mirror of
https://github.com/nwchemgit/nwchem.git
synced 2026-07-28 22:25:48 -04:00
NWChem: Open Source High-Performance Computational Chemistry
GitHub
commit
b5a513d7b0
64 changed files with 8568 additions and 1074 deletions
6
.github/workflows/github_actions.yml
vendored
6
.github/workflows/github_actions.yml
vendored
|
|
@ -209,12 +209,6 @@ jobs:
|
|||
armci_network: MPI-PR
|
||||
nwchem_modules: "tinyqmpw python"
|
||||
fc: ifx
|
||||
- os: ubuntu-latest
|
||||
experimental: true
|
||||
mpi_impl: intel
|
||||
armci_network: SOCKETS
|
||||
nwchem_modules: "tce"
|
||||
fc: ifx
|
||||
- os: ubuntu-22.04
|
||||
experimental: true
|
||||
mpi_impl: intel
|
||||
|
|
|
|||
|
|
@ -577,6 +577,9 @@ let "myexit+=$?"
|
|||
let "myexit+=$?"
|
||||
./runtests.mpi.unix procs $np gw_symmetry
|
||||
let "myexit+=$?"
|
||||
# BSE
|
||||
./runtests.mpi.unix procs $np bse_ethene
|
||||
let "myexit+=$?"
|
||||
# x2c
|
||||
./runtests.mpi.unix procs $np x2c-h2se
|
||||
let "myexit+=$?"
|
||||
|
|
|
|||
|
|
@ -433,6 +433,13 @@ sync
|
|||
set overall_status = 1
|
||||
continue
|
||||
fi
|
||||
# delete @GW lines for BSE tests
|
||||
if [[ `grep -c 'NWChem BSE Module' ${STUB}.out` > 0 ]] ; then
|
||||
rm -f sort.new
|
||||
cat ${STUB}.ok.out.nwparse |grep -v @GW > sort.new && mv sort.new ${STUB}.ok.out.nwparse
|
||||
rm -f sort.new
|
||||
cat ${STUB}.out.nwparse |grep -v @GW > sort.new && mv sort.new ${STUB}.out.nwparse
|
||||
fi
|
||||
#check if output is from EOMCCSD, since EOMCCSD output is non-deterministic
|
||||
if [[ `grep -c EOMCCSD ${STUB}.out.nwparse` > 0 ]] ; then
|
||||
rm -f sort.new
|
||||
|
|
|
|||
45
QA/tests/bse_ethene/bse_ethene.nw
Normal file
45
QA/tests/bse_ethene/bse_ethene.nw
Normal file
|
|
@ -0,0 +1,45 @@
|
|||
title "ethene BSE"
|
||||
memory total 2000 mb
|
||||
start
|
||||
|
||||
set lindep:n_dep 0
|
||||
|
||||
geometry
|
||||
H 0.000000 0.923274 1.238289
|
||||
H 0.000000 -0.923274 1.238289
|
||||
H 0.000000 0.923274 -1.238289
|
||||
H 0.000000 -0.923274 -1.238289
|
||||
C 0.000000 0.000000 0.668188
|
||||
C 0.000000 0.000000 -0.668188
|
||||
end
|
||||
|
||||
basis "ao basis" spherical bse
|
||||
* library aug-cc-pvtz
|
||||
end
|
||||
|
||||
basis "ri basis" spherical bse
|
||||
* library aug-cc-pvtz-rifit
|
||||
end
|
||||
|
||||
dft
|
||||
direct
|
||||
xc pbe0
|
||||
tolerances acccoul 10
|
||||
convergence energy 1d-8
|
||||
grid fine
|
||||
end
|
||||
|
||||
gw
|
||||
evgw 5
|
||||
method analytic
|
||||
states alpha occ -1 vir -1
|
||||
end
|
||||
|
||||
task gw
|
||||
|
||||
bse
|
||||
nroots 12
|
||||
method davidson
|
||||
end
|
||||
|
||||
task bse
|
||||
3934
QA/tests/bse_ethene/bse_ethene.out
Normal file
3934
QA/tests/bse_ethene/bse_ethene.out
Normal file
File diff suppressed because it is too large
Load diff
|
|
@ -35,11 +35,8 @@ end
|
|||
|
||||
gw
|
||||
core
|
||||
eta 0.01
|
||||
method cdgw
|
||||
states alpha occ 7
|
||||
solver newton 15
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
|
@ -52,27 +49,8 @@ end
|
|||
gw
|
||||
core
|
||||
first 4
|
||||
eta 0.01
|
||||
method cdgw 128
|
||||
solver newton 15
|
||||
method cdgw
|
||||
states alpha occ 4
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
||||
dft
|
||||
noscf
|
||||
end
|
||||
|
||||
unset gw:first
|
||||
|
||||
gw
|
||||
core
|
||||
eta 0.01
|
||||
solver newton 15
|
||||
states alpha occ 7
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
|
|
|||
|
|
@ -1,4 +1,5 @@
|
|||
argument 1 = gw_closedshell.nw
|
||||
argument 1 = /people/meji656/calculations/bse/QA/gw_closedshell.nw
|
||||
NWChem w/ OpenMP: maximum threads = 1
|
||||
|
||||
|
||||
|
||||
|
|
@ -21,11 +22,11 @@ H 2.5025 -1.2628 -0.0022
|
|||
H 0.9511 -2.0241 -0.0008
|
||||
end
|
||||
|
||||
basis "ao basis" spherical
|
||||
basis "ao basis" bse spherical
|
||||
* library cc-pvdz
|
||||
end
|
||||
|
||||
basis "cd basis" spherical
|
||||
basis "cd basis" bse spherical
|
||||
* library cc-pvdz-rifit
|
||||
end
|
||||
|
||||
|
|
@ -40,11 +41,8 @@ end
|
|||
|
||||
gw
|
||||
core
|
||||
eta 0.01
|
||||
method cdgw
|
||||
states alpha occ 7
|
||||
solver newton 15
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
|
@ -57,27 +55,8 @@ end
|
|||
gw
|
||||
core
|
||||
first 4
|
||||
eta 0.01
|
||||
method cdgw 128
|
||||
solver newton 15
|
||||
method cdgw
|
||||
states alpha occ 4
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
||||
dft
|
||||
noscf
|
||||
end
|
||||
|
||||
unset gw:first
|
||||
|
||||
gw
|
||||
core
|
||||
eta 0.01
|
||||
solver newton 15
|
||||
states alpha occ 7
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
|
@ -88,7 +67,7 @@ task dft gw
|
|||
|
||||
|
||||
|
||||
Northwest Computational Chemistry Package (NWChem) 7.0.1
|
||||
Northwest Computational Chemistry Package (NWChem) 7.3.0
|
||||
--------------------------------------------------------
|
||||
|
||||
|
||||
|
|
@ -96,7 +75,7 @@ task dft gw
|
|||
Pacific Northwest National Laboratory
|
||||
Richland, WA 99352
|
||||
|
||||
Copyright (c) 1994-2020
|
||||
Copyright (c) 1994-2025
|
||||
Pacific Northwest National Laboratory
|
||||
Battelle Memorial Institute
|
||||
|
||||
|
|
@ -121,21 +100,21 @@ task dft gw
|
|||
Job information
|
||||
---------------
|
||||
|
||||
hostname = dmejiar-VirtualBox
|
||||
program = nwchem
|
||||
date = Fri Feb 4 04:13:35 2022
|
||||
hostname = dc004
|
||||
program = /qfs/people/meji656/software/nwchem.bse/bin/LINUX64/nwchem
|
||||
date = Thu Oct 16 10:34:35 2025
|
||||
|
||||
compiled = Fri_Feb_04_04:04:58_2022
|
||||
source = /home/dmejiar/Software/nwchem_gfortran_gw
|
||||
nwchem branch = 7.0.0
|
||||
nwchem revision = nwchem_on_git-3497-g0c7e480c37
|
||||
ga revision = 5.8.1
|
||||
use scalapack = T
|
||||
input = gw_closedshell.nw
|
||||
compiled = Thu_Oct_16_10:07:15_2025
|
||||
source = /qfs/people/meji656/software/nwchem.bse
|
||||
nwchem branch = 7.3.0
|
||||
nwchem revision = v7.2.0-beta1-1624-g0dc011d0a0
|
||||
ga revision = 5.9.2
|
||||
use scalapack = F
|
||||
input = /people/meji656/calculations/bse/QA/gw_closedshell.nw
|
||||
prefix = gw_closedshell.
|
||||
data base = ./gw_closedshell.db
|
||||
status = startup
|
||||
nproc = 2
|
||||
nproc = 7
|
||||
time left = -1s
|
||||
|
||||
|
||||
|
|
@ -143,10 +122,10 @@ task dft gw
|
|||
Memory information
|
||||
------------------
|
||||
|
||||
heap = 65536000 doubles = 500.0 Mbytes
|
||||
stack = 65535997 doubles = 500.0 Mbytes
|
||||
global = 131072000 doubles = 1000.0 Mbytes (distinct from heap & stack)
|
||||
total = 262143997 doubles = 2000.0 Mbytes
|
||||
heap = 65535998 doubles = 500.0 Mbytes
|
||||
stack = 65535995 doubles = 500.0 Mbytes
|
||||
global = 131072000 doubles = 1000.0 Mbytes (distinct from heap & stack)
|
||||
total = 262143993 doubles = 2000.0 Mbytes
|
||||
verify = yes
|
||||
hardfail = no
|
||||
|
||||
|
|
@ -311,9 +290,6 @@ task dft gw
|
|||
|
||||
|
||||
|
||||
library name resolved from: environment
|
||||
library file name is: </home/dmejiar/Software/nwchem_gfortran_gw/src/basis/libraries.bse/>
|
||||
|
||||
|
||||
|
||||
Summary of "ao basis" -> "" (spherical)
|
||||
|
|
@ -323,9 +299,6 @@ task dft gw
|
|||
* cc-pvdz on all atoms
|
||||
|
||||
|
||||
library name resolved from: environment
|
||||
library file name is: </home/dmejiar/Software/nwchem_gfortran_gw/src/basis/libraries.bse/>
|
||||
|
||||
|
||||
|
||||
Summary of "cd basis" -> "" (spherical)
|
||||
|
|
@ -337,8 +310,8 @@ task dft gw
|
|||
|
||||
pbe96 is a nonlocal functional; adding pw91lda local functional.
|
||||
|
||||
NWChem GW@DFT Module
|
||||
--------------------
|
||||
NWChem GW Module
|
||||
----------------
|
||||
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
|
@ -796,48 +769,48 @@ task dft gw
|
|||
HOMO = -0.354463
|
||||
LUMO = 0.023799
|
||||
|
||||
Time after variat. SCF: 1.2
|
||||
Time after variat. SCF: 0.8
|
||||
|
||||
3 Center 2 Electron Integral Information
|
||||
----------------------------------------
|
||||
Maximum number of 3-center 2e- integrals is: 4898880.
|
||||
This is reduced with Schwarz screening to: 3084900.
|
||||
Incore requires a per proc buffer size of: 1377181.
|
||||
Incore requires a per proc buffer size of: 398581.
|
||||
Minimum dble words available (all nodes) is: 131060988
|
||||
This is reduced (for later use) to: 130785792
|
||||
proc 0 Suggested buffer size is: 1377181
|
||||
Max Suggested buffer size is: 1377181
|
||||
This is reduced (for later use) to: 130844112
|
||||
proc 0 Suggested buffer size is: 398581
|
||||
Max Suggested buffer size is: 398581
|
||||
no. integral batches is: 1000
|
||||
|
||||
Incore memory use for 3-center 2e- integrals is turned off.
|
||||
Time prior to 1st pass: 1.2
|
||||
Time prior to 1st pass: 0.8
|
||||
|
||||
Memory utilization after 1st SCF pass:
|
||||
Heap Space remaining (MW): 65.53 65525792
|
||||
Stack Space remaining (MW): 65.54 65535156
|
||||
Heap Space remaining (MW): 65.53 65525790
|
||||
Stack Space remaining (MW): 65.54 65535148
|
||||
|
||||
convergence iter energy DeltaE RMS-Dens Diis-err time
|
||||
---------------- ----- ----------------- --------- --------- --------- ------
|
||||
d= 0,ls=0.0,diis 1 -506.2138586363 -8.41D+02 1.85D-02 2.46D+00 5.4
|
||||
d= 0,ls=0.0,diis 2 -505.9428729180 2.71D-01 1.15D-02 5.90D+00 9.7
|
||||
d= 0,ls=0.0,diis 3 -506.4729357426 -5.30D-01 2.21D-03 1.04D-01 13.9
|
||||
d= 0,ls=0.0,diis 4 -506.4807116964 -7.78D-03 8.40D-04 2.68D-02 18.1
|
||||
d= 0,ls=0.0,diis 5 -506.4832631842 -2.55D-03 1.27D-04 5.68D-04 23.8
|
||||
d= 0,ls=0.0,diis 1 -506.2138586373 -8.41D+02 1.85D-02 2.46D+00 1.3
|
||||
d= 0,ls=0.0,diis 2 -505.9428729965 2.71D-01 1.15D-02 5.90D+00 1.9
|
||||
d= 0,ls=0.0,diis 3 -506.4729357440 -5.30D-01 2.21D-03 1.04D-01 2.4
|
||||
d= 0,ls=0.0,diis 4 -506.4807116867 -7.78D-03 8.40D-04 2.68D-02 3.0
|
||||
d= 0,ls=0.0,diis 5 -506.4832631820 -2.55D-03 1.27D-04 5.68D-04 3.7
|
||||
Resetting Diis
|
||||
d= 0,ls=0.0,diis 6 -506.4833170187 -5.38D-05 4.21D-05 2.12D-05 29.7
|
||||
d= 0,ls=0.0,diis 7 -506.4833201162 -3.10D-06 2.25D-05 3.93D-06 36.1
|
||||
d= 0,ls=0.0,diis 8 -506.4833202140 -9.77D-08 9.51D-06 4.40D-06 42.1
|
||||
d= 0,ls=0.0,diis 6 -506.4833170167 -5.38D-05 4.21D-05 2.12D-05 4.4
|
||||
d= 0,ls=0.0,diis 7 -506.4833201141 -3.10D-06 2.25D-05 3.93D-06 5.2
|
||||
d= 0,ls=0.0,diis 8 -506.4833202118 -9.77D-08 9.51D-06 4.40D-06 5.9
|
||||
|
||||
|
||||
Total DFT energy = -506.483320213950
|
||||
One electron energy = -1365.175909550851
|
||||
Coulomb energy = 581.765579688889
|
||||
Exchange-Corr. energy = -58.042890125742
|
||||
Total DFT energy = -506.483320211761
|
||||
One electron energy = -1365.175909547848
|
||||
Coulomb energy = 581.765579681333
|
||||
Exchange-Corr. energy = -58.042890118999
|
||||
Nuclear repulsion energy = 334.969899773753
|
||||
|
||||
Numeric. integr. density = 56.000003719920
|
||||
Numeric. integr. density = 56.000003716419
|
||||
|
||||
Total iterative time = 40.9s
|
||||
Total iterative time = 5.1s
|
||||
|
||||
|
||||
|
||||
|
|
@ -870,32 +843,36 @@ task dft gw
|
|||
2 0 0 2 -28.307350 -52.360374 -52.360374 76.413398
|
||||
|
||||
|
||||
***********************
|
||||
*** GW@DFT ***
|
||||
***********************
|
||||
***********************
|
||||
*** GW ***
|
||||
***********************
|
||||
|
||||
|
||||
Memory Information
|
||||
------------------
|
||||
Available GA per process is 999.9 MBs
|
||||
Available GA per process is 1000.0 MBs
|
||||
Available stack per process is 500.0 MBs
|
||||
Available heap per process is 499.9 MBs
|
||||
|
||||
Peak memory in Initialization
|
||||
Peak GA per process is 5.6 MBs
|
||||
Peak GA per process is 1.6 MBs
|
||||
Peak stack per process is 0.1 MBs
|
||||
Peak heap per process is 6.5 MBs
|
||||
Peak heap per process is 2.9 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Peak memory in CDGW Calculation
|
||||
Peak GA per process is 7.0 MBs
|
||||
Peak GA per process is 2.0 MBs
|
||||
Peak stack per process is 2.7 MBs
|
||||
Peak heap per process is 2.1 MBs
|
||||
Peak heap per process is 1.3 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Computing 2-center integrals
|
||||
Computing Inverse Cholesky factor
|
||||
Computing 3-center integrals
|
||||
Transform ERIs
|
||||
Distribute ERIs
|
||||
Read inverse
|
||||
Orthogonalize ERIs
|
||||
Computing Vxc
|
||||
Computing Sigma_x
|
||||
Computing W(iw) on the imaginary grid
|
||||
|
|
@ -904,28 +881,28 @@ task dft gw
|
|||
Alpha Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 1 -695.048 0.000
|
||||
@GW 2 -694.653 0.000
|
||||
@GW 3 -694.648 0.000
|
||||
@GW 4 -538.987 0.000
|
||||
@GW 5 -408.010 0.000
|
||||
@GW 6 -300.254 0.000
|
||||
@GW 7 -296.783 0.000
|
||||
@GW 1 -695.046 0.002
|
||||
@GW 2 -694.650 0.002
|
||||
@GW 3 -694.645 0.002
|
||||
@GW 4 -538.980 0.006
|
||||
@GW 5 -408.006 0.001
|
||||
@GW 6 -300.252 0.002
|
||||
@GW 7 -296.781 0.002
|
||||
|
||||
GW Timing Statistics
|
||||
------------------------
|
||||
Init. : 0.1s
|
||||
ERIs : 0.7s
|
||||
ERIs : 0.2s
|
||||
Distr.: 0.0s
|
||||
Vxc : 4.9s
|
||||
Vxc : 0.5s
|
||||
Vmo : 0.0s
|
||||
iW : 2.8s
|
||||
iW : 0.9s
|
||||
I_n : 0.0s
|
||||
R_n : 1.1s
|
||||
GW total: 9.6s
|
||||
R_n : 0.2s
|
||||
GW total: 2.0s
|
||||
|
||||
|
||||
Task times cpu: 51.4s wall: 51.6s
|
||||
Task times cpu: 7.6s wall: 7.6s
|
||||
|
||||
|
||||
NWChem Input Module
|
||||
|
|
@ -933,8 +910,8 @@ task dft gw
|
|||
|
||||
|
||||
|
||||
NWChem GW@DFT Module
|
||||
--------------------
|
||||
NWChem GW Module
|
||||
----------------
|
||||
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
|
@ -1034,34 +1011,34 @@ task dft gw
|
|||
|
||||
Loading old vectors from job with title :
|
||||
|
||||
GW Trifluoroacetamide
|
||||
W vals
|
||||
|
||||
Time after variat. SCF: 51.9
|
||||
Time after variat. SCF: 8.0
|
||||
|
||||
3 Center 2 Electron Integral Information
|
||||
----------------------------------------
|
||||
Maximum number of 3-center 2e- integrals is: 4898880.
