\n'
+for item in sp.cmfd_indices:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'k cmfd\n'
+for item in sp.k_cmfd:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd entropy\n'
+for item in sp.cmfd_entropy:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd balance\n'
+for item in sp.cmfd_balance:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd dominance ratio\n'
+for item in sp.cmfd_dominance:
+    outstr += "{0:10.3E}\n".format(item)
+outstr += 'cmfd openmc source comparison\n'
+for item in sp.cmfd_srccmp:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd source\n'
+cmfdsrc = np.reshape(sp.cmfd_src, np.product(sp.cmfd_indices), order='F')
+for item in cmfdsrc:
+    outstr += "{0:12.6E}\n".format(item)
+
+# write results to file
+with open('results_test.dat', 'w') as fh:
+    fh.write(outstr)
diff --git a/tests/test_cmfd_jfnk/results_true.dat b/tests/test_cmfd_jfnk/results_true.dat
new file mode 100644
index 0000000000..5e78660271
--- /dev/null
+++ b/tests/test_cmfd_jfnk/results_true.dat
@@ -0,0 +1,775 @@
+k-combined:
+1.167124E+00 1.217344E-02
+tally 1:
+1.126891E+01
+1.275756E+01
+2.086557E+01
+4.369504E+01
+2.853629E+01
+8.169563E+01
+3.404507E+01
+1.165020E+02
+3.723908E+01
+1.389881E+02
+3.760048E+01
+1.416691E+02
+3.455469E+01
+1.197186E+02
+2.837503E+01
+8.086287E+01
+2.151788E+01
+4.648312E+01
+1.194541E+01
+1.432330E+01
+tally 2:
+2.292174E+01
+2.650794E+01
+1.598184E+01
+1.288693E+01
+2.194144E+00
+2.510560E-01
+4.217817E+01
+8.938588E+01
+2.974466E+01
+4.447040E+01
+3.939271E+00
+7.882419E-01
+5.742290E+01
+1.654599E+02
+4.067582E+01
+8.308457E+01
+5.566956E+00
+1.570073E+00
+6.810587E+01
+2.331019E+02
+4.831544E+01
+1.174673E+02
+6.245000E+00
+1.967939E+00
+7.258131E+01
+2.646756E+02
+5.185146E+01
+1.351426E+02
+6.622489E+00
+2.231695E+00
+7.246115E+01
+2.641295E+02
+5.159541E+01
+1.339441E+02
+6.702239E+00
+2.277825E+00
+6.716396E+01
+2.266815E+02
+4.765062E+01
+1.141626E+02
+6.384097E+00
+2.063760E+00
+5.549541E+01
+1.545360E+02
+3.940819E+01
+7.794877E+01
+5.183373E+00
+1.369482E+00
+4.155684E+01
+8.679434E+01
+2.926743E+01
+4.307143E+01
+3.938851E+00
+7.877944E-01
+2.335179E+01
+2.758239E+01
+1.622682E+01
+1.331995E+01
+2.248920E+00
+2.594933E-01
+tally 3:
+1.537807E+01
+1.193746E+01
+1.044189E+00
+5.627797E-02
+2.861614E+01
+4.118324E+01
+1.846208E+00
+1.729007E-01
+3.921463E+01
+7.723579E+01
+2.427723E+00
+2.983153E-01
+4.653208E+01
+1.089818E+02
+3.128394E+00
+4.945409E-01
+4.989896E+01
+1.251503E+02
+3.242191E+00
+5.330743E-01
+4.967378E+01
+1.241657E+02
+3.184751E+00
+5.164524E-01
+4.592989E+01
+1.061019E+02
+2.936424E+00
+4.356307E-01
+3.783938E+01
+7.188486E+01
+2.485680E+00
+3.117618E-01
+2.819358E+01
+3.997585E+01
+1.875908E+00
+1.772000E-01
+1.562420E+01
+1.236020E+01
+1.013140E+00
+5.260420E-02
+tally 4:
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+3.126366E+00
+4.930686E-01
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+5.456412E+00
+1.499417E+00
+2.727756E+00
+3.788869E-01
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+7.453481E+00
+2.790160E+00
+5.240104E+00
+1.386896E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+8.741565E+00
+3.830226E+00
+7.137969E+00
+2.555043E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+9.221731E+00
+4.266573E+00
+8.474118E+00
+3.604648E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+9.054937E+00
+4.124536E+00
+9.066120E+00
+4.133834E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+8.362251E+00
+3.517244E+00
+9.086548E+00
+4.156118E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+6.871571E+00
+2.369189E+00
+8.399583E+00
+3.549783E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+4.998101E+00
+1.258193E+00
+7.145294E+00
+2.568114E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+2.716075E+00
+3.753566E-01
+5.427591E+00
+1.484153E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+3.056644E+00
+4.702345E-01
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+cmfd indices
+1.000000E+01
+1.000000E+00
+1.000000E+00
+1.000000E+00
+k cmfd
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+1.158665E+00
+1.173348E+00
+1.180514E+00
+1.168646E+00
+1.152485E+00
+1.154989E+00
+1.152285E+00
+1.152275E+00
+1.149476E+00
+1.154912E+00
+1.164393E+00
+1.170145E+00
+1.169161E+00
+1.167966E+00
+1.170153E+00
+1.170989E+00
+cmfd entropy
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+3.229801E+00
+3.226763E+00
+3.225011E+00
+3.230316E+00
+3.232043E+00
+3.228854E+00
+3.227439E+00
+3.228206E+00
+3.226749E+00
+3.222635E+00
+3.218882E+00
+3.219196E+00
+3.218057E+00
+3.220324E+00