|
||||
This is reduced with Schwarz screening to: 3084900.
|
||||
Incore requires a per proc buffer size of: 1377181.
|
||||
Incore requires a per proc buffer size of: 398581.
|
||||
Minimum dble words available (all nodes) is: 131060988
|
||||
This is reduced (for later use) to: 130586152
|
||||
proc 0 Suggested buffer size is: 1377181
|
||||
Max Suggested buffer size is: 1377181
|
||||
This is reduced (for later use) to: 130644472
|
||||
proc 0 Suggested buffer size is: 398581
|
||||
Max Suggested buffer size is: 398581
|
||||
no. integral batches is: 1000
|
||||
|
||||
Incore memory use for 3-center 2e- integrals is turned off.
|
||||
Time prior to 1st pass: 51.9
|
||||
Time prior to 1st pass: 8.0
|
||||
|
||||
|
||||
Total DFT energy = -506.483320618786
|
||||
One electron energy = -1365.176616645610
|
||||
Coulomb energy = 581.766179275143
|
||||
Exchange-Corr. energy = -58.042783022072
|
||||
Total DFT energy = -506.483320616637
|
||||
One electron energy = -1365.176616647789
|
||||
Coulomb energy = 581.766179277314
|
||||
Exchange-Corr. energy = -58.042783019915
|
||||
Nuclear repulsion energy = 334.969899773753
|
||||
|
||||
Numeric. integr. density = 56.000003720190
|
||||
Numeric. integr. density = 56.000003716815
|
||||
|
||||
Total iterative time = 5.9s
|
||||
Total iterative time = 0.7s
|
||||
|
||||
|
||||
|
||||
|
|
@ -1094,32 +1071,36 @@ GW Trifluoroacetamide
|
|||
2 0 0 2 -28.307350 -52.360374 -52.360374 76.413398
|
||||
|
||||
|
||||
***********************
|
||||
*** GW@DFT ***
|
||||
***********************
|
||||
***********************
|
||||
*** GW ***
|
||||
***********************
|
||||
|
||||
|
||||
Memory Information
|
||||
------------------
|
||||
Available GA per process is 999.9 MBs
|
||||
Available GA per process is 1000.0 MBs
|
||||
Available stack per process is 500.0 MBs
|
||||
Available heap per process is 499.9 MBs
|
||||
|
||||
Peak memory in Initialization
|
||||
Peak GA per process is 5.6 MBs
|
||||
Peak GA per process is 1.6 MBs
|
||||
Peak stack per process is 0.1 MBs
|
||||
Peak heap per process is 6.5 MBs
|
||||
Peak heap per process is 2.9 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Peak memory in CDGW Calculation
|
||||
Peak GA per process is 7.0 MBs
|
||||
Peak GA per process is 2.0 MBs
|
||||
Peak stack per process is 2.7 MBs
|
||||
Peak heap per process is 1.5 MBs
|
||||
Peak heap per process is 1.2 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Computing 2-center integrals
|
||||
Computing Inverse Cholesky factor
|
||||
Computing 3-center integrals
|
||||
Transform ERIs
|
||||
Distribute ERIs
|
||||
Read inverse
|
||||
Orthogonalize ERIs
|
||||
Computing Vxc
|
||||
Computing Sigma_x
|
||||
Computing W(iw) on the imaginary grid
|
||||
|
|
@ -1128,249 +1109,25 @@ GW Trifluoroacetamide
|
|||
Alpha Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 4 -538.987 0.000
|
||||
@GW 5 -408.010 0.000
|
||||
@GW 6 -300.254 0.000
|
||||
@GW 7 -296.783 0.000
|
||||
@GW 4 -538.980 0.006
|
||||
@GW 5 -408.006 0.001
|
||||
@GW 6 -300.252 0.002
|
||||
@GW 7 -296.781 0.002
|
||||
|
||||
GW Timing Statistics
|
||||
------------------------
|
||||
Init. : 0.1s
|
||||
ERIs : 0.8s
|
||||
ERIs : 0.2s
|
||||
Distr.: 0.0s
|
||||
Vxc : 4.9s
|
||||
Vxc : 0.5s
|
||||
Vmo : 0.0s
|
||||
iW : 1.6s
|
||||
iW : 0.9s
|
||||
I_n : 0.0s
|
||||
R_n : 0.6s
|
||||
GW total: 8.0s
|
||||
R_n : 0.1s
|
||||
GW total: 1.8s
|
||||
|
||||
|
||||
Task times cpu: 14.1s wall: 14.1s
|
||||
|
||||
|
||||
NWChem Input Module
|
||||
-------------------
|
||||
|
||||
|
||||
|
||||
NWChem GW@DFT Module
|
||||
--------------------
|
||||
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
||||
|
||||
|
||||
NWChem DFT Module
|
||||
-----------------
|
||||
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
||||
|
||||
|
||||
|
||||
Summary of "ao basis" -> "ao basis" (spherical)
|
||||
------------------------------------------------------------------------------
|
||||
Tag Description Shells Functions and Types
|
||||
---------------- ------------------------------ ------ ---------------------
|
||||
F cc-pvdz 6 14 3s2p1d
|
||||
O cc-pvdz 6 14 3s2p1d
|
||||
N cc-pvdz 6 14 3s2p1d
|
||||
C cc-pvdz 6 14 3s2p1d
|
||||
H cc-pvdz 3 5 2s1p
|
||||
|
||||
|
||||
Caching 1-el integrals
|
||||
|
||||
General Information
|
||||
-------------------
|
||||
SCF calculation type: DFT
|
||||
Wavefunction type: closed shell.
|
||||
No. of atoms : 9
|
||||
No. of electrons : 56
|
||||
Alpha electrons : 28
|
||||
Beta electrons : 28
|
||||
Charge : 0
|
||||
Spin multiplicity: 1
|
||||
Use of symmetry is: off; symmetry adaption is: off
|
||||
Maximum number of iterations: 50
|
||||
This is a Direct SCF calculation.
|
||||
AO basis - number of functions: 108
|
||||
number of shells: 48
|
||||
A Charge density fitting basis will be used.
|
||||
CD basis - number of functions: 420
|
||||
number of shells: 138
|
||||
Convergence on energy requested: 1.00D-06
|
||||
Convergence on density requested: 1.00D-05
|
||||
Convergence on gradient requested: 5.00D-04
|
||||
|
||||
XC Information
|
||||
--------------
|
||||
Hartree-Fock (Exact) Exchange 0.450
|
||||
PerdewBurkeErnzerhof Exchange Functional 0.550
|
||||
Perdew 1991 LDA Correlation Functional 1.000 local
|
||||
PerdewBurkeErnz. Correlation Functional 1.000 non-local
|
||||
|
||||
Grid Information
|
||||
----------------
|
||||
Grid used for XC integration: medium
|
||||
Radial quadrature: Mura-Knowles
|
||||
Angular quadrature: Lebedev.
|
||||
Tag B.-S. Rad. Rad. Pts. Rad. Cut. Ang. Pts.
|
||||
--- ---------- --------- --------- ---------
|
||||
F 0.50 49 11.0 434
|
||||
O 0.60 49 9.0 434
|
||||
N 0.65 49 8.0 434
|
||||
C 0.70 49 10.0 434
|
||||
H 0.35 45 10.0 434
|
||||
Grid pruning is: on
|
||||
Number of quadrature shells: 433
|
||||
Spatial weights used: Erf1
|
||||
|
||||
Convergence Information
|
||||
-----------------------
|
||||
Convergence aids based upon iterative change in
|
||||
total energy or number of iterations.
|
||||
Levelshifting, if invoked, occurs when the
|
||||
HOMO/LUMO gap drops below (HL_TOL): 1.00D-02
|
||||
DIIS, if invoked, will attempt to extrapolate
|
||||
using up to (NFOCK): 10 stored Fock matrices.
|
||||
|
||||
Damping( 0%) Levelshifting(0.5) DIIS
|
||||
--------------- ------------------- ---------------
|
||||
dE on: start ASAP start
|
||||
dE off: 2 iters 50 iters 50 iters
|
||||
|
||||
|
||||
Screening Tolerance Information
|
||||
-------------------------------
|
||||
Density screening/tol_rho: 1.00D-10
|
||||
AO Gaussian exp screening on grid/accAOfunc: 14
|
||||
CD Gaussian exp screening on grid/accCDfunc: 20
|
||||
XC Gaussian exp screening on grid/accXCfunc: 20
|
||||
Schwarz screening/accCoul: 1.00D-12
|
||||
|
||||
|
||||
Loading old vectors from job with title :
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
||||
Time after variat. SCF: 66.0
|
||||
|
||||
3 Center 2 Electron Integral Information
|
||||
----------------------------------------
|
||||
Maximum number of 3-center 2e- integrals is: 4898880.
|
||||
This is reduced with Schwarz screening to: 3084900.
|
||||
Incore requires a per proc buffer size of: 1377181.
|
||||
Minimum dble words available (all nodes) is: 131060988
|
||||
This is reduced (for later use) to: 130586152
|
||||
proc 0 Suggested buffer size is: 1377181
|
||||
Max Suggested buffer size is: 1377181
|
||||
no. integral batches is: 1000
|
||||
|
||||
Incore memory use for 3-center 2e- integrals is turned off.
|
||||
Time prior to 1st pass: 66.0
|
||||
|
||||
|
||||
Total DFT energy = -506.483320619392
|
||||
One electron energy = -1365.176616645610
|
||||
Coulomb energy = 581.766179275142
|
||||
Exchange-Corr. energy = -58.042783022677
|
||||
Nuclear repulsion energy = 334.969899773753
|
||||
|
||||
Numeric. integr. density = 56.000003721840
|
||||
|
||||
Total iterative time = 5.9s
|
||||
|
||||
|
||||
|
||||
center of mass
|
||||
--------------
|
||||
x = 0.05046829 y = 0.11377991 z = 0.00004386
|
||||
|
||||
moments of inertia (a.u.)
|
||||
------------------
|
||||
724.808949354496 0.000000000000 0.000000000000
|
||||
0.000000000000 484.336918953950 -0.000000000000
|
||||
0.000000000000 -0.000000000000 886.538753836127
|
||||
|
||||
Multipole analysis of the density
|
||||
---------------------------------
|
||||
|
||||
L x y z total alpha beta nuclear
|
||||
- - - - ----- ----- ---- -------
|
||||
0 0 0 0 -0.000000 -28.000000 -28.000000 56.000000
|
||||
|
||||
1 1 0 0 -1.489094 -0.744547 -0.744547 -0.000000
|
||||
1 0 1 0 -0.546971 -0.273486 -0.273486 0.000000
|
||||
1 0 0 1 0.000605 0.000302 0.000302 0.000000
|
||||
|
||||
2 2 0 0 -27.181301 -96.832025 -96.832025 166.482750
|
||||
2 1 1 0 4.886093 -1.980672 -1.980672 8.847437
|
||||
2 1 0 1 0.002838 -0.001223 -0.001223 0.005284
|
||||
2 0 2 0 -26.258758 -158.160655 -158.160655 290.062553
|
||||
2 0 1 1 -0.001360 -0.006340 -0.006340 0.011320
|
||||
2 0 0 2 -28.307350 -52.360374 -52.360374 76.413398
|
||||
|
||||
|
||||
***********************
|
||||
*** GW@DFT ***
|
||||
***********************
|
||||
|
||||
|
||||
Memory Information
|
||||
------------------
|
||||
Available GA per process is 999.9 MBs
|
||||
Available stack per process is 500.0 MBs
|
||||
Available heap per process is 499.9 MBs
|
||||
|
||||
Peak memory in Initialization
|
||||
Peak GA per process is 5.6 MBs
|
||||
Peak stack per process is 0.1 MBs
|
||||
Peak heap per process is 6.5 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Peak memory in AFF Calculation
|
||||
Peak GA per process is 27.6 MBs
|
||||
Peak stack per process is 38.3 MBs
|
||||
Peak heap per process is 0.0 MBs
|
||||
Memory not in MA per process 38.3 MBs
|
||||
|
||||
Computing 2-center integrals
|
||||
Computing Inverse Cholesky factor
|
||||
Computing 3-center integrals
|
||||
Computing V_xc ...
|
||||
Computing Sigma_x
|
||||
Computing RPA polarizability
|
||||
Solving quasiparticle equations
|
||||
|
||||
Alpha Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 1 -695.048 0.000
|
||||
@GW 2 -694.652 0.000
|
||||
@GW 3 -694.648 0.000
|
||||
@GW 4 -538.986 0.000
|
||||
@GW 5 -408.009 0.000
|
||||
@GW 6 -300.254 0.000
|
||||
@GW 7 -296.783 0.000
|
||||
|
||||
GW Timing Statistics
|
||||
------------------------
|
||||
Init. : 0.1s
|
||||
ERIs : 0.7s
|
||||
Distr.: 0.0s
|
||||
Vxc : 4.7s
|
||||
Sigma_x : 0.0s
|
||||
RPA : 2.6s
|
||||
Wmn : 0.1s
|
||||
Sigma_c : 0.0s
|
||||
GW total: 8.4s
|
||||
|
||||
|
||||
Task times cpu: 14.3s wall: 14.4s
|
||||
Task times cpu: 2.6s wall: 2.6s
|
||||
|
||||
|
||||
NWChem Input Module
|
||||
|
|
@ -1390,11 +1147,11 @@ MA usage statistics:
|
|||
heap stack
|
||||
---- -----
|
||||
current number of blocks 0 0
|
||||
maximum number of blocks 28 56
|
||||
maximum number of blocks 28 62
|
||||
current total bytes 0 0
|
||||
maximum total bytes 25980280 80584200
|
||||
maximum total K-bytes 25981 80585
|
||||
maximum total M-bytes 26 81
|
||||
maximum total bytes 22153480 30740728
|
||||
maximum total K-bytes 22154 30741
|
||||
maximum total M-bytes 23 31
|
||||
|
||||
|
||||
CITATION
|
||||
|
|
@ -1430,24 +1187,25 @@ MA usage statistics:
|
|||
|
||||
AUTHORS
|
||||
-------
|
||||
E. Apra, E. J. Bylaska, N. Govind, K. Kowalski, M. Valiev, W. A. de Jong,
|
||||
T. P. Straatsma, H. J. J. van Dam, D. Wang, T. L. Windus, N. P. Bauman,
|
||||
A. Panyala, J. Hammond, J. Autschbach, K. Bhaskaran-Nair, J. Brabec,
|
||||
K. Lopata, S. A. Fischer, S. Krishnamoorthy, M. Jacquelin, W. Ma, M. Klemm,
|
||||
O. Villa, Y. Chen, V. Anisimov, F. Aquino, S. Hirata, M. T. Hackler,
|
||||
Eric Hermes, L. Jensen, J. E. Moore, J. C. Becca, V. Konjkov,
|
||||
D. Mejia-Rodriguez, T. Risthaus, M. Malagoli, A. Marenich,
|
||||
A. Otero-de-la-Roza, J. Mullin, P. Nichols, R. Peverati, J. Pittner, Y. Zhao,
|
||||
P.-D. Fan, A. Fonari, M. J. Williamson, R. J. Harrison, J. R. Rehr,
|
||||
M. Dupuis, D. Silverstein, D. M. A. Smith, J. Nieplocha, V. Tipparaju,
|
||||
M. Krishnan, B. E. Van Kuiken, A. Vazquez-Mayagoitia, M. Swart, Q. Wu,
|
||||
T. Van Voorhis, A. A. Auer, M. Nooijen, L. D. Crosby, E. Brown, G. Cisneros,
|
||||
G. I. Fann, H. Fruchtl, J. Garza, K. Hirao, R. A. Kendall, J. A. Nichols,
|
||||
K. Tsemekhman, K. Wolinski, J. Anchell, D. E. Bernholdt, P. Borowski,
|
||||
T. Clark, D. Clerc, H. Dachsel, M. J. O. Deegan, K. Dyall, D. Elwood,
|
||||
E. Glendening, M. Gutowski, A. C. Hess, J. Jaffe, B. G. Johnson, J. Ju,
|
||||
R. Kobayashi, R. Kutteh, Z. Lin, R. Littlefield, X. Long, B. Meng,
|
||||
T. Nakajima, S. Niu, L. Pollack, M. Rosing, K. Glaesemann, G. Sandrone,
|
||||
M. Stave, H. Taylor, G. Thomas, J. H. van Lenthe, A. T. Wong, Z. Zhang.
|
||||
E. Apra, E. J. Bylaska, N. Govind, K. Kowalski, D. Mejia-Rodriguez,
|
||||
N. P. Bauman, A. Panyala, R. J. Harrison, M. Valiev, W. A. de Jong,
|
||||
T. P. Straatsma, H. J. J. van Dam, D. Wang, T. L. Windus, J. Hammond,
|
||||
J. Autschbach, R. de P. Soares, A. Kunitsa, K. Bhaskaran-Nair, J. Brabec,
|
||||
K. Lopata, S. A. Fischer, S. Krishnamoorthy, M. Jacquelin, R. Goswami,
|
||||
A. Woods, W. Ma, M. Klemm, O. Villa, Y. Chen, V. Anisimov, F. Aquino,
|
||||
S. Hirata, M. T. Hackler, E. Hermes, L. Jensen, J. E. Moore, J. C. Becca,
|
||||
V. Konjkov, T. Risthaus, M. Malagoli, A. Marenich, A. Otero-de-la-Roza,
|
||||
J. Mullin, P. Nichols, R. Peverati, J. Pittner, Y. Zhao, P.-D. Fan,
|
||||
A. Fonari, M. J. Williamson, J. R. Rehr, M. Dupuis, D. Silverstein,
|
||||
D. M. A. Smith, J. Nieplocha, V. Tipparaju, M. Krishnan, B. E. Van Kuiken,
|
||||
A. Vazquez-Mayagoitia, M. Swart, Q. Wu, T. Van Voorhis, A. A. Auer,
|
||||
M. Nooijen, L. D. Crosby, E. Brown, G. Cisneros, G. I. Fann, H. Fruchtl,
|
||||
J. Garza, K. Hirao, R. A. Kendall, J. A. Nichols, K. Tsemekhman, K. Wolinski,
|
||||
J. Anchell, D. E. Bernholdt, P. Borowski, T. Clark, D. Clerc, H. Dachsel,
|
||||
M. J. O. Deegan, K. Dyall, D. Elwood, E. Glendening, M. Gutowski, A. C. Hess,
|
||||
J. Jaffe, B. G. Johnson, J. Ju, R. Kobayashi, R. Kutteh, Z. Lin,
|
||||
R. Littlefield, X. Long, B. Meng, T. Nakajima, S. Niu, L. Pollack, M. Rosing,
|
||||
K. Glaesemann, G. Sandrone, M. Stave, H. Taylor, G. Thomas, J. H. van Lenthe,
|
||||
A. T. Wong, Z. Zhang.
|
||||
|
||||
Total times cpu: 79.8s wall: 80.2s
|
||||
Total times cpu: 10.5s wall: 10.5s
|
||||
|
|
|
|||
|
|
@ -24,35 +24,18 @@ end
|
|||
dft
|
||||
odft
|
||||
mult 2
|
||||
xc pbe96
|
||||
xc pbe0
|
||||
direct
|
||||
grid fine nodisk
|
||||
tolerances accCoul 14
|
||||
tolerances accCoul 12
|
||||
convergence energy 1d-8
|
||||
noprint "final vectors analysis"
|
||||
end
|
||||
|
||||
gw
|
||||
eta 0.001
|
||||
method cdgw
|
||||
solver newton 20
|
||||
states alpha occ 5 vir 5
|
||||
states beta occ 5 vir 5
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
||||
dft
|
||||
noscf
|
||||
end
|
||||
|
||||
gw
|
||||
eta 0.001
|
||||
solver newton 20
|
||||
states alpha occ 5 vir 5
|
||||
states beta occ 5 vir 5
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
|
|
|||
|
|
@ -1,4 +1,5 @@
|
|||
argument 1 = gw_openshell.nw
|
||||
argument 1 = /data/edo/nwchem/nwchem_bse-pull//QA/tests/gw_openshell/gw_openshell.nw
|
||||
NWChem w/ OpenMP: maximum threads = 1
|
||||
|
||||
|
||||
|
||||
|
|
@ -17,11 +18,11 @@ H 2.5025 -1.2628 -0.0022
|
|||
H 0.9511 -2.0241 -0.0008
|
||||
end
|
||||
|
||||
basis "ao basis" spherical
|
||||
basis "ao basis" bse spherical
|
||||
* library cc-pvdz
|
||||
end
|
||||
|
||||
basis "cd basis" spherical
|
||||
basis "cd basis" bse spherical
|
||||
* library cc-pvdz-rifit
|
||||
end
|
||||
|
||||
|
|
@ -29,35 +30,18 @@ end
|
|||
dft
|
||||
odft
|
||||
mult 2
|
||||
xc pbe96
|
||||
xc pbe0
|
||||
direct
|
||||
grid fine nodisk
|
||||
tolerances accCoul 14
|
||||
tolerances accCoul 12
|
||||
convergence energy 1d-8
|
||||
noprint "final vectors analysis"
|
||||
end
|
||||
|
||||
gw
|
||||
eta 0.001
|
||||
method cdgw
|
||||
solver newton 20
|
||||
states alpha occ 5 vir 5
|
||||
states beta occ 5 vir 5
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
||||
dft
|
||||
noscf
|
||||
end
|
||||
|
||||
gw
|
||||
eta 0.001
|
||||
solver newton 20
|
||||
states alpha occ 5 vir 5
|
||||
states beta occ 5 vir 5
|
||||
convergence 0.001 ev
|
||||
end
|
||||
|
||||
task dft gw
|
||||
|
|
@ -69,7 +53,7 @@ task dft gw
|
|||
|
||||
|
||||
|
||||
Northwest Computational Chemistry Package (NWChem) 7.0.1
|
||||
Northwest Computational Chemistry Package (NWChem) 7.3.0
|
||||
--------------------------------------------------------
|
||||
|
||||
|
||||
|
|
@ -77,7 +61,7 @@ task dft gw
|
|||
Pacific Northwest National Laboratory
|
||||
Richland, WA 99352
|
||||
|
||||
Copyright (c) 1994-2020
|
||||
Copyright (c) 1994-2025
|
||||
Pacific Northwest National Laboratory
|
||||
Battelle Memorial Institute
|
||||
|
||||
|
|
@ -102,21 +86,21 @@ task dft gw
|
|||
Job information
|
||||
---------------
|
||||
|
||||
hostname = dmejiar-VirtualBox
|
||||
program = nwchem
|
||||
date = Fri Feb 4 10:51:34 2022
|
||||
hostname = mica
|
||||
program = /data/edo/nwchem/nwchem_bse-pull//bin/LINUX64/nwchem
|
||||
date = Thu Oct 16 12:25:05 2025
|
||||
|
||||
compiled = Fri_Feb_04_04:17:52_2022
|
||||
source = /home/dmejiar/Software/nwchem_gfortran_gw
|
||||
nwchem branch = 7.0.0
|
||||
nwchem revision = nwchem_on_git-3497-g0c7e480c37
|
||||
ga revision = 5.8.1
|
||||
compiled = Thu_Oct_16_12:11:32_2025
|
||||
source = /data/edo/nwchem/nwchem_bse-pull/
|
||||
nwchem branch = 7.3.0
|
||||
nwchem revision = v7.2.0-beta1-1593-g3a8d752956
|
||||
ga revision = 5.9.2
|
||||
use scalapack = T
|
||||
input = gw_openshell.nw
|
||||
input = /data/edo/nwchem/nwchem_bse-pull//QA/tests/gw_openshell/gw_openshell.nw
|
||||
prefix = gw_openshell.
|
||||
data base = ./gw_openshell.db
|
||||
status = startup
|
||||
nproc = 2
|
||||
nproc = 1
|
||||
time left = -1s
|
||||
|
||||
|
||||
|
|
@ -124,10 +108,10 @@ task dft gw
|
|||
Memory information
|
||||
------------------
|
||||
|
||||
heap = 65535996 doubles = 500.0 Mbytes
|
||||
stack = 65536001 doubles = 500.0 Mbytes
|
||||
global = 131072000 doubles = 1000.0 Mbytes (distinct from heap & stack)
|
||||
total = 262143997 doubles = 2000.0 Mbytes
|
||||
heap = 65535996 doubles = 500.0 Mbytes
|
||||
stack = 65536001 doubles = 500.0 Mbytes
|
||||
global = 131072000 doubles = 1000.0 Mbytes (distinct from heap & stack)
|
||||
total = 262143997 doubles = 2000.0 Mbytes
|
||||
verify = yes
|
||||
hardfail = no
|
||||
|
||||
|
|
@ -251,9 +235,6 @@ task dft gw
|
|||
|
||||
|
||||
|
||||
library name resolved from: environment
|
||||
library file name is: </home/dmejiar/Software/nwchem_gfortran_gw/src/basis/libraries.bse/>
|
||||
|
||||
|
||||
|
||||
Summary of "ao basis" -> "" (spherical)
|
||||
|
|
@ -263,9 +244,6 @@ task dft gw
|
|||
* cc-pvdz on all atoms
|
||||
|
||||
|
||||
library name resolved from: environment
|
||||
library file name is: </home/dmejiar/Software/nwchem_gfortran_gw/src/basis/libraries.bse/>
|
||||
|
||||
|
||||
|
||||
Summary of "cd basis" -> "" (spherical)
|
||||
|
|
@ -276,8 +254,8 @@ task dft gw
|
|||
|
||||
|
||||
|
||||
NWChem GW@DFT Module
|
||||
--------------------
|
||||
NWChem GW Module
|
||||
----------------
|
||||
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
|
@ -603,8 +581,9 @@ task dft gw
|
|||
|
||||
XC Information
|
||||
--------------
|
||||
PBE96 Method XC Functional
|
||||
PerdewBurkeErnzerhof Exchange Functional 1.000
|
||||
PBE0 Method XC Functional
|
||||
Hartree-Fock (Exact) Exchange 0.250
|
||||
PerdewBurkeErnzerhof Exchange Functional 0.750
|
||||
Perdew 1991 LDA Correlation Functional 1.000 local
|
||||
PerdewBurkeErnz. Correlation Functional 1.000 non-local
|
||||
|
||||
|
|
@ -644,7 +623,7 @@ task dft gw
|
|||
AO Gaussian exp screening on grid/accAOfunc: 18
|
||||
CD Gaussian exp screening on grid/accCDfunc: 20
|
||||
XC Gaussian exp screening on grid/accXCfunc: 20
|
||||
Schwarz screening/accCoul: 1.00D-14
|
||||
Schwarz screening/accCoul: 1.00D-12
|
||||
|
||||
|
||||
Superposition of Atomic Density Guess
|
||||
|
|
@ -661,21 +640,21 @@ task dft gw
|
|||
HOMO = -0.154097
|
||||
LUMO = 0.070838
|
||||
|
||||
Time after variat. SCF: 0.6
|
||||
Time after variat. SCF: 0.7
|
||||
|
||||
3 Center 2 Electron Integral Information
|
||||
----------------------------------------
|
||||
Maximum number of 3-center 2e- integrals is: 529984.
|
||||
This is reduced with Schwarz screening to: 367500.
|
||||
Incore requires a per proc buffer size of: 175029.
|
||||
Incore requires a per proc buffer size of: 327517.
|
||||
Minimum dble words available (all nodes) is: 131065476
|
||||
This is reduced (for later use) to: 130813264
|
||||
proc 0 Suggested buffer size is: 175029
|
||||
Max Suggested buffer size is: 175029
|
||||
This is reduced (for later use) to: 130794336
|
||||
proc 0 Suggested buffer size is: 327517
|
||||
Max Suggested buffer size is: 327517
|
||||
no. integral batches is: 1000
|
||||
|
||||
Incore memory use for 3-center 2e- integrals is turned off.
|
||||
Time prior to 1st pass: 0.6
|
||||
Time prior to 1st pass: 0.7
|
||||
|
||||
Memory utilization after 1st SCF pass:
|
||||
Heap Space remaining (MW): 65.53 65530108
|
||||
|
|
@ -683,42 +662,42 @@ task dft gw
|
|||
|
||||
convergence iter energy DeltaE RMS-Dens Diis-err time
|
||||
---------------- ----- ----------------- --------- --------- --------- ------
|
||||
d= 0,ls=0.0,diis 1 -168.9561121185 -2.32D+02 2.05D-02 2.18D-01 1.5
|
||||
1.84D-02 2.17D-01
|
||||
d= 0,ls=0.0,diis 2 -167.9504224824 1.01D+00 1.76D-02 2.63D+00 2.5
|
||||
1.47D-02 2.33D+00
|
||||
d= 0,ls=0.0,diis 3 -169.0537934402 -1.10D+00 3.84D-03 2.91D-02 3.3
|
||||
2.75D-03 2.86D-02
|
||||
d= 0,ls=0.0,diis 4 -169.0564442937 -2.65D-03 2.06D-03 2.17D-02 4.1
|
||||
1.46D-03 2.00D-02
|
||||
d= 0,ls=0.0,diis 5 -169.0660484759 -9.60D-03 2.88D-04 5.93D-04 4.8
|
||||
2.45D-04 5.80D-04
|
||||
d= 0,ls=0.0,diis 1 -168.9657539780 -2.32D+02 1.39D-02 1.93D-01 1.6
|
||||
1.31D-02 1.92D-01
|
||||
d= 0,ls=0.0,diis 2 -168.6178261291 3.48D-01 1.00D-02 1.13D+00 2.5
|
||||
9.06D-03 1.06D+00
|
||||
d= 0,ls=0.0,diis 3 -169.0613313213 -4.44D-01 1.36D-03 7.44D-03 3.5
|
||||
1.32D-03 8.46D-03
|
||||
d= 0,ls=0.0,diis 4 -169.0630054782 -1.67D-03 9.29D-04 4.19D-03 4.5
|
||||
5.76D-04 3.46D-03
|
||||
d= 0,ls=0.0,diis 5 -169.0646959497 -1.69D-03 1.55D-04 1.38D-04 5.5
|
||||
1.31D-04 1.63D-04
|
||||
Resetting Diis
|
||||
d= 0,ls=0.0,diis 6 -169.0663002306 -2.52D-04 1.01D-04 2.23D-05 5.6
|
||||
6.70D-05 2.22D-05
|
||||
d= 0,ls=0.0,diis 7 -169.0663113153 -1.11D-05 8.27D-05 2.77D-06 6.5
|
||||
5.41D-05 1.73D-06
|
||||
d= 0,ls=0.0,diis 8 -169.0663035508 7.76D-06 4.90D-05 2.02D-05 7.2
|
||||
4.55D-05 2.15D-05
|
||||
d= 0,ls=0.0,diis 9 -169.0663123641 -8.81D-06 1.02D-05 3.98D-07 8.1
|
||||
7.29D-06 3.32D-07
|
||||
d= 0,ls=0.0,diis 10 -169.0663125308 -1.67D-07 2.91D-06 3.07D-08 8.9
|
||||
2.57D-06 3.01D-08
|
||||
d= 0,ls=0.0,diis 11 -169.0663125423 -1.15D-08 1.15D-06 1.20D-08 9.6
|
||||
1.36D-06 1.15D-08
|
||||
d= 0,ls=0.0,diis 12 -169.0663125471 -4.87D-09 6.52D-07 1.79D-09 10.4
|
||||
2.64D-07 1.06D-09
|
||||
d= 0,ls=0.0,diis 6 -169.0647672024 -7.13D-05 4.96D-05 7.61D-06 6.5
|
||||
5.91D-05 7.90D-06
|
||||
d= 0,ls=0.0,diis 7 -169.0647730165 -5.81D-06 5.03D-05 2.16D-06 7.5
|
||||
1.81D-05 8.25D-07
|
||||
d= 0,ls=0.0,diis 8 -169.0647729683 4.81D-08 1.78D-05 2.36D-06 8.4
|
||||
2.14D-05 3.64D-06
|
||||
d= 0,ls=0.0,diis 9 -169.0647741451 -1.18D-06 9.22D-06 5.57D-07 9.4
|
||||
4.55D-06 2.30D-07
|
||||
d= 0,ls=0.0,diis 10 -169.0647743526 -2.08D-07 9.00D-06 4.08D-08 10.4
|
||||
8.19D-06 2.42D-08
|
||||
d= 0,ls=0.0,diis 11 -169.0647744288 -7.62D-08 2.13D-06 7.37D-09 11.4
|
||||
1.35D-06 1.44D-09
|
||||
d= 0,ls=0.0,diis 12 -169.0647744339 -5.12D-09 7.95D-07 7.61D-10 12.4
|
||||
6.20D-07 8.83D-10
|
||||
|
||||
|
||||
Total DFT energy = -169.066312547124
|
||||
One electron energy = -355.393819271438
|
||||
Coulomb energy = 144.424914623888
|
||||
Exchange-Corr. energy = -21.405786398232
|
||||
Total DFT energy = -169.064774433938
|
||||
One electron energy = -355.358774140560
|
||||
Coulomb energy = 144.394070505297
|
||||
Exchange-Corr. energy = -21.408449297333
|
||||
Nuclear repulsion energy = 63.308378498658
|
||||
|
||||
Numeric. integr. density = 22.999999954116
|
||||
Numeric. integr. density = 22.999999950854
|
||||
|
||||
Total iterative time = 9.8s
|
||||
Total iterative time = 11.7s
|
||||
|
||||
|
||||
|
||||
|
|
@ -728,27 +707,27 @@ task dft gw
|
|||
|
||||
alpha 1 2 3 4 5 6 7 8 9 10
|
||||
beta 1 2 3 4 5 6 7 8 9 10
|
||||
overlap 1.000 1.000 1.000 1.000 1.000 0.996 0.996 0.990 1.000 0.985
|
||||
overlap 1.000 1.000 1.000 1.000 0.999 0.993 0.991 0.974 1.000 0.968
|
||||
|
||||
|
||||
alpha 11 12 13 14 15 16 17 18 19 20
|
||||
beta 11 12 13 14 15 16 17 18 19 20
|
||||
overlap 1.000 0.993 0.999 0.994 0.995 0.999 0.999 0.952 1.000 0.945
|
||||
overlap 1.000 0.985 0.999 0.987 0.993 0.998 0.997 0.946 1.000 0.920
|
||||
|
||||
|
||||
alpha 21 22 23 24 25 26 27 28 29 30
|
||||