+3.219099E+00
+3.219847E+00
+cmfd balance
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+6.244292E-03
+5.210436E-03
+4.212224E-03
+3.321578E-03
+3.067254E-03
+3.076634E-03
+2.601637E-03
+2.666441E-03
+2.490817E-03
+2.571411E-03
+3.013819E-03
+2.802836E-03
+2.725356E-03
+2.475281E-03
+2.279123E-03
+2.110816E-03
+cmfd dominance ratio
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+cmfd openmc source comparison
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+1.136193E-02
+1.059214E-02
+9.675100E-03
+7.296571E-03
+5.812485E-03
+6.138879E-03
+6.070297E-03
+5.791569E-03
+6.020032E-03
+7.287183E-03
+7.068423E-03
+6.380521E-03
+5.828010E-03
+4.741191E-03
+4.848115E-03
+4.686773E-03
+cmfd source
+4.367444E-02
+7.866919E-02
+1.050799E-01
+1.366581E-01
+1.421824E-01
+1.393453E-01
+1.269612E-01
+1.060419E-01
+7.920472E-02
+4.218277E-02
diff --git a/tests/test_cmfd_jfnk/settings.xml b/tests/test_cmfd_jfnk/settings.xml
new file mode 100644
index 0000000000..41de07f569
--- /dev/null
+++ b/tests/test_cmfd_jfnk/settings.xml
@@ -0,0 +1,32 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<settings>
+
+  <!-- Parameters for criticality calculation -->
+  <eigenvalue>
+    <batches>20</batches>
+    <inactive>10</inactive>
+    <particles>1000</particles>
+  </eigenvalue>
+
+  <!-- How verbose output should be -->
+  <verbosity value="7" />
+
+  <!-- Starting source -->
+  <source>
+    <space>
+      <type>box</type>
+      <parameters>-10 -1 -1 10 1 1</parameters>
+    </space>
+  </source>
+
+  <!-- Shannon Entropy --> 
+  <entropy>
+    <dimension> 10 1 1 </dimension>
+    <lower_left> -10.0 -1.0 -1.0 </lower_left>
+    <upper_right> 10.0  1.0  1.0 </upper_right>
+  </entropy>
+
+ <!-- Run CMFD -->
+ <run_cmfd> true </run_cmfd>
+
+</settings>
diff --git a/tests/test_cmfd_jfnk/tallies.xml b/tests/test_cmfd_jfnk/tallies.xml
new file mode 100644
index 0000000000..b20c0ad613
--- /dev/null
+++ b/tests/test_cmfd_jfnk/tallies.xml
@@ -0,0 +1,16 @@
+<?xml version="1.0"?>
+<tallies>
+
+  <mesh id="1">
+    <type>rectangular</type>
+    <lower_left>-10 -1 -1 </lower_left>
+    <upper_right>10 1 1</upper_right>
+    <dimension>10 1 1</dimension>
+  </mesh>
+
+  <tally id="1">
+    <filter type="mesh" bins="1" />
+    <scores>flux</scores>
+  </tally>
+
+</tallies>
diff --git a/tests/test_cmfd_jfnk/test_cmfd_jfnk.py b/tests/test_cmfd_jfnk/test_cmfd_jfnk.py
new file mode 100644
index 0000000000..1388ca85a5
--- /dev/null
+++ b/tests/test_cmfd_jfnk/test_cmfd_jfnk.py
@@ -0,0 +1,73 @@
+#!/usr/bin/env python
+
+import os
+from subprocess import Popen, STDOUT, PIPE, call
+import filecmp
+import glob
+
+from nose.plugins.skip import SkipTest
+
+from nose_mpi import NoseMPI
+
+
+pwd = os.path.dirname(__file__)
+
+skipAll = False
+
+
+def setup():
+    os.putenv('PWD', pwd)
+    os.chdir(pwd)
+    global skipAll
+    skipAll = False
+
+
+def test_run():
+    openmc_path = pwd + '/../../src/openmc'
+    if int(NoseMPI.mpi_np) > 0:
+        proc = Popen([NoseMPI.mpi_exec, '-np', NoseMPI.mpi_np, openmc_path],
+                     stderr=STDOUT, stdout=PIPE)
+    else:
+        proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
+    output = proc.communicate()[0]
+    print(output)
+    returncode = proc.returncode
+    if 'CMFD is not available' in output:
+        global skipAll
+        skipAll = True
+        raise SkipTest
+    assert returncode == 0
+
+
+def test_created_statepoint():
+    if skipAll:
+        raise SkipTest
+    statepoint = glob.glob(pwd + '/statepoint.20.*')
+    assert len(statepoint) == 1
+    assert statepoint[0].endswith('binary') or statepoint[0].endswith('h5')
+
+
+def test_output_exists():
+    if skipAll:
+        raise SkipTest
+    assert os.path.exists(pwd + '/tallies.out')
+
+
+def test_results():
+    if skipAll:
+        raise SkipTest
+    statepoint = glob.glob(pwd + '/statepoint.20.*')
+    call(['python', 'results.py', statepoint[0]])
+    compare = filecmp.cmp('results_test.dat', 'results_true.dat')
+    if not compare:
+        os.rename('results_test.dat', 'results_error.dat')
+    assert compare
+
+
+def teardown():
+    output = glob.glob(pwd + '/statepoint.20.*')
+    output.append(pwd + '/tallies.out')
+    output.append(pwd + '/results_test.dat')
+    for f in output:
+        if os.path.exists(f):
+            os.remove(f)
diff --git a/tests/test_cmfd_nofeed/results.py b/tests/test_cmfd_nofeed/results.py
index 533a624558..26bcb36d9f 100644
--- a/tests/test_cmfd_nofeed/results.py
+++ b/tests/test_cmfd_nofeed/results.py
@@ -34,7 +34,7 @@ results4 = np.reshape(results4, size4)
 
 # set up output string
 outstr = ''
- 
+
 # write out k-combined
 outstr += 'k-combined:\n'
 outstr += "{0:12.6E} {1:12.6E}\n".format(sp.k_combined[0], sp.k_combined[1])
@@ -42,17 +42,41 @@ outstr += "{0:12.6E} {1:12.6E}\n".format(sp.k_combined[0], sp.k_combined[1])
 # write out tally results
 outstr += 'tally 1:\n'
 for item in results1:
-  outstr += "{0:12.6E}\n".format(item)
+    outstr += "{0:12.6E}\n".format(item)
 outstr += 'tally 2:\n'
 for item in results2:
-  outstr += "{0:12.6E}\n".format(item)
+    outstr += "{0:12.6E}\n".format(item)
 outstr += 'tally 3:\n'
 for item in results3:
-  outstr += "{0:12.6E}\n".format(item)
+    outstr += "{0:12.6E}\n".format(item)
 outstr += 'tally 4:\n'
 for item in results4:
-  outstr += "{0:12.6E}\n".format(item)
+    outstr += "{0:12.6E}\n".format(item)
+
+# write out cmfd answers
+outstr += 'cmfd indices\n'
+for item in sp.cmfd_indices:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'k cmfd\n'
+for item in sp.k_cmfd:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd entropy\n'
+for item in sp.cmfd_entropy:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd balance\n'
+for item in sp.cmfd_balance:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd dominance ratio\n'
+for item in sp.cmfd_dominance:
+    outstr += "{0:10.3E}\n".format(item)
+outstr += 'cmfd openmc source comparison\n'
+for item in sp.cmfd_srccmp:
+    outstr += "{0:12.6E}\n".format(item)
+outstr += 'cmfd source\n'
+cmfdsrc = np.reshape(sp.cmfd_src, np.product(sp.cmfd_indices), order='F')
+for item in cmfdsrc:
+    outstr += "{0:12.6E}\n".format(item)
 