beta 21 22 23 24 25 26 27 28 29 30
|
||||
overlap 0.958 0.999 0.998 1.000 0.998 0.999 0.998 0.996 0.997 0.980
|
||||
beta 21 22 23 24 25 26 27 28 30 29
|
||||
overlap 0.928 0.997 0.998 1.000 0.997 0.999 0.996 0.993 0.996 0.977
|
||||
|
||||
|
||||
alpha 31 32 33 34 35 36 37 38 39 40
|
||||
beta 31 32 33 34 35 36 37 38 39 40
|
||||
overlap 0.985 0.996 0.991 0.992 0.999 1.000 0.997 0.994 0.997 0.999
|
||||
overlap 0.983 0.995 0.992 0.993 0.999 1.000 0.986 0.979 0.993 1.000
|
||||
|
||||
|
||||
alpha 41 42 43 44 45 46 47 48 49 50
|
||||
beta 41 42 43 44 45 46 47 48 49 50
|
||||
overlap 0.999 0.996 0.994 0.999 0.998 0.998 1.000 0.999 0.999 1.000
|
||||
overlap 0.999 0.999 0.992 0.994 0.998 0.997 1.000 0.998 0.999 1.000
|
||||
|
||||
|
||||
alpha 51 52
|
||||
|
|
@ -758,7 +737,7 @@ task dft gw
|
|||
--------------------------
|
||||
Expectation value of S2:
|
||||
--------------------------
|
||||
<S2> = 0.7519 (Exact = 0.7500)
|
||||
<S2> = 0.7537 (Exact = 0.7500)
|
||||
|
||||
|
||||
center of mass
|
||||
|
|
@ -778,21 +757,21 @@ task dft gw
|
|||
- - - - ----- ----- ---- -------
|
||||
0 0 0 0 -0.000000 -12.000000 -11.000000 23.000000
|
||||
|
||||
1 1 0 0 0.238427 0.541429 -0.303002 0.000000
|
||||
1 0 1 0 -1.321737 -0.969708 -0.352030 0.000000
|
||||
1 0 0 1 0.000663 -0.000054 0.000717 0.000000
|
||||
1 1 0 0 0.206401 0.547712 -0.341311 0.000000
|
||||
1 0 1 0 -1.390130 -1.013204 -0.376927 0.000000
|
||||
1 0 0 1 0.000816 0.000001 0.000815 0.000000
|
||||
|
||||
2 2 0 0 -11.884337 -13.068472 -9.699498 10.883633
|
||||
2 1 1 0 -0.250985 0.506239 0.213062 -0.970286
|
||||
2 1 0 1 0.006019 -0.000243 -0.000755 0.007017
|
||||
2 0 2 0 -11.275818 -52.448505 -49.703057 90.875744
|
||||
2 0 1 1 0.001863 0.000480 -0.000509 0.001891
|
||||
2 0 0 2 -13.100006 -6.900236 -6.199822 0.000052
|
||||
2 2 0 0 -11.728128 -12.959803 -9.651957 10.883633
|
||||
2 1 1 0 -0.304140 0.447969 0.218176 -0.970286
|
||||
2 1 0 1 0.006364 0.000093 -0.000746 0.007017
|
||||
2 0 2 0 -11.396501 -52.445939 -49.826305 90.875744
|
||||
2 0 1 1 0.001878 0.000424 -0.000437 0.001891
|
||||
2 0 0 2 -13.115703 -6.888717 -6.227038 0.000052
|
||||
|
||||
|
||||
***********************
|
||||
*** GW@DFT ***
|
||||
***********************
|
||||
***********************
|
||||
*** GW ***
|
||||
***********************
|
||||
|
||||
|
||||
Memory Information
|
||||
|
|
@ -802,15 +781,15 @@ task dft gw
|
|||
Available heap per process is 500.0 MBs
|
||||
|
||||
Peak memory in Initialization
|
||||
Peak GA per process is 1.4 MBs
|
||||
Peak GA per process is 2.7 MBs
|
||||
Peak stack per process is 0.0 MBs
|
||||
Peak heap per process is 1.6 MBs
|
||||
Peak heap per process is 2.9 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Peak memory in CDGW Calculation
|
||||
Peak GA per process is 1.7 MBs
|
||||
Peak GA per process is 3.4 MBs
|
||||
Peak stack per process is 0.6 MBs
|
||||
Peak heap per process is 1.5 MBs
|
||||
Peak heap per process is 1.9 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Computing 2-center integrals
|
||||
|
|
@ -824,323 +803,48 @@ task dft gw
|
|||
Alpha Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 8 -13.232 0.003
|
||||
@GW 9 -12.555 0.001
|
||||
@GW 10 -12.357 0.000
|
||||
@GW 11 -9.991 0.000
|
||||
@GW 12 -7.779 0.001
|
||||
@GW 13 3.732 0.000
|
||||
@GW 14 3.851 0.001
|
||||
@GW 15 5.314 0.000
|
||||
@GW 16 9.836 0.001
|
||||
@GW 17 11.891 0.000
|
||||
@GW 8 -14.622 0.000
|
||||
@GW 9 -13.641 0.004
|
||||
@GW 10 -12.648 0.005
|
||||
@GW 11 -10.354 0.002
|
||||
@GW 12 -8.453 0.002
|
||||
@GW 13 3.960 0.003
|
||||
@GW 14 3.965 0.005
|
||||
@GW 15 5.443 0.002
|
||||
@GW 16 10.094 0.004
|
||||
@GW 17 12.237 0.004
|
||||
|
||||
Beta Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 7 -15.994 0.001
|
||||
@GW 8 -14.090 0.000
|
||||
@GW 9 -12.312 0.000
|
||||
@GW 10 -11.111 0.000
|
||||
@GW 11 -9.813 0.000
|
||||
@GW 12 0.782 0.000
|
||||
@GW 13 3.891 0.001
|
||||
@GW 14 3.844 0.000
|
||||
@GW 15 5.494 0.001
|
||||
@GW 16 9.576 0.001
|
||||
@GW 7 -17.756 0.002
|
||||
@GW 8 -14.841 0.005
|
||||
@GW 9 -13.401 0.002
|
||||
@GW 10 -11.674 0.002
|
||||
@GW 11 -10.239 0.002
|
||||
@GW 12 1.142 0.002
|
||||
@GW 13 4.416 0.002
|
||||
@GW 14 4.181 0.005
|
||||
@GW 15 5.574 0.003
|
||||
@GW 16 10.343 0.017 *
|
||||
|
||||
* Result did not converge
|
||||
|
||||
|
||||
GW Timing Statistics
|
||||
------------------------
|
||||
Init. : 0.1s
|
||||
ERIs : 0.2s
|
||||
Init. : 0.0s
|
||||
ERIs : 0.1s
|
||||
Distr.: 0.0s
|
||||
Vxc : 0.7s
|
||||
Vxc : 0.8s
|
||||
Vmo : 0.0s
|
||||
iW : 0.7s
|
||||
iW : 0.5s
|
||||
I_n : 0.0s
|
||||
R_n : 0.3s
|
||||
GW total: 2.0s
|
||||
GW total: 1.8s
|
||||
|
||||
|
||||
Task times cpu: 12.2s wall: 12.3s
|
||||
|
||||
|
||||
NWChem Input Module
|
||||
-------------------
|
||||
|
||||
|
||||
|
||||
NWChem GW@DFT Module
|
||||
--------------------
|
||||
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
||||
|
||||
|
||||
NWChem DFT Module
|
||||
-----------------
|
||||
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
||||
|
||||
|
||||
|
||||
Summary of "ao basis" -> "ao basis" (spherical)
|
||||
------------------------------------------------------------------------------
|
||||
Tag Description Shells Functions and Types
|
||||
---------------- ------------------------------ ------ ---------------------
|
||||
O cc-pvdz 6 14 3s2p1d
|
||||
N cc-pvdz 6 14 3s2p1d
|
||||
C cc-pvdz 6 14 3s2p1d
|
||||
H cc-pvdz 3 5 2s1p
|
||||
|
||||
|
||||
Caching 1-el integrals
|
||||
|
||||
General Information
|
||||
-------------------
|
||||
SCF calculation type: DFT
|
||||
Wavefunction type: spin polarized.
|
||||
No. of atoms : 5
|
||||
No. of electrons : 23
|
||||
Alpha electrons : 12
|
||||
Beta electrons : 11
|
||||
Charge : 0
|
||||
Spin multiplicity: 2
|
||||
Use of symmetry is: off; symmetry adaption is: off
|
||||
Maximum number of iterations: 50
|
||||
This is a Direct SCF calculation.
|
||||
AO basis - number of functions: 52
|
||||
number of shells: 24
|
||||
A Charge density fitting basis will be used.
|
||||
CD basis - number of functions: 196
|
||||
number of shells: 66
|
||||
Convergence on energy requested: 1.00D-08
|
||||
Convergence on density requested: 1.00D-05
|
||||
Convergence on gradient requested: 5.00D-04
|
||||
|
||||
XC Information
|
||||
--------------
|
||||
PBE96 Method XC Functional
|
||||
PerdewBurkeErnzerhof Exchange Functional 1.000
|
||||
Perdew 1991 LDA Correlation Functional 1.000 local
|
||||
PerdewBurkeErnz. Correlation Functional 1.000 non-local
|
||||
|
||||
Grid Information
|
||||
----------------
|
||||
Grid used for XC integration: fine
|
||||
Radial quadrature: Mura-Knowles
|
||||
Angular quadrature: Lebedev.
|
||||
Tag B.-S. Rad. Rad. Pts. Rad. Cut. Ang. Pts.
|
||||
--- ---------- --------- --------- ---------
|
||||
O 0.60 70 10.0 590
|
||||
N 0.65 70 7.0 590
|
||||
C 0.70 70 8.0 590
|
||||
H 0.35 60 8.0 590
|
||||
Grid pruning is: on
|
||||
Number of quadrature shells: 330
|
||||
Spatial weights used: Erf1
|
||||
|
||||
Convergence Information
|
||||
-----------------------
|
||||
Convergence aids based upon iterative change in
|
||||
total energy or number of iterations.
|
||||
Levelshifting, if invoked, occurs when the
|
||||
HOMO/LUMO gap drops below (HL_TOL): 1.00D-02
|
||||
DIIS, if invoked, will attempt to extrapolate
|
||||
using up to (NFOCK): 10 stored Fock matrices.
|
||||
|
||||
Damping( 0%) Levelshifting(0.5) DIIS
|
||||
--------------- ------------------- ---------------
|
||||
dE on: start ASAP start
|
||||
dE off: 2 iters 50 iters 50 iters
|
||||
|
||||
|
||||
Screening Tolerance Information
|
||||
-------------------------------
|
||||
Density screening/tol_rho: 1.00D-11
|
||||
AO Gaussian exp screening on grid/accAOfunc: 18
|
||||
CD Gaussian exp screening on grid/accCDfunc: 20
|
||||
XC Gaussian exp screening on grid/accXCfunc: 20
|
||||
Schwarz screening/accCoul: 1.00D-14
|
||||
|
||||
|
||||
Loading old vectors from job with title :
|
||||
|
||||
GW Trifluoroacetamide
|
||||
|
||||
Time after variat. SCF: 12.5
|
||||
|
||||
3 Center 2 Electron Integral Information
|
||||
----------------------------------------
|
||||
Maximum number of 3-center 2e- integrals is: 529984.
|
||||
This is reduced with Schwarz screening to: 367500.
|
||||
Incore requires a per proc buffer size of: 175029.
|
||||
Minimum dble words available (all nodes) is: 131065476
|
||||
This is reduced (for later use) to: 130683464
|
||||
proc 0 Suggested buffer size is: 175029
|
||||
Max Suggested buffer size is: 175029
|
||||
no. integral batches is: 1000
|
||||
|
||||
Incore memory use for 3-center 2e- integrals is turned off.
|
||||
Time prior to 1st pass: 12.5
|
||||
|
||||
|
||||
Total DFT energy = -169.066312547888
|
||||
One electron energy = -355.393692322893
|
||||
Coulomb energy = 144.424774055736
|
||||
Exchange-Corr. energy = -21.405772779389
|
||||
Nuclear repulsion energy = 63.308378498658
|
||||
|
||||
Numeric. integr. density = 22.999999954126
|
||||
|
||||
Total iterative time = 0.9s
|
||||
|
||||
|
||||
|
||||
alpha - beta orbital overlaps
|
||||
-----------------------------
|
||||
|
||||
|
||||
alpha 1 2 3 4 5 6 7 8 9 10
|
||||
beta 1 2 3 4 5 6 7 8 9 10
|
||||
overlap 1.000 1.000 1.000 1.000 1.000 0.996 0.996 0.990 1.000 0.985
|
||||
|
||||
|
||||
alpha 11 12 13 14 15 16 17 18 19 20
|
||||
beta 11 12 13 14 15 16 17 18 19 20
|
||||
overlap 1.000 0.993 0.999 0.994 0.995 0.999 0.999 0.952 1.000 0.945
|
||||
|
||||
|
||||
alpha 21 22 23 24 25 26 27 28 29 30
|
||||
beta 21 22 23 24 25 26 27 28 29 30
|
||||
overlap 0.958 0.999 0.998 1.000 0.998 0.999 0.998 0.996 0.997 0.980
|
||||
|
||||
|
||||
alpha 31 32 33 34 35 36 37 38 39 40
|
||||
beta 31 32 33 34 35 36 37 38 39 40
|
||||
overlap 0.985 0.996 0.991 0.992 0.999 1.000 0.997 0.994 0.997 0.999
|
||||
|
||||
|
||||
alpha 41 42 43 44 45 46 47 48 49 50
|
||||
beta 41 42 43 44 45 46 47 48 49 50
|
||||
overlap 0.999 0.996 0.994 0.999 0.998 0.998 1.000 0.999 0.999 1.000
|
||||
|
||||
|
||||
alpha 51 52
|
||||
beta 51 52
|
||||
overlap 0.999 0.999
|
||||
|
||||
--------------------------
|
||||
Expectation value of S2:
|
||||
--------------------------
|
||||
<S2> = 0.7519 (Exact = 0.7500)
|
||||
|
||||
|
||||
center of mass
|
||||
--------------
|
||||
x = -0.02948626 y = 0.13290476 z = -0.00001949
|
||||
|
||||
moments of inertia (a.u.)
|
||||
------------------
|
||||
162.621484064576 0.000000000000 -0.000000000000
|
||||
0.000000000000 16.578708728791 -0.000000000000
|
||||
-0.000000000000 -0.000000000000 179.200070028499
|
||||
|
||||
Multipole analysis of the density
|
||||
---------------------------------
|
||||
|
||||
L x y z total alpha beta nuclear
|
||||
- - - - ----- ----- ---- -------
|
||||
0 0 0 0 -0.000000 -12.000000 -11.000000 23.000000
|
||||
|
||||
1 1 0 0 0.238427 0.541429 -0.303002 0.000000
|
||||
1 0 1 0 -1.321737 -0.969708 -0.352030 0.000000
|
||||
1 0 0 1 0.000663 -0.000054 0.000717 0.000000
|
||||
|
||||
2 2 0 0 -11.884337 -13.068472 -9.699498 10.883633
|
||||
2 1 1 0 -0.250985 0.506239 0.213062 -0.970286
|
||||
2 1 0 1 0.006019 -0.000243 -0.000755 0.007017
|
||||
2 0 2 0 -11.275818 -52.448505 -49.703057 90.875744
|
||||
2 0 1 1 0.001863 0.000480 -0.000509 0.001891
|
||||
2 0 0 2 -13.100006 -6.900236 -6.199822 0.000052
|
||||
|
||||
|
||||
***********************
|
||||
*** GW@DFT ***
|
||||
***********************
|
||||
|
||||
|
||||
Memory Information
|
||||
------------------
|
||||
Available GA per process is 1000.0 MBs
|
||||
Available stack per process is 500.0 MBs
|
||||
Available heap per process is 500.0 MBs
|
||||
|
||||
Peak memory in Initialization
|
||||
Peak GA per process is 1.4 MBs
|
||||
Peak stack per process is 0.0 MBs
|
||||
Peak heap per process is 1.6 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Peak memory in AFF Calculation
|
||||
Peak GA per process is 5.2 MBs
|
||||
Peak stack per process is 6.6 MBs
|
||||
Peak heap per process is 0.0 MBs
|
||||
Memory not in MA per process 6.6 MBs
|
||||
|
||||
Computing 2-center integrals
|
||||
Computing Inverse Cholesky factor
|
||||
Computing 3-center integrals
|
||||
Computing V_xc ...
|
||||
Computing Sigma_x
|
||||
Computing RPA polarizability
|
||||
Solving quasiparticle equations
|
||||
|
||||
Alpha Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 8 -13.237 0.001
|
||||
@GW 9 -12.542 0.000
|
||||
@GW 10 -12.360 0.000
|
||||
@GW 11 -9.991 0.000
|
||||
@GW 12 -7.780 0.001
|
||||
@GW 13 3.732 0.000
|
||||
@GW 14 3.851 0.000
|
||||
@GW 15 5.314 0.000
|
||||
@GW 16 9.843 0.001
|
||||
@GW 17 11.907 0.000
|
||||
|
||||
Beta Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 7 -16.569 0.000
|
||||
@GW 8 -14.089 0.000
|
||||
@GW 9 -12.853 0.000
|
||||
@GW 10 -11.113 0.000
|
||||
@GW 11 -9.813 0.000
|
||||
@GW 12 0.782 0.000
|
||||
@GW 13 4.374 0.000
|
||||
@GW 14 4.112 0.000
|
||||
@GW 15 5.494 0.000
|
||||
@GW 16 10.421 0.000
|
||||
|
||||
GW Timing Statistics
|
||||
------------------------
|
||||
Init. : 0.1s
|
||||
ERIs : 0.2s
|
||||
Distr.: 0.0s
|
||||
Vxc : 0.7s
|
||||
Sigma_x : 0.0s
|
||||
RPA : 0.3s
|
||||
Wmn : 0.0s
|
||||
Sigma_c : 0.0s
|
||||
GW total: 1.3s
|
||||
|
||||
|
||||
Task times cpu: 2.3s wall: 2.3s
|
||||
Task times cpu: 13.8s wall: 13.8s
|
||||
|
||||
|
||||
NWChem Input Module
|
||||
|
|
@ -1160,11 +864,11 @@ MA usage statistics:
|
|||
heap stack
|
||||
---- -----
|
||||
current number of blocks 0 0
|
||||
maximum number of blocks 28 58
|
||||
maximum number of blocks 28 57
|
||||
current total bytes 0 0
|
||||
maximum total bytes 21812064 30217960
|
||||
maximum total K-bytes 21813 30218
|
||||
maximum total M-bytes 22 31
|
||||
maximum total bytes 23306336 30269032
|
||||
maximum total K-bytes 23307 30270
|
||||
maximum total M-bytes 24 31
|
||||
|
||||
|
||||
CITATION
|
||||
|
|
@ -1200,24 +904,25 @@ MA usage statistics:
|
|||
|
||||
AUTHORS
|
||||
-------
|
||||
E. Apra, E. J. Bylaska, N. Govind, K. Kowalski, M. Valiev, W. A. de Jong,
|
||||
T. P. Straatsma, H. J. J. van Dam, D. Wang, T. L. Windus, N. P. Bauman,
|
||||
A. Panyala, J. Hammond, J. Autschbach, K. Bhaskaran-Nair, J. Brabec,
|
||||
K. Lopata, S. A. Fischer, S. Krishnamoorthy, M. Jacquelin, W. Ma, M. Klemm,
|
||||
O. Villa, Y. Chen, V. Anisimov, F. Aquino, S. Hirata, M. T. Hackler,
|
||||
Eric Hermes, L. Jensen, J. E. Moore, J. C. Becca, V. Konjkov,
|
||||
D. Mejia-Rodriguez, T. Risthaus, M. Malagoli, A. Marenich,
|
||||
A. Otero-de-la-Roza, J. Mullin, P. Nichols, R. Peverati, J. Pittner, Y. Zhao,
|
||||
P.-D. Fan, A. Fonari, M. J. Williamson, R. J. Harrison, J. R. Rehr,
|
||||
M. Dupuis, D. Silverstein, D. M. A. Smith, J. Nieplocha, V. Tipparaju,
|
||||
M. Krishnan, B. E. Van Kuiken, A. Vazquez-Mayagoitia, M. Swart, Q. Wu,
|
||||
T. Van Voorhis, A. A. Auer, M. Nooijen, L. D. Crosby, E. Brown, G. Cisneros,
|
||||
G. I. Fann, H. Fruchtl, J. Garza, K. Hirao, R. A. Kendall, J. A. Nichols,
|
||||
K. Tsemekhman, K. Wolinski, J. Anchell, D. E. Bernholdt, P. Borowski,
|
||||
T. Clark, D. Clerc, H. Dachsel, M. J. O. Deegan, K. Dyall, D. Elwood,
|
||||
E. Glendening, M. Gutowski, A. C. Hess, J. Jaffe, B. G. Johnson, J. Ju,
|
||||
R. Kobayashi, R. Kutteh, Z. Lin, R. Littlefield, X. Long, B. Meng,
|
||||
T. Nakajima, S. Niu, L. Pollack, M. Rosing, K. Glaesemann, G. Sandrone,
|
||||
M. Stave, H. Taylor, G. Thomas, J. H. van Lenthe, A. T. Wong, Z. Zhang.
|
||||
E. Apra, E. J. Bylaska, N. Govind, K. Kowalski, D. Mejia-Rodriguez,
|
||||
N. P. Bauman, A. Panyala, R. J. Harrison, M. Valiev, W. A. de Jong,
|
||||
T. P. Straatsma, H. J. J. van Dam, D. Wang, T. L. Windus, J. Hammond,
|
||||
J. Autschbach, R. de P. Soares, A. Kunitsa, K. Bhaskaran-Nair, J. Brabec,
|
||||
K. Lopata, S. A. Fischer, S. Krishnamoorthy, M. Jacquelin, A. Woods, W. Ma,
|
||||
M. Klemm, O. Villa, Y. Chen, V. Anisimov, F. Aquino, S. Hirata,
|
||||
M. T. Hackler, E. Hermes, L. Jensen, J. E. Moore, J. C. Becca, V. Konjkov,
|
||||
T. Risthaus, M. Malagoli, A. Marenich, A. Otero-de-la-Roza, J. Mullin,
|
||||
P. Nichols, R. Peverati, J. Pittner, Y. Zhao, P.-D. Fan, A. Fonari,
|
||||
M. J. Williamson, J. R. Rehr, M. Dupuis, D. Silverstein, D. M. A. Smith,
|
||||
J. Nieplocha, V. Tipparaju, M. Krishnan, B. E. Van Kuiken,
|
||||
A. Vazquez-Mayagoitia, M. Swart, Q. Wu, T. Van Voorhis, A. A. Auer,
|
||||
M. Nooijen, L. D. Crosby, E. Brown, G. Cisneros, G. I. Fann, H. Fruchtl,
|
||||
J. Garza, K. Hirao, R. A. Kendall, J. A. Nichols, K. Tsemekhman, K. Wolinski,
|
||||
J. Anchell, D. E. Bernholdt, P. Borowski, T. Clark, D. Clerc, H. Dachsel,
|
||||
M. J. O. Deegan, K. Dyall, D. Elwood, E. Glendening, M. Gutowski, A. C. Hess,
|
||||
J. Jaffe, B. G. Johnson, J. Ju, R. Kobayashi, R. Kutteh, Z. Lin,
|
||||
R. Littlefield, X. Long, B. Meng, T. Nakajima, S. Niu, L. Pollack, M. Rosing,
|
||||
K. Glaesemann, G. Sandrone, M. Stave, H. Taylor, G. Thomas, J. H. van Lenthe,
|
||||
A. T. Wong, Z. Zhang.
|
||||
|
||||
Total times cpu: 14.5s wall: 14.7s
|
||||
Total times cpu: 14.2s wall: 14.2s
|
||||
|
|
|
|||
|
|
@ -1,5 +1,5 @@
|
|||
argument 1 = gw_symmetry.nw
|
||||
NWChem w/ OpenMP: maximum threads = 8
|
||||
argument 1 = /people/meji656/calculations/bse/QA/gw_symmetry.nw
|
||||
NWChem w/ OpenMP: maximum threads = 1
|
||||
|
||||
|
||||
|
||||
|
|
@ -53,7 +53,7 @@ task dft gw
|
|||
|
||||
|
||||
|
||||
Northwest Computational Chemistry Package (NWChem) 7.2.0
|
||||
Northwest Computational Chemistry Package (NWChem) 7.3.0
|
||||
--------------------------------------------------------
|
||||
|
||||
|
||||
|
|
@ -61,7 +61,7 @@ task dft gw
|
|||
Pacific Northwest National Laboratory
|
||||
Richland, WA 99352
|
||||
|
||||
Copyright (c) 1994-2022
|
||||
Copyright (c) 1994-2025
|
||||
Pacific Northwest National Laboratory
|
||||
Battelle Memorial Institute
|
||||
|
||||
|
|
@ -86,21 +86,21 @@ task dft gw
|
|||
Job information
|
||||
---------------
|
||||
|
||||
hostname = WE41476
|
||||
program = nwchem
|
||||
date = Wed Jul 12 13:50:26 2023
|
||||
hostname = dc005
|
||||
program = /qfs/people/meji656/software/nwchem.bse/bin/LINUX64/nwchem
|
||||
date = Thu Oct 16 10:35:32 2025
|
||||
|
||||
compiled = Wed_Jul_12_13:50:03_2023
|
||||
source = /Users/meji656/Sources/nwchem_bse
|
||||
nwchem branch = 7.2.0
|
||||
nwchem revision = nwchem_on_git-4850-g011c901253
|
||||
ga revision = 5.8.0
|
||||
use scalapack = T
|
||||
input = gw_symmetry.nw
|
||||
compiled = Thu_Oct_16_10:07:15_2025
|
||||
source = /qfs/people/meji656/software/nwchem.bse
|
||||
nwchem branch = 7.3.0
|
||||
nwchem revision = v7.2.0-beta1-1624-g0dc011d0a0
|
||||
ga revision = 5.9.2
|
||||
use scalapack = F
|
||||
input = /people/meji656/calculations/bse/QA/gw_symmetry.nw
|
||||
prefix = gw_symmetry.
|
||||
data base = ./gw_symmetry.db
|
||||
status = startup
|
||||
nproc = 1
|
||||
nproc = 7
|
||||
time left = -1s
|
||||
|
||||
|
||||
|
|
@ -108,10 +108,10 @@ task dft gw
|
|||
Memory information
|
||||
------------------
|
||||
|
||||
heap = 65535996 doubles = 500.0 Mbytes
|
||||
stack = 65536001 doubles = 500.0 Mbytes
|
||||
global = 131072000 doubles = 1000.0 Mbytes (distinct from heap & stack)
|
||||
total = 262143997 doubles = 2000.0 Mbytes
|
||||
heap = 65535998 doubles = 500.0 Mbytes
|
||||
stack = 65535995 doubles = 500.0 Mbytes
|
||||
global = 131072000 doubles = 1000.0 Mbytes (distinct from heap & stack)
|
||||
total = 262143993 doubles = 2000.0 Mbytes
|
||||
verify = yes
|
||||
hardfail = no
|
||||
|
||||
|
|
@ -462,33 +462,32 @@ task dft gw
|
|||
6 a1 7 t2 8 t2 9 t2 10 t2
|
||||
11 t2 12 t2 13 a1 14 t2 15 t2
|
||||
|
||||
WARNING: movecs_in_org=atomic not equal to movecs_in=./gw_symmetry.movecs
|
||||
Time after variat. SCF: 0.3
|
||||
Time prior to 1st pass: 0.3
|
||||
Time after variat. SCF: 0.4
|
||||
Time prior to 1st pass: 0.4
|
||||
|
||||
Memory utilization after 1st SCF pass:
|
||||
Heap Space remaining (MW): 65.53 65532636
|
||||
Stack Space remaining (MW): 65.54 65535396
|
||||
Heap Space remaining (MW): 65.53 65532638
|
||||
Stack Space remaining (MW): 65.54 65535388
|
||||
|
||||
convergence iter energy DeltaE RMS-Dens Diis-err time
|
||||
---------------- ----- ----------------- --------- --------- --------- ------
|
||||
d= 0,ls=0.0,diis 1 -40.4526766257 -5.40D+01 2.73D-03 1.58D-01 0.6
|
||||
d= 0,ls=0.0,diis 2 -40.4605212533 -7.84D-03 1.18D-03 2.13D-02 0.8
|
||||
d= 0,ls=0.0,diis 3 -40.4604458885 7.54D-05 6.05D-04 2.16D-02 1.1
|
||||
d= 0,ls=0.0,diis 4 -40.4630635417 -2.62D-03 2.24D-05 6.66D-06 1.3
|
||||
d= 0,ls=0.0,diis 5 -40.4630646869 -1.15D-06 2.13D-06 1.39D-07 1.6
|
||||
d= 0,ls=0.0,diis 6 -40.4630647071 -2.03D-08 7.01D-08 1.45D-10 1.9
|
||||
d= 0,ls=0.0,diis 1 -40.4526766257 -5.40D+01 2.73D-03 1.58D-01 0.5
|
||||
d= 0,ls=0.0,diis 2 -40.4605212533 -7.84D-03 1.18D-03 2.13D-02 0.5
|
||||
d= 0,ls=0.0,diis 3 -40.4604458885 7.54D-05 6.05D-04 2.16D-02 0.6
|
||||
d= 0,ls=0.0,diis 4 -40.4630635417 -2.62D-03 2.24D-05 6.66D-06 0.6
|
||||
d= 0,ls=0.0,diis 5 -40.4630646869 -1.15D-06 2.13D-06 1.39D-07 0.7
|
||||
d= 0,ls=0.0,diis 6 -40.4630647071 -2.03D-08 7.01D-08 1.45D-10 0.7
|
||||
|
||||
|
||||
Total DFT energy = -40.463064707119
|
||||
One electron energy = -79.987919842739
|
||||
Coulomb energy = 32.854590174254
|
||||
Total DFT energy = -40.463064707118
|
||||
One electron energy = -79.987919842738
|
||||
Coulomb energy = 32.854590174255
|
||||
Exchange-Corr. energy = -6.837400210970
|
||||
Nuclear repulsion energy = 13.507665172336
|
||||
|
||||
Numeric. integr. density = 10.000008758007
|
||||
|
||||
Total iterative time = 1.5s
|
||||
Total iterative time = 0.3s
|
||||
|
||||
|
||||
|
||||
|
|
@ -523,11 +522,11 @@ task dft gw
|
|||
|
||||
1 1 0 0 -0.000000 -0.000000 -0.000000 0.000000
|
||||
1 0 1 0 0.000000 0.000000 0.000000 0.000000
|
||||
1 0 0 1 -0.000000 -0.000000 -0.000000 0.000000
|
||||
1 0 0 1 0.000000 0.000000 0.000000 0.000000
|
||||
|
||||
2 2 0 0 -6.283208 -5.939937 -5.939937 5.596666
|
||||
2 1 1 0 0.000000 0.000000 0.000000 0.000000
|
||||
2 1 0 1 0.000000 0.000000 0.000000 0.000000
|
||||
2 1 1 0 -0.000000 -0.000000 -0.000000 0.000000
|
||||
2 1 0 1 -0.000000 -0.000000 -0.000000 0.000000
|
||||
2 0 2 0 -6.283208 -5.939937 -5.939937 5.596666
|
||||
2 0 1 1 -0.000000 -0.000000 -0.000000 0.000000
|
||||
2 0 0 2 -6.283208 -5.939937 -5.939937 5.596666
|
||||
|
|
@ -626,18 +625,18 @@ task dft gw
|
|||
|
||||
Memory Information
|
||||
------------------
|
||||
Available GA per process is 999.9 MBs
|
||||
Available GA per process is 1000.0 MBs
|
||||
Available stack per process is 500.0 MBs
|
||||
Available heap per process is 499.9 MBs
|
||||
|
||||
Peak memory in Initialization
|
||||
Peak GA per process is 1.1 MBs
|
||||
Peak GA per process is 0.2 MBs
|
||||
Peak stack per process is 0.1 MBs
|
||||
Peak heap per process is 1.2 MBs
|
||||
Peak heap per process is 0.5 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Peak memory in AFF Calculation
|
||||
Peak GA per process is 2.5 MBs
|
||||
Peak GA per process is 0.4 MBs
|
||||
Peak stack per process is 1.3 MBs
|
||||
Peak heap per process is 0.0 MBs
|
||||
Memory not in MA per process 1.3 MBs
|
||||
|
|
@ -657,25 +656,25 @@ task dft gw
|
|||
Alpha Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 5 -13.866 0.000
|
||||
@GW 5 -13.864 0.003
|
||||
|
||||
Applying -4.4168 eV scissor-shift to rest of particle states
|
||||
Applying -4.4144 eV scissor-shift to rest of particle states
|
||||
|
||||
|
||||
GW Timing Statistics
|
||||
------------------------
|
||||
Init. : 0.1s
|
||||
ERIs : 0.3s
|
||||
ERIs : 0.1s
|
||||
Distr.: 0.0s
|
||||
Vxc : 0.2s
|
||||
Vxc : 0.0s
|
||||
Sigma_x : 0.0s
|
||||
RPA : 0.0s
|
||||
Wmn : 0.0s
|
||||
Sigma_c : 0.0s
|
||||
GW total: 0.5s
|
||||
GW total: 0.2s
|
||||
|
||||
|
||||
Task times cpu: 2.4s wall: 2.4s
|
||||
Task times cpu: 0.6s wall: 0.6s
|
||||
|
||||
|
||||
NWChem Input Module
|
||||
|
|
@ -784,7 +783,7 @@ task dft gw
|
|||
|
||||
Loading old vectors from job with title :
|
||||
|
||||
GW Methane Td
|
||||
W vals
|
||||
|
||||
|
||||
Symmetry analysis of molecular orbitals - initial
|
||||
|
|
@ -800,28 +799,28 @@ GW Methane Td
|
|||
6 a1 7 t2 8 t2 9 t2 10 t2
|
||||
11 t2 12 t2 13 t2 14 t2 15 t2
|
||||
|
||||
Time after variat. SCF: 2.5
|
||||
Time prior to 1st pass: 2.5
|
||||
Time after variat. SCF: 0.9
|
||||
Time prior to 1st pass: 0.9
|
||||
|
||||
Memory utilization after 1st SCF pass:
|
||||
Heap Space remaining (MW): 65.53 65532636
|
||||
Stack Space remaining (MW): 65.54 65535396
|
||||
Heap Space remaining (MW): 65.53 65532638
|
||||
Stack Space remaining (MW): 65.54 65535388
|
||||
|
||||
convergence iter energy DeltaE RMS-Dens Diis-err time
|
||||
---------------- ----- ----------------- --------- --------- --------- ------
|
||||
d= 0,ls=0.0,diis 1 -40.4630647077 -5.40D+01 1.34D-08 1.68D-12 3.5
|
||||
d= 0,ls=0.0,diis 2 -40.4630647077 -2.15D-12 7.98D-09 3.73D-12 4.4
|
||||
d= 0,ls=0.0,diis 1 -40.4630647076 -5.40D+01 1.34D-08 1.68D-12 1.0
|
||||
d= 0,ls=0.0,diis 2 -40.4630647076 2.13D-13 7.98D-09 3.73D-12 1.1
|
||||
|
||||
|
||||
Total DFT energy = -40.463064707739
|
||||
One electron energy = -79.987938823681
|
||||
Coulomb energy = 32.854611046058
|
||||
Exchange-Corr. energy = -6.837402102452
|
||||
Total DFT energy = -40.463064707584
|
||||
One electron energy = -79.987938823837
|
||||
Coulomb energy = 32.854611046046
|
||||
Exchange-Corr. energy = -6.837402102129
|
||||
Nuclear repulsion energy = 13.507665172336
|
||||
|
||||
Numeric. integr. density = 10.000008760592
|
||||
Numeric. integr. density = 10.000008760422
|
||||
|
||||
Total iterative time = 1.9s
|
||||
Total iterative time = 0.3s
|
||||
|
||||
|
||||
|
||||
|
|
@ -854,15 +853,15 @@ GW Methane Td