 # write results to file
-with open('results_test.dat','w') as fh:
+with open('results_test.dat', 'w') as fh:
     fh.write(outstr)
diff --git a/tests/test_cmfd_nofeed/results_true.dat b/tests/test_cmfd_nofeed/results_true.dat
index 6f00290374..683f6a6ec3 100644
--- a/tests/test_cmfd_nofeed/results_true.dat
+++ b/tests/test_cmfd_nofeed/results_true.dat
@@ -652,3 +652,124 @@ tally 4:
 0.000000E+00
 0.000000E+00
 0.000000E+00
+cmfd indices
+1.000000E+01
+1.000000E+00
+1.000000E+00
+1.000000E+00
+k cmfd
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+1.158665E+00
+1.177323E+00
+1.179341E+00
+1.169660E+00
+1.179827E+00
+1.180220E+00
+1.177699E+00
+1.179321E+00
+1.180728E+00
+1.180996E+00
+1.186680E+00
+1.181455E+00
+1.178393E+00
+1.176470E+00
+1.172593E+00
+1.167042E+00
+cmfd entropy
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+3.229801E+00
+3.227279E+00
+3.224715E+00
+3.226705E+00
+3.226404E+00
+3.227329E+00
+3.227407E+00
+3.228601E+00
+3.230423E+00
+3.228093E+00
+3.226358E+00
+3.226226E+00
+3.227324E+00
+3.228594E+00
+3.229513E+00
+3.229820E+00
+cmfd balance
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+6.244292E-03
+5.300885E-03
+5.198276E-03
+3.759109E-03
+3.994559E-03
+3.064086E-03
+2.749389E-03
+2.388483E-03
+2.304428E-03
+2.188627E-03
+2.269076E-03
+2.016776E-03
+1.622043E-03
+1.470984E-03
+1.367144E-03
+1.402525E-03
+cmfd dominance ratio
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 0.000E+00
+ 5.509E-01
+ 5.474E-01
+ 5.434E-01
+ 5.469E-01
+ 5.461E-01
+ 5.474E-01
+ 5.486E-01
+ 5.485E-01
+ 5.491E-01
+ 5.474E-01
+ 5.467E-01
+ 5.471E-01
+ 5.480E-01
+ 5.486E-01
+ 5.500E-01
+ 5.502E-01
+cmfd openmc source comparison
+0.000000E+00
+0.000000E+00
+0.000000E+00
+0.000000E+00
+1.136193E-02
+1.007349E-02
+8.776995E-03
+6.729971E-03
+6.921055E-03
+5.867589E-03
+4.366276E-03
+4.051899E-03
+3.406846E-03
+4.043983E-03
+4.864903E-03
+3.711687E-03
+3.124413E-03
+2.675579E-03
+2.523549E-03
+2.724884E-03
+cmfd source
+4.534769E-02
+8.593793E-02
+1.059054E-01
+1.317554E-01
+1.358070E-01
+1.352224E-01
+1.290227E-01
+1.093443E-01
+7.943056E-02
+4.222668E-02
diff --git a/tests/test_cmfd_nofeed/test_cmfd_nofeed.py b/tests/test_cmfd_nofeed/test_cmfd_nofeed.py
index feb4a0657e..be7ec194a5 100644
--- a/tests/test_cmfd_nofeed/test_cmfd_nofeed.py
+++ b/tests/test_cmfd_nofeed/test_cmfd_nofeed.py
@@ -13,28 +13,31 @@ pwd = os.path.dirname(__file__)
 
 skipAll = False
 
+
 def setup():
     os.putenv('PWD', pwd)
     os.chdir(pwd)
     global skipAll
     skipAll = False
 
+
 def test_run():
     openmc_path = pwd + '/../../src/openmc'
     if int(NoseMPI.mpi_np) > 0:
         proc = Popen([NoseMPI.mpi_exec, '-np', NoseMPI.mpi_np, openmc_path],
-               stderr=STDOUT, stdout=PIPE)
+                     stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     output = proc.communicate()[0]
     print(output)
+    returncode = proc.returncode
     if 'CMFD is not available' in output:
         global skipAll
         skipAll = True
         raise SkipTest
     assert returncode == 0
 
+
 def test_created_statepoint():
     if skipAll:
         raise SkipTest
@@ -42,11 +45,13 @@ def test_created_statepoint():
     assert len(statepoint) == 1
     assert statepoint[0].endswith('binary') or statepoint[0].endswith('h5')
 
+
 def test_output_exists():
     if skipAll:
         raise SkipTest
     assert os.path.exists(pwd + '/tallies.out')
 
+
 def test_results():
     if skipAll:
         raise SkipTest
@@ -54,9 +59,10 @@ def test_results():
     call(['python', 'results.py', statepoint[0]])
     compare = filecmp.cmp('results_test.dat', 'results_true.dat')
     if not compare:
-      os.rename('results_test.dat', 'results_error.dat')
+        os.rename('results_test.dat', 'results_error.dat')
     assert compare
 
+
 def teardown():
     output = glob.glob(pwd + '/statepoint.20.*')
     output.append(pwd + '/tallies.out')
@@ -64,4 +70,3 @@ def teardown():
     for f in output:
         if os.path.exists(f):
             os.remove(f)
-
diff --git a/tests/test_compile/test_compile.py b/tests/test_compile/test_compile.py
index 66368c08ad..8d411abd91 100644
--- a/tests/test_compile/test_compile.py
+++ b/tests/test_compile/test_compile.py
@@ -60,6 +60,20 @@ def test_mpi():
     shutil.move('openmc', 'openmc-mpi')
 
 
+def test_mpi_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-mpi-debug')
+
+
+def test_mpi_optimize():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'OPTIMIZE=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-mpi-optimize')
+
+
 def test_omp():
     returncode = run(['make', 'distclean'])
     returncode = run(['make', compiler, 'OPENMP=yes'])
@@ -67,6 +81,20 @@ def test_omp():
     shutil.move('openmc', 'openmc-omp')
 
 
+def test_omp_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'OPENMP=yes', 'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-omp-debug')
+
+
+def test_omp_optimize():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'OPENMP=yes', 'OPTIMIZE=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-omp-optimize')
+
+
 def test_hdf5():
     returncode = run(['make', 'distclean'])
     returncode = run(['make', compiler, 'HDF5=yes'])
@@ -74,6 +102,41 @@ def test_hdf5():
     shutil.move('openmc', 'openmc-hdf5')
 
 
+def test_hdf5_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'HDF5=yes', 'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-hdf5-debug')
+
+
+def test_hdf5_optimize():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'HDF5=yes', 'OPTIMIZE=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-hdf5-optimize')
+
+
+def test_omp_hdf5():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'OPENMP=yes', 'HDF5=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-omp-hdf5')
+
+
+def test_omp_hdf5_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'OPENMP=yes', 'HDF5=yes', 'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-omp-hdf5-debug')
+
+
+def test_omp_hdf5_optimize():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'OPENMP=yes', 'HDF5=yes', 'OPTIMIZE=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-omp-hdf5-optimize')
+
+
 def test_petsc():
     returncode = run(['make', 'distclean'])
     returncode = run(['make', compiler, 'MPI=yes', 'PETSC=yes'])
@@ -81,6 +144,21 @@ def test_petsc():
     shutil.move('openmc', 'openmc-petsc')
 