|
|||
- - - - ----- ----- ---- -------
|
||||
0 0 0 0 -0.000000 -5.000000 -5.000000 10.000000
|
||||
|
||||
1 1 0 0 -0.000000 -0.000000 -0.000000 0.000000
|
||||
1 0 1 0 -0.000000 -0.000000 -0.000000 0.000000
|
||||
1 0 0 1 -0.000000 -0.000000 -0.000000 0.000000
|
||||
1 1 0 0 0.000000 0.000000 0.000000 0.000000
|
||||
1 0 1 0 0.000000 0.000000 0.000000 0.000000
|
||||
1 0 0 1 0.000000 0.000000 0.000000 0.000000
|
||||
|
||||
2 2 0 0 -6.283210 -5.939938 -5.939938 5.596666
|
||||
2 1 1 0 -0.000000 -0.000000 -0.000000 0.000000
|
||||
2 1 0 1 -0.000000 -0.000000 -0.000000 0.000000
|
||||
2 1 1 0 0.000000 0.000000 0.000000 0.000000
|
||||
2 1 0 1 0.000000 0.000000 0.000000 0.000000
|
||||
2 0 2 0 -6.283210 -5.939938 -5.939938 5.596666
|
||||
2 0 1 1 -0.000000 -0.000000 -0.000000 0.000000
|
||||
2 0 1 1 0.000000 0.000000 0.000000 0.000000
|
||||
2 0 0 2 -6.283210 -5.939938 -5.939938 5.596666
|
||||
|
||||
|
||||
|
|
@ -873,18 +872,18 @@ GW Methane Td
|
|||
|
||||
Memory Information
|
||||
------------------
|
||||
Available GA per process is 999.9 MBs
|
||||
Available GA per process is 1000.0 MBs
|
||||
Available stack per process is 500.0 MBs
|
||||
Available heap per process is 499.9 MBs
|
||||
|
||||
Peak memory in Initialization
|
||||
Peak GA per process is 1.1 MBs
|
||||
Peak GA per process is 0.2 MBs
|
||||
Peak stack per process is 0.1 MBs
|
||||
Peak heap per process is 1.2 MBs
|
||||
Peak heap per process is 0.5 MBs
|
||||
Memory not in MA per process 0.0 MBs
|
||||
|
||||
Peak memory in AFF Calculation
|
||||
Peak GA per process is 2.5 MBs
|
||||
Peak GA per process is 0.4 MBs
|
||||
Peak stack per process is 1.3 MBs
|
||||
Peak heap per process is 0.0 MBs
|
||||
Memory not in MA per process 1.3 MBs
|
||||
|
|
@ -904,25 +903,25 @@ GW Methane Td
|
|||
Alpha Orbitals
|
||||
State Energy (eV) Error (eV)
|
||||
-----------------------------------
|
||||
@GW 5 -13.866 0.000
|
||||
@GW 5 -13.864 0.003
|
||||
|
||||
Applying -4.4168 eV scissor-shift to rest of particle states
|
||||
Applying -4.4144 eV scissor-shift to rest of particle states
|
||||
|
||||
|
||||
GW Timing Statistics
|
||||
------------------------
|
||||
Init. : 0.0s
|
||||
ERIs : 0.2s
|
||||
ERIs : 0.0s
|
||||
Distr.: 0.0s
|
||||
Vxc : 0.5s
|
||||
Vxc : 0.1s
|
||||
Sigma_x : 0.0s
|
||||
RPA : 0.0s
|
||||
Wmn : 0.0s
|
||||
Sigma_c : 0.0s
|
||||
GW total: 0.8s
|
||||
GW total: 0.2s
|
||||
|
||||
|
||||
Task times cpu: 2.8s wall: 2.8s
|
||||
Task times cpu: 0.5s wall: 0.5s
|
||||
|
||||
|
||||
NWChem Input Module
|
||||
|
|
@ -944,9 +943,9 @@ MA usage statistics:
|
|||
current number of blocks 0 0
|
||||
maximum number of blocks 29 62
|
||||
current total bytes 0 0
|
||||
maximum total bytes 32636240 32078504
|
||||
maximum total K-bytes 32637 32079
|
||||
maximum total M-bytes 33 33
|
||||
maximum total bytes 31929024 32078520
|
||||
maximum total K-bytes 31930 32079
|
||||
maximum total M-bytes 32 33
|
||||
|
||||
|
||||
CITATION
|
||||
|
|
@ -982,24 +981,25 @@ MA usage statistics:
|
|||
|
||||
AUTHORS
|
||||
-------
|
||||
E. Apra, E. J. Bylaska, N. Govind, K. Kowalski, M. Valiev, D. Mejia-Rodriguez,
|
||||
A. Kunitsa, N. P. Bauman, A. Panyala, W. A. de Jong, T. P. Straatsma,
|
||||
H. J. J. van Dam, D. Wang, T. L. Windus, J. Hammond, J. Autschbach, A. Woods,
|
||||
K. Bhaskaran-Nair, J. Brabec, K. Lopata, S. A. Fischer, S. Krishnamoorthy,
|
||||
M. Jacquelin, W. Ma, M. Klemm, O. Villa, Y. Chen, V. Anisimov, F. Aquino,
|
||||
E. Apra, E. J. Bylaska, N. Govind, K. Kowalski, D. Mejia-Rodriguez,
|
||||
N. P. Bauman, A. Panyala, R. J. Harrison, M. Valiev, W. A. de Jong,
|
||||
T. P. Straatsma, H. J. J. van Dam, D. Wang, T. L. Windus, J. Hammond,
|
||||
J. Autschbach, R. de P. Soares, A. Kunitsa, K. Bhaskaran-Nair, J. Brabec,
|
||||
K. Lopata, S. A. Fischer, S. Krishnamoorthy, M. Jacquelin, R. Goswami,
|
||||
A. Woods, W. Ma, M. Klemm, O. Villa, Y. Chen, V. Anisimov, F. Aquino,
|
||||
S. Hirata, M. T. Hackler, E. Hermes, L. Jensen, J. E. Moore, J. C. Becca,
|
||||
V. Konjkov, T. Risthaus, M. Malagoli, A. Marenich, A. Otero-de-la-Roza,
|
||||
J. Mullin, P. Nichols, R. Peverati, J. Pittner, Y. Zhao, P.-D. Fan,
|
||||
A. Fonari, M. J. Williamson, R. J. Harrison, J. R. Rehr, M. Dupuis,
|
||||
D. Silverstein, D. M. A. Smith, J. Nieplocha, V. Tipparaju, M. Krishnan,
|
||||
B. E. Van Kuiken, A. Vazquez-Mayagoitia, M. Swart, Q. Wu, T. Van Voorhis,
|
||||
A. A. Auer, M. Nooijen, L. D. Crosby, E. Brown, G. Cisneros, G. I. Fann,
|
||||
H. Fruchtl, J. Garza, K. Hirao, R. A. Kendall, J. A. Nichols, K. Tsemekhman,
|
||||
K. Wolinski, J. Anchell, D. E. Bernholdt, P. Borowski, T. Clark, D. Clerc,
|
||||
H. Dachsel, M. J. O. Deegan, K. Dyall, D. Elwood, E. Glendening, M. Gutowski,
|
||||
A. C. Hess, J. Jaffe, B. G. Johnson, J. Ju, R. Kobayashi, R. Kutteh, Z. Lin,
|
||||
A. Fonari, M. J. Williamson, J. R. Rehr, M. Dupuis, D. Silverstein,
|
||||
D. M. A. Smith, J. Nieplocha, V. Tipparaju, M. Krishnan, B. E. Van Kuiken,
|
||||
A. Vazquez-Mayagoitia, M. Swart, Q. Wu, T. Van Voorhis, A. A. Auer,
|
||||
M. Nooijen, L. D. Crosby, E. Brown, G. Cisneros, G. I. Fann, H. Fruchtl,
|
||||
J. Garza, K. Hirao, R. A. Kendall, J. A. Nichols, K. Tsemekhman, K. Wolinski,
|
||||
J. Anchell, D. E. Bernholdt, P. Borowski, T. Clark, D. Clerc, H. Dachsel,
|
||||
M. J. O. Deegan, K. Dyall, D. Elwood, E. Glendening, M. Gutowski, A. C. Hess,
|
||||
J. Jaffe, B. G. Johnson, J. Ju, R. Kobayashi, R. Kutteh, Z. Lin,
|
||||
R. Littlefield, X. Long, B. Meng, T. Nakajima, S. Niu, L. Pollack, M. Rosing,
|
||||
K. Glaesemann, G. Sandrone, M. Stave, H. Taylor, G. Thomas, J. H. van Lenthe,
|
||||
A. T. Wong, Z. Zhang.
|
||||
|
||||
Total times cpu: 5.2s wall: 5.2s
|
||||
Total times cpu: 1.3s wall: 1.3s
|
||||
|
|
|
|||
44
src/bsemol/GNUmakefile
Normal file
44
src/bsemol/GNUmakefile
Normal file
|
|
@ -0,0 +1,44 @@
|
|||
OBJ = bsemol.o bse_input.o bse_init.o bse_analytic.o \
|
||||
bse_davidson.o bse_lanczos.o bse_defaults.o \
|
||||
bse_finalize.o bse_ri_init.o bse_wmn.o \
|
||||
bse_analytic_cis.o bse_analytic_tda.o \
|
||||
bse_oscstr.o bse_puteris.o bse_get_eia.o \
|
||||
bse_davidson_diagonal.o \
|
||||
bse_davidson_kvec.o bse_davidson_guess.o \
|
||||
bse_davidson_kortho.o bse_davidson_restart.o \
|
||||
bse_davidson_newvecs.o bse_davidson_residual.o \
|
||||
bse_s2.o
|
||||
|
||||
USES_BLAS = bse_buildw.F bse_ri_init.F bse_wmn.F \
|
||||
bse_davidson.F bse_davidson_diagonal.F bse_davidson_kvec.F bse_davidson_kortho.F \
|
||||
bse_init.F bse_puteris.F bse_s2.F
|
||||
|
||||
OBJ_OPTIMIZE = bse_buildw.o
|
||||
|
||||
LIBRARY = libbsemol.a
|
||||
|
||||
HEADERS = ../nwdft/include/cdft.fh ../nwdft/grid/cgridfile.fh ../nwdft/include/dft_scaleig.fh
|
||||
|
||||
LIB_DEFINES =
|
||||
LIB_INCLUDES = -I./
|
||||
|
||||
include ../config/makefile.h
|
||||
|
||||
|
||||
ifeq ($(_USE_SCALAPACK),Y)
|
||||
_USE_ELPA = $(shell ${GA_PATH}/bin/ga-config --use_elpa| awk ' /1/ {print "Y"}')
|
||||
|
||||
ifeq ($(_USE_ELPA),Y)
|
||||
LIB_DEFINES += -DUSE_ELPA
|
||||
endif
|
||||
#assuming we have GAPZHEEVD
|
||||
GOTGAPZHEEVD= Y
|
||||
ifeq ($(GOTGAPZHEEVD),Y)
|
||||
LIB_DEFINES += -DUSE_GAPZHEEVD
|
||||
endif
|
||||
_PDSYEVR = $(shell [ -e ${NWCHEM_TOP}/src/tools/install/lib/libga.a ] && nm ${NWCHEM_TOP}/src/tools/install/lib/libga.a|egrep T |egrep ga_pdsyevr |awk '/ga_pdsyevr/ {print "Y"}')
|
||||
ifeq ($(_PDSYEVR),Y)
|
||||
LIB_DEFINES += -DUSE_PDSYEVR
|
||||
endif
|
||||
endif
|
||||
include ../config/makelib.h
|
||||
44
src/bsemol/bse.fh
Normal file
44
src/bsemol/bse.fh
Normal file
|
|
@ -0,0 +1,44 @@
|
|||
|
||||
double precision, parameter :: ha2ev = 27.211386245988d0
|
||||
|
||||
type :: bse_params_t
|
||||
sequence
|
||||
integer :: rtdb
|
||||
integer :: me, nprocs, iMaxThreads
|
||||
integer :: ipol
|
||||
logical :: main, odft
|
||||
c
|
||||
integer :: maxiter, nspace
|
||||
integer :: natoms
|
||||
integer :: nmo, nri
|
||||
integer :: nocc(2),nvir(2),npoles(2),mynpoles(2)
|
||||
integer :: oolo(2),oohi(2)
|
||||
integer :: ovlo(2),ovhi(2)
|
||||
integer :: vvlo(2),vvhi(2)
|
||||
integer :: llmo(2),nqp(2)
|
||||
integer :: nroots,nmax
|
||||
double precision :: tol2e, eta
|
||||
c
|
||||
integer :: l_mf_evals, k_mf_evals
|
||||
integer :: l_gw_evals, k_gw_evals
|
||||
integer :: l_occ, k_occ
|
||||
integer :: l_irs(2),k_irs(2)
|
||||
integer :: l_map,k_map
|
||||
integer :: nblocks
|
||||
c
|
||||
integer :: fd
|
||||
c
|
||||
integer :: g_movecs(2), g_moeris, g_sols
|
||||
integer :: g_erioo(2), g_eriov(2), g_erivv(2), g_erim
|
||||
integer :: g_wov(2), g_apb, g_dipa(3), g_dipb(3), g_amb
|
||||
integer :: g_quada(6), g_quadb(6)
|
||||
integer :: g_dipmaga(3), g_dipmagb(3)
|
||||
logical :: noio
|
||||
c
|
||||
logical :: ncap, davidson, lanczos, analytic, tda
|
||||
logical :: singlet, triplet, dowindow
|
||||
double precision :: thresh, fdstep, vxddp(2), vxddm(2)
|
||||
double precision :: elower, eupper
|
||||
|
||||
end type bse_params_t
|
||||
|
||||
82
src/bsemol/bse_analytic.F
Normal file
82
src/bsemol/bse_analytic.F
Normal file
|
|
@ -0,0 +1,82 @@
|
|||
subroutine bse_analytic(pars)
|
||||
|
||||
implicit none
|
||||
|
||||
#include "global.fh"
|
||||
#include "errquit.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "bse.fh"
|
||||
#include "cdft.fh"
|
||||
type(bse_params_t) :: pars
|
||||
character(*), parameter :: pname = 'bse_analytic: '
|
||||
|
||||
integer lSize,isp
|
||||
integer l_omega, k_omega
|
||||
integer l_dia,k_dia
|
||||
integer lW,kW
|
||||
integer maxpoles,maxnpoles,totnpoles,nri
|
||||
|
||||
! synchronize parallel processes
|
||||
call ga_sync
|
||||
|
||||
!initialize useful variables
|
||||
nri = pars%nri
|
||||
maxpoles = maxval(pars%mynpoles(1:pars%ipol))
|
||||
maxnpoles = maxval(pars%npoles(1:pars%ipol))
|
||||
totnpoles = maxnpoles*pars%ipol
|
||||
|
||||
! allocation
|
||||
if(.not.ma_push_get(mt_dbl,totnpoles,'omega',l_omega,k_omega))
|
||||
& call errquit(pname//'failed to allocate omega',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,totnpoles,'dia',l_dia,k_dia))
|
||||
& call errquit(pname//'failed to allocate Delta_ia',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,nri**2,'W',lW,kW))
|
||||
& call errquit(pname//'failed to allocate W',0,MA_ERR)
|
||||
|
||||
! obtain "old" eigenvalue difference to build W
|
||||
do isp=1,pars%ipol
|
||||
lSize = (isp-1)*maxpoles
|
||||
call bse_get_eia(dbl_mb(pars%k_mf_evals+(isp-1)*nbf_ao),
|
||||
& dbl_mb(k_dia+lSize),pars%nocc(isp),
|
||||
& pars%nvir(isp),pars%ovlo(isp),pars%ovhi(isp))
|
||||
enddo
|
||||
|
||||
! obtain screened Coulomb matrix
|
||||
call bse_buildw(pars,dbl_mb(k_dia),dbl_mb(kW),pars%nmo,nri,
|
||||
& maxpoles,pars%ipol)
|
||||
|
||||
! transform ERIs using screened Coulomb matrix
|
||||
call bse_wmn(pars,dbl_mb(kW),pars%nmo,nri,pars%ipol)
|
||||
if(.not.ma_chop_stack(lW))
|
||||
& call errquit(pname//'failed to chop stack',0,MA_ERR)
|
||||
|
||||
! obtain "GW" eigenvalue difference to build Casida matrix
|
||||
do isp=1,pars%ipol
|
||||
lSize = (isp-1)*maxpoles
|
||||
call bse_get_eia(dbl_mb(pars%k_gw_evals+(isp-1)*nbf_ao),
|
||||
& dbl_mb(k_dia+lSize),pars%nocc(isp),
|
||||
& pars%nvir(isp),pars%ovlo(isp),
|
||||
& pars%ovhi(isp))
|
||||
enddo
|
||||
|
||||
! Build and diagonalize Casida-like matrix
|
||||
if (pars%tda) then
|
||||
call bse_analytic_tda(pars,dbl_mb(k_dia),dbl_mb(k_omega),
|
||||
& maxpoles,totnpoles,pars%singlet)
|
||||
else
|
||||
call bse_analytic_cis(pars,dbl_mb(k_dia),dbl_mb(k_omega),
|
||||
& maxpoles,totnpoles,pars%singlet)
|
||||
endif
|
||||
|
||||
! Compute oscillator strengths
|
||||
call bse_oscstr(pars,dbl_mb(k_omega),totnpoles)
|
||||
|
||||
! Destroy GAs
|
||||
if (.not.ga_destroy(pars%g_apb))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
! chop stack
|
||||
if(.not.ma_chop_stack(l_omega))
|
||||
$ call errquit(pname//'failed to chop stack',0,MA_ERR)
|
||||
|
||||
end subroutine bse_analytic
|
||||
359
src/bsemol/bse_analytic_cis.F
Normal file
359
src/bsemol/bse_analytic_cis.F
Normal file
|
|
@ -0,0 +1,359 @@
|
|||
subroutine bse_analytic_cis(pars,wia,omega,maxpoles,totpoles,
|
||||
$ singlet)
|
||||
implicit none
|
||||
#include "mafdecls.fh"
|
||||
#include "errquit.fh"
|
||||
#include "global.fh"
|
||||
#include "bse.fh"
|
||||
#ifdef SCALAPACK
|
||||
#include "scaleig.fh"
|
||||
integer ga_cholesky,ga_llt_i
|
||||
external ga_cholesky,ga_llt_i
|
||||
#endif
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
|
||||
logical singlet
|
||||
integer maxpoles, totpoles, npoles
|
||||
double precision wia(maxpoles,pars%ipol)
|
||||
double precision omega(totpoles)
|
||||
|
||||
character(*),parameter :: pname = 'bse_casida: '
|
||||
|
||||
integer isp,jsp,nri,off1,off2
|
||||
integer ilo,ihi,jlo,jhi
|
||||
integer nocc2(2),nvir2(2)
|
||||
integer ipole,imo,jmo,amo,bmo
|
||||
integer k,l,kproc,lproc,klocal,llocal,kglobal,lglobal
|
||||
integer g_diag, g_tmp, g_amb, ktmp, ld, info, i, j
|
||||
integer ij,jb,llia,ulia
|
||||
|
||||
#ifdef USE_OPENMP
|
||||
call util_blas_set_num_threads(pars%iMaxthreads)
|
||||
#endif
|
||||
|
||||
nri = pars%nri
|
||||
npoles = pars%npoles(1)
|
||||
if (pars%ipol.gt.1) npoles = npoles + pars%npoles(2)
|
||||
|
||||
if(.not.ga_create(mt_dbl,npoles,npoles,'apb',0,npoles,
|
||||
& pars%g_apb))
|
||||
& call errquit(pname//'could not create A+B GA',0,GA_ERR)
|
||||
if(.not.ga_create(mt_dbl,npoles,npoles,'amb',0,npoles,
|
||||
& g_amb))
|
||||
& call errquit(pname//'could not create A-B GA',0,GA_ERR)
|
||||
if (.not.nga_create(mt_dbl,1,npoles,'wia',0,g_diag))
|
||||
$ call errquit(pname//'could not create wia',0,GA_ERR)
|
||||
|
||||
if (pars%ipol.eq.1) then
|
||||
nocc2(1) = pars%nocc(1)**2
|
||||
nvir2(1) = pars%nvir(1)**2
|
||||
|
||||
call ga_distribution(pars%g_apb,pars%me,ilo,ihi,jlo,jhi)
|
||||
if (jhi.ge.jlo) call nga_put(g_diag,ilo,ihi,wia,ihi-ilo+1)
|
||||
|
||||
! Screened Coulomb contribution from A
|
||||
if(.not.ga_create(mt_dbl,nvir2(1),nocc2(1),'tmp',nvir2(1),0,
|
||||
& g_tmp))
|
||||
& call errquit(pname//'could not create temp GA',1,GA_ERR)
|
||||
call ga_dgemm('t','n',nvir2(1),nocc2(1),pars%nri,-1d0,
|
||||
% pars%g_erivv(1),pars%g_erioo(1),0d0,g_tmp)
|
||||
call ga_distribution(g_tmp,pars%me,ilo,ihi,jlo,jhi)
|
||||
call ga_access(g_tmp,ilo,ihi,jlo,jhi,ktmp,ld)
|
||||
! Need to take care about indexing
|
||||
do ij=jlo,jhi
|
||||
jmo = (ij-1)/pars%nocc(1) + 1
|
||||
imo = ij - (jmo-1)*pars%nocc(1)
|
||||
llia = (imo-1)*pars%nvir(1) + 1
|
||||
ulia = imo*pars%nvir(1)
|
||||
do bmo = 1,pars%nvir(1)
|
||||
jb = (jmo-1)*pars%nvir(1) + bmo
|
||||
call ga_put(pars%g_apb,llia,ulia,jb,jb,
|
||||
& dbl_mb(ktmp+(ij-jlo)*nvir2(1)+(bmo-1)*pars%nvir(1)),
|
||||
& pars%nvir(1))
|
||||
enddo
|
||||
enddo
|
||||
call ga_sync()
|
||||
call ga_copy(pars%g_apb,g_amb)
|
||||
call ga_release(g_tmp,ilo,ihi,jlo,jhi)
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
! Screened-Coulomb contribution from B
|
||||
if(.not.ga_create(mt_dbl,npoles,npoles,'tmp',0,npoles,
|
||||
& g_tmp))
|
||||
& call errquit(pname//'could not create temp GA',1,GA_ERR)
|
||||
call ga_dgemm('t','n',npoles,npoles,pars%nri,-1d0,
|
||||
& pars%g_wov(1),pars%g_wov(1),0d0,g_tmp)
|
||||
call ga_add(1d0,g_tmp,1d0,pars%g_apb,pars%g_apb)
|
||||
call ga_add(-1d0,g_tmp,1d0,g_amb,g_amb)
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
if (singlet) then
|
||||
! Hartree contribution
|
||||
call ga_dgemm('t','n',npoles,npoles,pars%nri,4d0,
|
||||
& pars%g_eriov(1),pars%g_eriov(1),1d0,pars%g_apb)
|
||||
endif
|
||||
|
||||
! Diagonal term
|
||||
call ga_add_diagonal(pars%g_apb,g_diag)
|
||||
call ga_add_diagonal(g_amb,g_diag)
|
||||
|
||||
! Destroy ERIs
|
||||
if (.not.ga_destroy(pars%g_erioo(1)))
|
||||
& call errquit(pname//'could not destroy erioo GA',1,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_erivv(1)))
|
||||
& call errquit(pname//'could not destroy erivv GA',1,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_eriov(1)))
|
||||
& call errquit(pname//'could not destroy eriov GA',1,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_wov(1)))
|
||||
& call errquit(pname//'could not destroy wov GA',1,GA_ERR)
|
||||
|
||||
! First step is to get (A-B)^(1/2)
|
||||
call ga_sync()
|
||||
#ifdef SCALAPACK
|
||||
info= ga_cholesky('L',g_amb)
|
||||
#elif defined(PEIGS)
|
||||
call ga_chol(g_amb, g_amb, info)
|
||||
#else
|
||||
call ga_chol_seq(g_amb, g_amb, info)
|
||||
#endif
|
||||
|
||||
! Transform (A+B) matrix with Cholesky factors
|
||||
if(.not.ga_create(mt_dbl,npoles,npoles,'tmp',0,npoles,
|
||||
& g_tmp))
|
||||
& call errquit(pname//'could not create temp GA',1,GA_ERR)
|
||||
call ga_dgemm('n','n',npoles,npoles,npoles,1d0,
|
||||
& pars%g_apb,g_amb,0d0,g_tmp)
|
||||
call ga_dgemm('t','n',npoles,npoles,npoles,1d0,
|
||||
& g_amb,g_tmp,0d0,pars%g_apb)
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
! Diagonalize (A+B) matrix
|
||||
#ifdef PARALLEL_DIAG
|
||||
#ifdef SCALAPACK
|
||||
call dft_scaleig(pars%g_apb,pars%g_apb,omega,0)
|
||||
#else
|
||||
call ga_diag_std(pars%g_apb,pars%g_apb,omega)
|
||||
#endif
|
||||
#else
|
||||
call ga_diag_std_seq(pars%g_apb,pars%g_apb,omega)
|
||||
#endif
|
||||
|
||||
!
|
||||
! Since we transformed the eigenvalue equation to an Hermitian
|
||||
! form, we obtain the square of the neutral excitations
|
||||
! (Equation (32)).
|
||||
!
|
||||
!$omp parallel do
|
||||
do ipole=1,npoles
|
||||
omega(ipole) = dsqrt(omega(ipole))
|
||||
enddo
|
||||
!$omp end parallel do
|
||||
|
||||
! Transform eigenvectors with (A-B)^(1/2)
|
||||
if(.not.ga_create(mt_dbl,npoles,npoles,'tmp',0,npoles,
|
||||
& g_tmp))
|
||||
& call errquit(pname//'could not create temp GA',1,GA_ERR)
|
||||
call ga_dgemm('n','n',npoles,npoles,npoles,1d0,
|
||||
& g_amb,pars%g_apb,0d0,g_tmp)
|
||||
|
||||
! Gather neutral excitations vectors
|
||||
call nga_distribution(g_diag,pars%me,ilo,ihi)
|
||||
if(ihi.ge.ilo) then
|
||||
call nga_put(g_diag,ilo,ihi,1d0/dsqrt(omega(ilo:ihi)),
|
||||
& ihi-ilo+1)
|
||||
endif
|
||||
call ga_sync()
|
||||
call ga_scale_cols(g_tmp,g_diag)
|
||||
|
||||
! Destroy unused GAs
|
||||
call ga_copy(g_tmp,pars%g_apb)
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
if (.not.ga_destroy(g_amb))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
if (.not.ga_destroy(g_diag))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
else
|
||||
nocc2(1) = pars%nocc(1)**2
|
||||
nvir2(1) = pars%nvir(1)**2
|
||||
nocc2(2) = pars%nocc(2)**2
|
||||
nvir2(2) = pars%nvir(2)**2
|
||||
|
||||
call ga_zero(pars%g_apb)
|
||||
|
||||
! Gather eigenvalue differences
|
||||
call ga_distribution(pars%g_eriov(1),pars%me,ilo,ihi,jlo,jhi)
|
||||
call nga_put(g_diag,jlo,jhi,wia(1,1),jhi-jlo+1)
|
||||
call ga_distribution(pars%g_eriov(2),pars%me,ilo,ihi,jlo,jhi)
|
||||
call nga_put(g_diag,jlo+pars%npoles(1),jhi+pars%npoles(1),
|
||||
$ wia(1,2),jhi-jlo+1)
|
||||
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
|
||||
! Add Screened Coulomb contribution from A
|
||||
if (.not.ga_create(mt_dbl,nvir2(isp),nocc2(isp),'tmp',
|
||||
$ nvir2(isp),0,g_tmp))
|
||||
$ call errquit('bse_analytic_tda: could not create tmp',145,
|
||||
$ GA_ERR)
|
||||
|
||||
call ga_dgemm('t','n',nvir2(isp),nocc2(isp),nri,-1d0,
|
||||
$ pars%g_erivv(isp),pars%g_erioo(isp),0d0,g_tmp)
|
||||
call ga_distribution(g_tmp,pars%me,ilo,ihi,jlo,jhi)
|
||||
call ga_access(g_tmp,ilo,ihi,jlo,jhi,ktmp,ihi-ilo+1)
|
||||
do ij=jlo,jhi
|
||||
jmo = (ij-1)/pars%nocc(isp) + 1
|
||||
imo = ij - (jmo-1)*pars%nocc(isp)
|
||||
do bmo = 1,pars%nvir(isp)
|
||||
llia = (imo-1)*pars%nvir(isp) + 1 + off1
|
||||
ulia = imo*pars%nvir(isp) + off1
|
||||
jb = (jmo-1)*pars%nvir(isp) + bmo + off1
|
||||
call ga_put(pars%g_apb,llia,ulia,jb,jb,
|
||||
& dbl_mb(ktmp+(ij-jlo)*nvir2(isp)+
|
||||
& (bmo-1)*pars%nvir(isp)),pars%nvir(isp))
|
||||
enddo
|
||||
enddo
|
||||
call ga_release(g_tmp,ilo,ihi,jlo,jhi)
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',172,GA_ERR)
|
||||
|
||||
! Destroy ERIs
|
||||
if (.not.ga_destroy(pars%g_erioo(isp)))
|
||||
& call errquit(pname//'could not destroy erioo GA',175,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_erivv(isp)))
|
||||
& call errquit(pname//'could not destroy erivv GA',177,GA_ERR)
|
||||
|
||||
call ga_copy_patch('n',pars%g_apb,1+off1,
|
||||
& pars%npoles(isp)+off1,1+off1,pars%npoles(isp)+off1,
|
||||
& g_amb,1+off1,pars%npoles(isp)+off1,1+off1,
|
||||
& pars%npoles(isp)+off1)
|
||||
|
||||
|
||||
! Add Screened Coulomb contribution from B
|
||||
if(.not.ga_create(mt_dbl,pars%npoles(isp),pars%npoles(isp),
|
||||
& 'tmp',0,pars%npoles(isp),g_tmp))
|
||||
& call errquit(pname//'could not create temp GA',234,GA_ERR)
|
||||
call ga_dgemm('t','n',pars%npoles(isp),pars%npoles(isp),
|
||||
& pars%nri,-1d0,pars%g_wov(isp),pars%g_wov(isp),
|
||||
& 0d0,g_tmp)
|
||||
call ga_add_patch(1d0,g_tmp,1,pars%npoles(isp),
|
||||
& 1,pars%npoles(isp),
|
||||
& 1d0,pars%g_apb,1+off1,pars%npoles(isp)+off1,
|
||||
& 1+off1,pars%npoles(isp)+off1,
|
||||
& pars%g_apb,1+off1,pars%npoles(isp)+off1,
|
||||
& 1+off1,pars%npoles(isp)+off1)
|
||||
call ga_add_patch(-1d0,g_tmp,1,pars%npoles(isp),
|
||||
& 1,pars%npoles(isp),
|
||||
& 1d0,g_amb,1+off1,pars%npoles(isp)+off1,
|
||||
& 1+off1,pars%npoles(isp)+off1,
|
||||
& g_amb,1+off1,pars%npoles(isp)+off1,
|
||||
& 1+off1,pars%npoles(isp)+off1)
|
||||
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_wov(isp)))
|
||||
& call errquit(pname//'could not destroy wov GA',1,GA_ERR)
|
||||
enddo
|
||||
|
||||
! Form RPA matrix
|
||||
if (singlet) then
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
do jsp=1,pars%ipol
|
||||
off2 = (jsp-1)*pars%npoles(1)
|
||||
call ga_matmul_patch('t','n',2d0,1d0,
|
||||
$ pars%g_eriov(isp),1,pars%npoles(isp),1,nri,
|
||||
$ pars%g_eriov(jsp),1,nri,1,pars%npoles(jsp),
|
||||
$ pars%g_apb,1+off1,pars%npoles(isp)+off1,
|
||||
$ 1+off2,pars%npoles(jsp)+off2)
|
||||
enddo
|
||||
enddo
|
||||
endif
|
||||
call ga_add_diagonal(pars%g_apb,g_diag)
|
||||
call ga_add_diagonal(g_amb,g_diag)
|
||||
|
||||
|
||||
! Destroy ERIs
|
||||
if (.not.ga_destroy(pars%g_eriov(1)))
|
||||
& call errquit(pname//'could not destroy eriov GA',275,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_eriov(2)))
|
||||
& call errquit(pname//'could not destroy eriov GA',277,GA_ERR)
|
||||
|
||||
|
||||
! First step is to get (A-B)^(1/2)
|
||||
call ga_sync()
|
||||
#ifdef SCALAPACK
|
||||
info= ga_cholesky('L',g_amb)
|
||||
#elif defined(PEIGS)
|
||||
call ga_chol(g_amb, g_amb, info)
|
||||
#else
|
||||
call ga_chol_seq(g_amb, g_amb, info)
|
||||
#endif
|
||||
if (info.ne.0) call errquit("Cholesky Failed",298,GA_ERR)
|
||||
|
||||
! Transform (A+B) matrix with Cholesky factors
|
||||
if(.not.ga_create(mt_dbl,npoles,npoles,'tmp',0,npoles,
|
||||
& g_tmp))
|
||||
& call errquit(pname//'could not create temp GA',1,GA_ERR)
|
||||
call ga_dgemm('n','n',npoles,npoles,npoles,1d0,
|
||||
& pars%g_apb,g_amb,0d0,g_tmp)
|
||||
call ga_dgemm('t','n',npoles,npoles,npoles,1d0,
|
||||
& g_amb,g_tmp,0d0,pars%g_apb)
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
! Diagonalize (A+B) matrix
|
||||
#ifdef PARALLEL_DIAG
|
||||
#ifdef SCALAPACK
|
||||
call dft_scaleig(pars%g_apb,pars%g_apb,omega,0)
|
||||
#else
|
||||
call ga_diag_std(pars%g_apb,pars%g_apb,omega)
|
||||
#endif
|
||||
#else
|
||||
call ga_diag_std_seq(pars%g_apb,pars%g_apb,omega)
|
||||
#endif
|
||||
|
||||
!
|
||||
! Since we transformed the eigenvalue equation to an Hermitian
|
||||
! form, we obtain the square of the neutral excitations
|
||||
! (Equation (32)).
|
||||
!