 
+def test_petsc_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'PETSC=yes', 'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-petsc-debug')
+
+
+def test_petsc_optimize():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'PETSC=yes',
+                      'OPTIMIZE=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-petsc-optimize')
+
+
 def test_mpi_omp():
     returncode = run(['make', 'distclean'])
     returncode = run(['make', compiler, 'MPI=yes', 'OPENMP=yes'])
@@ -88,11 +166,18 @@ def test_mpi_omp():
     shutil.move('openmc', 'openmc-mpi-omp')
 
 
-def test_omp_hdf5():
+def test_mpi_omp_debug():
     returncode = run(['make', 'distclean'])
-    returncode = run(['make', compiler, 'MPI=yes', 'OPENMP=yes', 'HDF5=yes'])
+    returncode = run(['make', compiler, 'MPI=yes', 'OPENMP=yes', 'DEBUG=yes'])
     assert returncode == 0
-    shutil.move('openmc', 'openmc-omp-hdf5')
+    shutil.move('openmc', 'openmc-mpi-omp-debug')
+
+
+def test_mpi_omp_optimize():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'OPENMP=yes', 'OPTIMIZE=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-mpi-omp-optimize')
 
 
 def test_mpi_hdf5():
@@ -102,6 +187,43 @@ def test_mpi_hdf5():
     shutil.move('openmc', 'openmc-phdf5')
 
 
+def test_mpi_hdf5_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'HDF5=yes', 'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-phdf5-debug')
+
+
+def test_mpi_hdf5_optimize():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'HDF5=yes', 'OPTIMIZE=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-phdf5-optimize')
+
+
+def test_mpi_omp_hdf5():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'OPENMP=yes', 'HDF5=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-phdf5-omp')
+
+
+def test_mpi_omp_hdf5_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'OPENMP=yes', 'HDF5=yes',
+                      'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-phdf5-omp-debug')
+
+
+def test_mpi_omp_hdf5_optimize():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'OPENMP=yes', 'HDF5=yes',
+                      'OPTIMIZE=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-phdf5-omp-optimize')
+
+
 def test_mpi_hdf5_petsc():
     returncode = run(['make', 'distclean'])
     returncode = run(['make', compiler, 'MPI=yes', 'HDF5=yes', 'PETSC=yes'])
@@ -109,6 +231,14 @@ def test_mpi_hdf5_petsc():
     shutil.move('openmc', 'openmc-phdf5-petsc')
 
 
+def test_mpi_hdf5_petsc_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'HDF5=yes', 'PETSC=yes',
+                      'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-phdf5-petsc-debug')
+
+
 def test_mpi_hdf5_petsc_optimize():
     returncode = run(['make', 'distclean'])
     returncode = run(['make', compiler, 'MPI=yes', 'HDF5=yes', 'PETSC=yes',
@@ -117,6 +247,22 @@ def test_mpi_hdf5_petsc_optimize():
     shutil.move('openmc', 'openmc-phdf5-petsc-optimize')
 
 
+def test_mpi_omp_hdf5_petsc():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'OMP=yes', 'HDF5=yes',
+                      'PETSC=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-omp-phdf5-petsc')
+
+
+def test_mpi_omp_hdf5_petsc_debug():
+    returncode = run(['make', 'distclean'])
+    returncode = run(['make', compiler, 'MPI=yes', 'OMP=yes', 'HDF5=yes',
+                      'PETSC=yes', 'DEBUG=yes'])
+    assert returncode == 0
+    shutil.move('openmc', 'openmc-omp-phdf5-petsc-debug')
+
+
 def test_mpi_omp_hdf5_petsc_optimize():
     returncode = run(['make', 'distclean'])
     returncode = run(['make', compiler, 'MPI=yes', 'OMP=yes', 'HDF5=yes',
@@ -127,8 +273,8 @@ def test_mpi_omp_hdf5_petsc_optimize():
 
 def run(commands):
     proc = Popen(commands, stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     return returncode
 
 
diff --git a/tests/test_confidence_intervals/test_confidence_intervals.py b/tests/test_confidence_intervals/test_confidence_intervals.py
index a0dda8b468..a8372834f7 100644
--- a/tests/test_confidence_intervals/test_confidence_intervals.py
+++ b/tests/test_confidence_intervals/test_confidence_intervals.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_density_atombcm/test_density_atombcm.py b/tests/test_density_atombcm/test_density_atombcm.py
index d13f404722..2fde1e6312 100644
--- a/tests/test_density_atombcm/test_density_atombcm.py
+++ b/tests/test_density_atombcm/test_density_atombcm.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_density_atomcm3/test_density_atomcm3.py b/tests/test_density_atomcm3/test_density_atomcm3.py
index 860f858d34..74a644b59d 100644
--- a/tests/test_density_atomcm3/test_density_atomcm3.py
+++ b/tests/test_density_atomcm3/test_density_atomcm3.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_density_kgm3/test_density_kgm3.py b/tests/test_density_kgm3/test_density_kgm3.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_density_kgm3/test_density_kgm3.py
+++ b/tests/test_density_kgm3/test_density_kgm3.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_density_sum/test_density_sum.py b/tests/test_density_sum/test_density_sum.py
index d1bc7d3a69..fcafcf0f1f 100644
--- a/tests/test_density_sum/test_density_sum.py
+++ b/tests/test_density_sum/test_density_sum.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_eigenvalue_genperbatch/test_eigenvalue_genperbatch.py b/tests/test_eigenvalue_genperbatch/test_eigenvalue_genperbatch.py
index eb4e206c1b..24077bd591 100644
--- a/tests/test_eigenvalue_genperbatch/test_eigenvalue_genperbatch.py
+++ b/tests/test_eigenvalue_genperbatch/test_eigenvalue_genperbatch.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_eigenvalue_no_inactive/test_eigenvalue_no_inactive.py b/tests/test_eigenvalue_no_inactive/test_eigenvalue_no_inactive.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_eigenvalue_no_inactive/test_eigenvalue_no_inactive.py
+++ b/tests/test_eigenvalue_no_inactive/test_eigenvalue_no_inactive.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_energy_grid/test_energy_grid.py b/tests/test_energy_grid/test_energy_grid.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_energy_grid/test_energy_grid.py
+++ b/tests/test_energy_grid/test_energy_grid.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_entropy/test_entropy.py b/tests/test_entropy/test_entropy.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_entropy/test_entropy.py
+++ b/tests/test_entropy/test_entropy.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_cell/test_filter_cell.py b/tests/test_filter_cell/test_filter_cell.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_cell/test_filter_cell.py
+++ b/tests/test_filter_cell/test_filter_cell.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_cellborn/test_filter_cellborn.py b/tests/test_filter_cellborn/test_filter_cellborn.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_cellborn/test_filter_cellborn.py
+++ b/tests/test_filter_cellborn/test_filter_cellborn.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_energy/test_filter_energy.py b/tests/test_filter_energy/test_filter_energy.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_energy/test_filter_energy.py
+++ b/tests/test_filter_energy/test_filter_energy.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_energyout/test_filter_energyout.py b/tests/test_filter_energyout/test_filter_energyout.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_energyout/test_filter_energyout.py
+++ b/tests/test_filter_energyout/test_filter_energyout.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_group_transfer/test_filter_group_transfer.py b/tests/test_filter_group_transfer/test_filter_group_transfer.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_group_transfer/test_filter_group_transfer.py
+++ b/tests/test_filter_group_transfer/test_filter_group_transfer.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_material/test_filter_material.py b/tests/test_filter_material/test_filter_material.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_material/test_filter_material.py
+++ b/tests/test_filter_material/test_filter_material.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_mesh_2d/test_filter_mesh_2d.py b/tests/test_filter_mesh_2d/test_filter_mesh_2d.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_mesh_2d/test_filter_mesh_2d.py
+++ b/tests/test_filter_mesh_2d/test_filter_mesh_2d.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_mesh_3d/test_filter_mesh_3d.py b/tests/test_filter_mesh_3d/test_filter_mesh_3d.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_mesh_3d/test_filter_mesh_3d.py
+++ b/tests/test_filter_mesh_3d/test_filter_mesh_3d.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_filter_universe/test_filter_universe.py b/tests/test_filter_universe/test_filter_universe.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_filter_universe/test_filter_universe.py
+++ b/tests/test_filter_universe/test_filter_universe.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_fixed_source/test_fixed_source.py b/tests/test_fixed_source/test_fixed_source.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_fixed_source/test_fixed_source.py
+++ b/tests/test_fixed_source/test_fixed_source.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_lattice/test_lattice.py b/tests/test_lattice/test_lattice.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_lattice/test_lattice.py
+++ b/tests/test_lattice/test_lattice.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_lattice_multiple/test_lattice_multiple.py b/tests/test_lattice_multiple/test_lattice_multiple.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_lattice_multiple/test_lattice_multiple.py
+++ b/tests/test_lattice_multiple/test_lattice_multiple.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_natural_element/test_natural_element.py b/tests/test_natural_element/test_natural_element.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_natural_element/test_natural_element.py
+++ b/tests/test_natural_element/test_natural_element.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_output/test_output.py b/tests/test_output/test_output.py
index 9024d6fb19..dc48dc37b9 100644
--- a/tests/test_output/test_output.py
+++ b/tests/test_output/test_output.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_summary_exists():
diff --git a/tests/test_plot_background/test_plot_background.py b/tests/test_plot_background/test_plot_background.py
index a34b5bbfb7..d8fc18f755 100644
--- a/tests/test_plot_background/test_plot_background.py
+++ b/tests/test_plot_background/test_plot_background.py
@@ -17,8 +17,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path, '-p'], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_plot_exists():
diff --git a/tests/test_plot_basis/test_plot_basis.py b/tests/test_plot_basis/test_plot_basis.py
index fe0a13ccac..86a4159200 100644
--- a/tests/test_plot_basis/test_plot_basis.py
+++ b/tests/test_plot_basis/test_plot_basis.py
@@ -17,8 +17,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path, '-p'], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_plots_exists():
diff --git a/tests/test_plot_colspec/test_plot_colspec.py b/tests/test_plot_colspec/test_plot_colspec.py
index a34b5bbfb7..d8fc18f755 100644
--- a/tests/test_plot_colspec/test_plot_colspec.py
+++ b/tests/test_plot_colspec/test_plot_colspec.py
@@ -17,8 +17,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path, '-p'], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_plot_exists():
diff --git a/tests/test_plot_mask/test_plot_mask.py b/tests/test_plot_mask/test_plot_mask.py
index fe0a13ccac..86a4159200 100644
--- a/tests/test_plot_mask/test_plot_mask.py
+++ b/tests/test_plot_mask/test_plot_mask.py
@@ -17,8 +17,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path, '-p'], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_plots_exists():
diff --git a/tests/test_ptables_off/test_ptables_off.py b/tests/test_ptables_off/test_ptables_off.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_ptables_off/test_ptables_off.py
+++ b/tests/test_ptables_off/test_ptables_off.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_reflective_cone/test_reflective_cone.py b/tests/test_reflective_cone/test_reflective_cone.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_reflective_cone/test_reflective_cone.py
+++ b/tests/test_reflective_cone/test_reflective_cone.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_reflective_cylinder/test_reflective_cylinder.py b/tests/test_reflective_cylinder/test_reflective_cylinder.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_reflective_cylinder/test_reflective_cylinder.py
+++ b/tests/test_reflective_cylinder/test_reflective_cylinder.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_reflective_plane/test_reflective_plane.py b/tests/test_reflective_plane/test_reflective_plane.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_reflective_plane/test_reflective_plane.py
+++ b/tests/test_reflective_plane/test_reflective_plane.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_reflective_sphere/test_reflective_sphere.py b/tests/test_reflective_sphere/test_reflective_sphere.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_reflective_sphere/test_reflective_sphere.py
+++ b/tests/test_reflective_sphere/test_reflective_sphere.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_rotation/test_rotation.py b/tests/test_rotation/test_rotation.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_rotation/test_rotation.py
+++ b/tests/test_rotation/test_rotation.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_salphabeta/test_salphabeta.py b/tests/test_salphabeta/test_salphabeta.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_salphabeta/test_salphabeta.py
+++ b/tests/test_salphabeta/test_salphabeta.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_salphabeta_multiple/test_salphabeta_multiple.py b/tests/test_salphabeta_multiple/test_salphabeta_multiple.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_salphabeta_multiple/test_salphabeta_multiple.py
+++ b/tests/test_salphabeta_multiple/test_salphabeta_multiple.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_MT/test_score_MT.py b/tests/test_score_MT/test_score_MT.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_MT/test_score_MT.py
+++ b/tests/test_score_MT/test_score_MT.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_absorption/test_score_absorption.py b/tests/test_score_absorption/test_score_absorption.py
index b49faa8177..335d649dd7 100644
--- a/tests/test_score_absorption/test_score_absorption.py
+++ b/tests/test_score_absorption/test_score_absorption.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_current/test_score_current.py b/tests/test_score_current/test_score_current.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_current/test_score_current.py
+++ b/tests/test_score_current/test_score_current.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_events/test_score_events.py b/tests/test_score_events/test_score_events.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_events/test_score_events.py
+++ b/tests/test_score_events/test_score_events.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_fission/test_score_fission.py b/tests/test_score_fission/test_score_fission.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_fission/test_score_fission.py
+++ b/tests/test_score_fission/test_score_fission.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_flux/test_score_flux.py b/tests/test_score_flux/test_score_flux.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_flux/test_score_flux.py
+++ b/tests/test_score_flux/test_score_flux.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_kappafission/test_score_kappafission.py b/tests/test_score_kappafission/test_score_kappafission.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_kappafission/test_score_kappafission.py
+++ b/tests/test_score_kappafission/test_score_kappafission.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_nufission/test_score_nufission.py b/tests/test_score_nufission/test_score_nufission.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_nufission/test_score_nufission.py
+++ b/tests/test_score_nufission/test_score_nufission.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_nuscatter/test_score_nuscatter.py b/tests/test_score_nuscatter/test_score_nuscatter.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_nuscatter/test_score_nuscatter.py
+++ b/tests/test_score_nuscatter/test_score_nuscatter.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_scatter/test_score_scatter.py b/tests/test_score_scatter/test_score_scatter.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_scatter/test_score_scatter.py
+++ b/tests/test_score_scatter/test_score_scatter.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_scatter_n/test_score_scatter_n.py b/tests/test_score_scatter_n/test_score_scatter_n.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_scatter_n/test_score_scatter_n.py
+++ b/tests/test_score_scatter_n/test_score_scatter_n.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_scatter_pn/test_score_scatter_pn.py b/tests/test_score_scatter_pn/test_score_scatter_pn.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_scatter_pn/test_score_scatter_pn.py
+++ b/tests/test_score_scatter_pn/test_score_scatter_pn.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_score_total/test_score_total.py b/tests/test_score_total/test_score_total.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_score_total/test_score_total.py
+++ b/tests/test_score_total/test_score_total.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_seed/test_seed.py b/tests/test_seed/test_seed.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_seed/test_seed.py
+++ b/tests/test_seed/test_seed.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_source_angle_mono/test_source_angle_mono.py b/tests/test_source_angle_mono/test_source_angle_mono.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_source_angle_mono/test_source_angle_mono.py
+++ b/tests/test_source_angle_mono/test_source_angle_mono.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_source_energy_maxwell/test_source_energy_maxwell.py b/tests/test_source_energy_maxwell/test_source_energy_maxwell.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_source_energy_maxwell/test_source_energy_maxwell.py
+++ b/tests/test_source_energy_maxwell/test_source_energy_maxwell.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_source_energy_mono/test_source_energy_mono.py b/tests/test_source_energy_mono/test_source_energy_mono.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_source_energy_mono/test_source_energy_mono.py
+++ b/tests/test_source_energy_mono/test_source_energy_mono.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_source_point/test_source_point.py b/tests/test_source_point/test_source_point.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_source_point/test_source_point.py
+++ b/tests/test_source_point/test_source_point.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_statepoint_batch/test_statepoint_batch.py b/tests/test_statepoint_batch/test_statepoint_batch.py
index 57d00dbe8f..ed4f7b2a7a 100644
--- a/tests/test_statepoint_batch/test_statepoint_batch.py
+++ b/tests/test_statepoint_batch/test_statepoint_batch.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_statepoints_exist():
diff --git a/tests/test_statepoint_interval/test_statepoint_interval.py b/tests/test_statepoint_interval/test_statepoint_interval.py
index 5c3c87fbc5..263ab37907 100644
--- a/tests/test_statepoint_interval/test_statepoint_interval.py
+++ b/tests/test_statepoint_interval/test_statepoint_interval.py
@@ -20,8 +20,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_statepoints_exist():
diff --git a/tests/test_statepoint_restart/test_statepoint_restart.py b/tests/test_statepoint_restart/test_statepoint_restart.py
index 7a880d59b0..544c9197ec 100644
--- a/tests/test_statepoint_restart/test_statepoint_restart.py
+++ b/tests/test_statepoint_restart/test_statepoint_restart.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
@@ -45,8 +45,8 @@ def test_restart_form1():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path, '-r', statepoint[0]], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint_form1():
@@ -72,8 +72,8 @@ def test_restart_form2():
     else:
         proc = Popen([openmc_path, '--restart', statepoint[0]], 
                      stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint_form2():
@@ -94,8 +94,8 @@ def test_restart_serial():
     openmc_path = pwd + '/../../src/openmc'
     proc = Popen([openmc_path, '--restart', statepoint[0]], 
                  stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint_serial():
diff --git a/tests/test_statepoint_sourcesep/test_statepoint_sourcesep.py b/tests/test_statepoint_sourcesep/test_statepoint_sourcesep.py
index 61191c57ae..ca1c9966a4 100644
--- a/tests/test_statepoint_sourcesep/test_statepoint_sourcesep.py
+++ b/tests/test_statepoint_sourcesep/test_statepoint_sourcesep.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_statepoint_exists():
diff --git a/tests/test_survival_biasing/test_survival_biasing.py b/tests/test_survival_biasing/test_survival_biasing.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_survival_biasing/test_survival_biasing.py
+++ b/tests/test_survival_biasing/test_survival_biasing.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_tally_assumesep/test_tally_assumesep.py b/tests/test_tally_assumesep/test_tally_assumesep.py
index 62c4d7316a..151df366bd 100644
--- a/tests/test_tally_assumesep/test_tally_assumesep.py
+++ b/tests/test_tally_assumesep/test_tally_assumesep.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_trace/test_trace.py b/tests/test_trace/test_trace.py
index d618dfc5e8..a518ab7731 100644
--- a/tests/test_trace/test_trace.py
+++ b/tests/test_trace/test_trace.py
@@ -19,9 +19,9 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     stdout = proc.communicate()[0]
     print(stdout)
+    returncode = proc.returncode
     assert returncode == 0
     assert stdout.find('Simulating Particle 453') != -1
 
diff --git a/tests/test_translation/test_translation.py b/tests/test_translation/test_translation.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_translation/test_translation.py
+++ b/tests/test_translation/test_translation.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_uniform_fs/test_uniform_fs.py b/tests/test_uniform_fs/test_uniform_fs.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_uniform_fs/test_uniform_fs.py
+++ b/tests/test_uniform_fs/test_uniform_fs.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_universe/test_universe.py b/tests/test_universe/test_universe.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_universe/test_universe.py
+++ b/tests/test_universe/test_universe.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():
diff --git a/tests/test_void/test_void.py b/tests/test_void/test_void.py
index fe8344c15c..64d20b6115 100644
--- a/tests/test_void/test_void.py
+++ b/tests/test_void/test_void.py
@@ -19,8 +19,8 @@ def test_run():
                stderr=STDOUT, stdout=PIPE)
     else:
         proc = Popen([openmc_path], stderr=STDOUT, stdout=PIPE)
-    returncode = proc.wait()
     print(proc.communicate()[0])
+    returncode = proc.returncode
     assert returncode == 0
 
 def test_created_statepoint():