|
||||
!$omp parallel do
|
||||
do ipole=1,npoles
|
||||
omega(ipole) = dsqrt(omega(ipole))
|
||||
enddo
|
||||
!$omp end parallel do
|
||||
|
||||
! Transform eigenvectors with (A-B)^(1/2)
|
||||
if(.not.ga_create(mt_dbl,npoles,npoles,'tmp',0,npoles,
|
||||
& g_tmp))
|
||||
& call errquit(pname//'could not create temp GA',1,GA_ERR)
|
||||
call ga_dgemm('n','n',npoles,npoles,npoles,1d0,
|
||||
& g_amb,pars%g_apb,0d0,g_tmp)
|
||||
|
||||
! Gather neutral excitations vectors
|
||||
call nga_distribution(g_diag,pars%me,ilo,ihi)
|
||||
if(ihi.ge.ilo) then
|
||||
call nga_put(g_diag,ilo,ihi,1d0/dsqrt(omega(ilo:ihi)),
|
||||
& ihi-ilo+1)
|
||||
endif
|
||||
call ga_sync()
|
||||
call ga_scale_cols(g_tmp,g_diag)
|
||||
call ga_copy(g_tmp,pars%g_apb)
|
||||
|
||||
! Destroy unused GAs
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',344,GA_ERR)
|
||||
if (.not.ga_destroy(g_amb))
|
||||
& call errquit(pname//'could not destroy temp GA',346,GA_ERR)
|
||||
if (.not.ga_destroy(g_diag))
|
||||
& call errquit(pname//'could not destroy temp GA',348,GA_ERR)
|
||||
|
||||
endif
|
||||
|
||||
end subroutine
|
||||
200
src/bsemol/bse_analytic_tda.F
Normal file
200
src/bsemol/bse_analytic_tda.F
Normal file
|
|
@ -0,0 +1,200 @@
|
|||
subroutine bse_analytic_tda(pars,wia,omega,maxpoles,totpoles,
|
||||
$ singlet)
|
||||
implicit none
|
||||
#include "mafdecls.fh"
|
||||
#include "errquit.fh"
|
||||
#include "global.fh"
|
||||
#include "bse.fh"
|
||||
#ifdef SCALAPACK
|
||||
#include "scaleig.fh"
|
||||
integer ga_cholesky,ga_llt_i
|
||||
external ga_cholesky,ga_llt_i
|
||||
#endif
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
|
||||
integer maxpoles, totpoles
|
||||
double precision wia(maxpoles,pars%ipol)
|
||||
double precision omega(totpoles)
|
||||
logical singlet
|
||||
|
||||
character(*),parameter :: pname = 'bse_analytic_tda: '
|
||||
|
||||
integer nri,isp,jsp,npoles
|
||||
integer ilo,ihi,jlo,jhi
|
||||
integer nocc2(2),nvir2(2)
|
||||
integer ipole,imo,jmo,amo,bmo
|
||||
integer k,l,kproc,lproc,klocal,llocal,kglobal,lglobal
|
||||
integer g_diag, g_tmp, g_amb, ktmp, ld, info, i, j
|
||||
integer modtn,tovern,off1,off2
|
||||
double precision alpha
|
||||
integer ij,jb,llia,ulia
|
||||
|
||||
#ifdef USE_OPENMP
|
||||
call util_blas_set_num_threads(pars%iMaxthreads)
|
||||
#endif
|
||||
|
||||
npoles = pars%npoles(1)
|
||||
if (pars%ipol.gt.1) npoles = npoles + pars%npoles(2)
|
||||
nri = pars%nri
|
||||
|
||||
if(.not.ga_create(mt_dbl,totpoles,totpoles,'apb',0,npoles,
|
||||
& pars%g_apb))
|
||||
& call errquit(pname//'could not create A+B GA',0,GA_ERR)
|
||||
if (.not.nga_create(mt_dbl,1,npoles,'wia',0,g_diag))
|
||||
$ call errquit(pname//'could not create wia',0,GA_ERR)
|
||||
|
||||
if (pars%ipol.eq.1) then
|
||||
nocc2(1) = pars%nocc(1)**2
|
||||
nvir2(1) = pars%nvir(1)**2
|
||||
|
||||
call ga_distribution(pars%g_apb,pars%me,ilo,ihi,jlo,jhi)
|
||||
if (jhi.ge.jlo) call nga_put(g_diag,ilo,ihi,wia,ihi-ilo+1)
|
||||
|
||||
|
||||
! Screened Coulomb contribution
|
||||
if(.not.ga_create(mt_dbl,nvir2(1),nocc2(1),'tmp',nvir2(1),0,
|
||||
& g_tmp))
|
||||
& call errquit(pname//'could not create temp GA',1,GA_ERR)
|
||||
call ga_dgemm('t','n',nvir2(1),nocc2(1),pars%nri,-1d0,
|
||||
% pars%g_erivv(1),pars%g_erioo(1),0d0,g_tmp)
|
||||
call ga_distribution(g_tmp,pars%me,ilo,ihi,jlo,jhi)
|
||||
call ga_access(g_tmp,ilo,ihi,jlo,jhi,ktmp,ld)
|
||||
! Need to take care about indexing
|
||||
do ij=jlo,jhi
|
||||
jmo = (ij-1)/pars%nocc(1) + 1
|
||||
imo = ij - (jmo-1)*pars%nocc(1)
|
||||
do bmo = 1,pars%nvir(1)
|
||||
llia = (imo-1)*pars%nvir(1) + 1
|
||||
ulia = imo*pars%nvir(1)
|
||||
jb = (jmo-1)*pars%nvir(1) + bmo
|
||||
call ga_put(pars%g_apb,llia,ulia,jb,jb,
|
||||
& dbl_mb(ktmp+(ij-jlo)*nvir2(1)+(bmo-1)*pars%nvir(1)),
|
||||
& pars%nvir(1))
|
||||
enddo
|
||||
enddo
|
||||
call ga_release(g_tmp,ilo,ihi,jlo,jhi)
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
! Hartree contribution
|
||||
if (singlet) then
|
||||
call ga_dgemm('t','n',npoles,npoles,pars%nri,2d0,
|
||||
& pars%g_eriov(1),pars%g_eriov(1),1d0,pars%g_apb)
|
||||
endif
|
||||
|
||||
! Diagonal term
|
||||
call ga_add_diagonal(pars%g_apb,g_diag)
|
||||
|
||||
! Destroy ERIs
|
||||
if (.not.ga_destroy(pars%g_erioo(1)))
|
||||
& call errquit(pname//'could not destroy erioo GA',1,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_erivv(1)))
|
||||
& call errquit(pname//'could not destroy erivv GA',1,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_eriov(1)))
|
||||
& call errquit(pname//'could not destroy eriov GA',1,GA_ERR)
|
||||
|
||||
|
||||
! Diagonalize (A+B) matrix
|
||||
#ifdef PARALLEL_DIAG
|
||||
#ifdef SCALAPACK
|
||||
call dft_scaleig(pars%g_apb,pars%g_apb,omega,0)
|
||||
#else
|
||||
call ga_diag_std(pars%g_apb,pars%g_apb,omega)
|
||||
#endif
|
||||
#else
|
||||
call ga_diag_std_seq(pars%g_apb,pars%g_apb,omega)
|
||||
#endif
|
||||
|
||||
! Destroy unused GAs
|
||||
if (.not.ga_destroy(g_diag))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
else
|
||||
nocc2(1) = pars%nocc(1)**2
|
||||
nvir2(1) = pars%nvir(1)**2
|
||||
nocc2(2) = pars%nocc(2)**2
|
||||
nvir2(2) = pars%nvir(2)**2
|
||||
|
||||
|
||||
! Gather eigenvalue differences
|
||||
call ga_distribution(pars%g_eriov(1),pars%me,ilo,ihi,jlo,jhi)
|
||||
call nga_put(g_diag,jlo,jhi,wia(1,1),jhi-jlo+1)
|
||||
call ga_distribution(pars%g_eriov(2),pars%me,ilo,ihi,jlo,jhi)
|
||||
call nga_put(g_diag,jlo+pars%npoles(1),jhi+pars%npoles(1),
|
||||
$ wia(1,2),jhi-jlo+1)
|
||||
|
||||
! Form RPA matrix
|
||||
if (.not.singlet) then
|
||||
call ga_zero(pars%g_apb)
|
||||
else
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
do jsp=1,pars%ipol
|
||||
off2 = (jsp-1)*pars%npoles(1)
|
||||
call ga_matmul_patch('t','n',1d0,0d0,
|
||||
$ pars%g_eriov(isp),1,pars%npoles(isp),1,nri,
|
||||
$ pars%g_eriov(jsp),1,nri,1,pars%npoles(jsp),
|
||||
$ pars%g_apb,1+off1,pars%npoles(isp)+off1,
|
||||
$ 1+off2,pars%npoles(jsp)+off2)
|
||||
enddo
|
||||
enddo
|
||||
endif
|
||||
call ga_add_diagonal(pars%g_apb,g_diag)
|
||||
|
||||
!Add Screened Coulomb contribution
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
if (.not.ga_destroy(pars%g_eriov(isp)))
|
||||
& call errquit(pname//'could not destroy eriov GA',147,GA_ERR)
|
||||
|
||||
if (.not.ga_create(mt_dbl,nvir2(isp),nocc2(isp),'tmp',
|
||||
$ nvir2(isp),0,g_tmp))
|
||||
$ call errquit('bse_analytic_tda: could not create tmp',145,
|
||||
$ GA_ERR)
|
||||
|
||||
call ga_dgemm('t','n',nvir2(isp),nocc2(isp),nri,-1d0,
|
||||
$ pars%g_erivv(isp),pars%g_erioo(isp),0d0,g_tmp)
|
||||
call ga_distribution(g_tmp,pars%me,ilo,ihi,jlo,jhi)
|
||||
call ga_access(g_tmp,ilo,ihi,jlo,jhi,ktmp,ihi-ilo+1)
|
||||
do ij=jlo,jhi
|
||||
jmo = (ij-1)/pars%nocc(isp) + 1
|
||||
imo = ij - (jmo-1)*pars%nocc(isp)
|
||||
do bmo = 1,pars%nvir(isp)
|
||||
llia = (imo-1)*pars%nvir(isp) + 1 + off1
|
||||
ulia = imo*pars%nvir(isp) + off1
|
||||
jb = (jmo-1)*pars%nvir(isp) + bmo + off1
|
||||
call ga_acc(pars%g_apb,llia,ulia,jb,jb,
|
||||
& dbl_mb(ktmp+(ij-jlo)*nvir2(isp)+
|
||||
& (bmo-1)*pars%nvir(isp)),pars%nvir(isp),1d0)
|
||||
enddo
|
||||
enddo
|
||||
call ga_release(g_tmp,ilo,ihi,jlo,jhi)
|
||||
if (.not.ga_destroy(g_tmp))
|
||||
& call errquit(pname//'could not destroy temp GA',172,GA_ERR)
|
||||
! Destroy ERIs
|
||||
if (.not.ga_destroy(pars%g_erioo(isp)))
|
||||
& call errquit(pname//'could not destroy erioo GA',175,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_erivv(isp)))
|
||||
& call errquit(pname//'could not destroy erivv GA',177,GA_ERR)
|
||||
enddo
|
||||
|
||||
! Diagonalize (A+B) matrix
|
||||
#ifdef PARALLEL_DIAG
|
||||
#ifdef SCALAPACK
|
||||
call dft_scaleig(pars%g_apb,pars%g_apb,omega,0)
|
||||
#else
|
||||
call ga_diag_std(pars%g_apb,pars%g_apb,omega)
|
||||
#endif
|
||||
#else
|
||||
call ga_diag_std_seq(pars%g_apb,pars%g_apb,omega)
|
||||
#endif
|
||||
|
||||
! Destroy unused GAs
|
||||
if (.not.ga_destroy(g_diag))
|
||||
& call errquit(pname//'could not destroy temp GA',1,GA_ERR)
|
||||
|
||||
|
||||
endif
|
||||
|
||||
end subroutine
|
||||
78
src/bsemol/bse_buildw.F
Normal file
78
src/bsemol/bse_buildw.F
Normal file
|
|
@ -0,0 +1,78 @@
|
|||
subroutine bse_buildW(pars,wia,w,nmo,nri,maxpoles,ipol)
|
||||
implicit none
|
||||
#include "errquit.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "global.fh"
|
||||
#include "stdio.fh"
|
||||
#include "bse.fh"
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
integer nmo, nri, maxpoles, ipol
|
||||
|
||||
character(*), parameter :: pname = 'bse_buildw: '
|
||||
|
||||
double precision wia(maxpoles,ipol)
|
||||
double precision w(nri,nri)
|
||||
|
||||
integer :: ovlo,ovhi,kov(2),isp
|
||||
integer :: ldpi, sizepi, iri, jri
|
||||
integer :: alloc, info, ld, ipole
|
||||
double precision,allocatable :: factor(:,:)
|
||||
|
||||
allocate(factor(maxpoles,ipol),stat=alloc)
|
||||
if (alloc.ne.0)
|
||||
& call errquit(pname//'allocation failed',0,MA_ERR)
|
||||
C
|
||||
C Decide the leading dimension of the \Pi array according to
|
||||
C the rectangular full-packed (RFP) format
|
||||
C
|
||||
if (mod(nri,2).eq.0) then
|
||||
ldpi = nri + 1
|
||||
else
|
||||
ldpi = nri
|
||||
endif
|
||||
sizepi = (nri*(nri+1))/2
|
||||
|
||||
! Get pointer to local integrals
|
||||
do isp=1,pars%ipol
|
||||
kov(isp) = 1
|
||||
if (pars%mynpoles(isp).lt.1) cycle
|
||||
|
||||
ovlo = pars%ovlo(isp)
|
||||
ovhi = pars%ovhi(isp)
|
||||
call ga_access(pars%g_eriov(isp),1,nri,ovlo,ovhi,kov(isp),ld)
|
||||
|
||||
!$omp parallel do
|
||||
do ipole=1,pars%mynpoles(isp)
|
||||
factor(ipole,isp) = dsqrt(1.0d0/wia(ipole,isp))
|
||||
enddo
|
||||
!$omp end parallel do
|
||||
enddo
|
||||
|
||||
! Build the Polarizability matrix and transform it to the
|
||||
! dielectric matrix \epsilon = 1 - \pi
|
||||
call gw_cdgw_buildpi('w',dbl_mb(kov(1)),dbl_mb(kov(2)),w,factor,
|
||||
$ ldpi,nri,pars%mynpoles,pars%me.eq.0,
|
||||
$ ipol,maxpoles)
|
||||
call ga_dgop(1038,w,sizepi,'+')
|
||||
|
||||
!Factorize dieletric matrix for further use
|
||||
call dpftrf('n','l',nri,w,info)
|
||||
if(info.ne.0)
|
||||
& call errquit(pname//'Failed to decompose dielectric',info,0)
|
||||
|
||||
!Deallocate
|
||||
deallocate(factor,stat=alloc)
|
||||
if (alloc.ne.0)
|
||||
& call errquit(pname//'deallocation failed',0,MA_ERR)
|
||||
|
||||
!Release GAs
|
||||
do isp=1,ipol
|
||||
if (pars%mynpoles(isp).lt.1) cycle
|
||||
ovlo = pars%ovlo(isp)
|
||||
ovhi = pars%ovhi(isp)
|
||||
call ga_release(pars%g_eriov(isp),1,nri,ovlo,ovhi)
|
||||
enddo
|
||||
|
||||
end
|
||||
|
||||
569
src/bsemol/bse_davidson.F
Normal file
569
src/bsemol/bse_davidson.F
Normal file
|
|
@ -0,0 +1,569 @@
|
|||
subroutine bse_davidson(pars)
|
||||
|
||||
implicit none
|
||||
|
||||
#include "global.fh"
|
||||
#include "errquit.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "bse.fh"
|
||||
#include "cdft.fh"
|
||||
type(bse_params_t) :: pars
|
||||
character(*), parameter :: pname = 'bse_analytic: '
|
||||
|
||||
integer lSize,isp
|
||||
integer l_omega, k_omega
|
||||
integer l_dia,k_dia
|
||||
integer lW,kW
|
||||
integer maxpoles,maxnpoles,totnpoles,nri
|
||||
|
||||
integer ntrials,g_trials,npoles,ndim
|
||||
integer g_x,g_y,g_ym,g_s,g_ks,g_mks,g_w,g_kw,g_mkw,off1
|
||||
integer g_wia,g_tmp,info,lwork,kwork,idum,mdim,k_trials
|
||||
integer k_x,k_y,k_ym,k_s,k_ks,k_mks,k_w,k_kw,k_mkw,k_r,l_r
|
||||
integer l_map,k_map,l_tmp1,k_tmp1,l_tmp2,k_tmp2,k_x2,k_trials2
|
||||
integer k_w2,l_tmp3,k_tmp3,l_tmp4,k_tmp4,k_kw2,k_mkw2
|
||||
integer myld,vkv,l_vkv,myhi,mylo,l_ksw,ksw,jdum,myld2
|
||||
integer subscript(2)
|
||||
double precision minimum,norm(pars%nroots),rdum,factor
|
||||
|
||||
integer inode,nprocs,nrestart
|
||||
integer ipole,jpole,itrial,imo,amo,idavidson
|
||||
logical found,dohartree,stat,debug
|
||||
double precision,external :: dasum
|
||||
|
||||
!debug = pars%me.eq.0
|
||||
debug = .false.
|
||||
|
||||
! synchronize parallel processes
|
||||
call ga_sync()
|
||||
|
||||
!initialize useful variables
|
||||
nri = pars%nri
|
||||
maxpoles = maxval(pars%mynpoles(1:pars%ipol))
|
||||
maxnpoles = maxval(pars%npoles(1:pars%ipol))
|
||||
totnpoles = maxnpoles*pars%ipol
|
||||
npoles = sum(pars%npoles(1:pars%ipol))
|
||||
ntrials = pars%nroots*pars%nspace
|
||||
nprocs = ga_nnodes()
|
||||
|
||||
! allocation
|
||||
if(.not.ma_push_get(mt_dbl,totnpoles,'dia',l_dia,k_dia))
|
||||
& call errquit(pname//'failed to allocate Delta_ia',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,npoles,'omega',l_omega,k_omega))
|
||||
& call errquit(pname//'failed to allocate omega',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,npoles,'residual',l_r,k_r))
|
||||
& call errquit(pname//'failed to allocate residual',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,maxpoles*pars%nmax,'tmp1',
|
||||
& l_tmp1,k_tmp1))
|
||||
& call errquit(pname//'failed to allocate tmp1',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,maxpoles*pars%nmax,'tmp2',
|
||||
& l_tmp2,k_tmp2))
|
||||
& call errquit(pname//'failed to allocate tmp2',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,maxpoles*pars%nmax,'tmp3',
|
||||
& l_tmp3,k_tmp3))
|
||||
& call errquit(pname//'failed to allocate tmp3',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,maxpoles*pars%nmax,'tmp4',
|
||||
& l_tmp4,k_tmp4))
|
||||
& call errquit(pname//'failed to allocate tmp4',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_int,2*ga_nnodes()+1,'map',l_map,k_map))
|
||||
& call errquit(pname//'failed to allocate map',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,ntrials*ntrials,'vkv',l_vkv,vkv))
|
||||
& call errquit(pname//'failed to allocate vkv',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,nri**2,'W',lW,kW))
|
||||
& call errquit(pname//'failed to allocate W',0,MA_ERR)
|
||||
|
||||
! Get map for irreg array
|
||||
pars%nblocks = ga_nnodes()
|
||||
do inode=0,ga_nnodes()-1
|
||||
idum = pars%ovlo(1)
|
||||
call ga_brdcst(mt_int,idum,ma_sizeof(mt_int,1,mt_byte),inode)
|
||||
int_mb(k_map+inode) = idum
|
||||
enddo
|
||||
if (pars%ipol.gt.1) then
|
||||
pars%nblocks = pars%nblocks + ga_nnodes()
|
||||
do inode=0,ga_nnodes()-1
|
||||
idum = pars%ovlo(2)
|
||||
call ga_brdcst(mt_int,idum,ma_sizeof(mt_int,1,mt_byte),inode)
|
||||
int_mb(k_map+ga_nnodes()+inode) = idum + pars%npoles(1)
|
||||
enddo
|
||||
endif
|
||||
int_mb(k_map+pars%nblocks) = 1
|
||||
|
||||
! Allocate GA arrays
|
||||
g_x = ga_create_handle()
|
||||
call ga_set_data(g_x,2,(/npoles,ntrials/),mt_dbl)
|
||||
call ga_set_array_name(g_x,'x')
|
||||
call ga_set_tiled_irreg_proc_grid(g_x,int_mb(k_map),
|
||||
$ (/pars%nblocks,1/),(/ga_nnodes(),1/))
|
||||
stat = ga_allocate(g_x)
|
||||
stat = stat.and.ga_duplicate(g_x,g_y,'y').and.
|
||||
$ ga_duplicate(g_x,g_ym,'ym').and.
|
||||
$ ga_duplicate(g_x,g_s,'s').and.
|
||||
$ ga_duplicate(g_x,g_ks,'ks').and.
|
||||
$ ga_duplicate(g_x,g_mks,'mks').and.
|
||||
$ ga_duplicate(g_x,g_trials,'trials')
|
||||
if(.not.stat) call errquit('ga create failed',103,GA_ERR)
|
||||
call ga_zero(g_y)
|
||||
call ga_zero(g_ym)
|
||||
call ga_zero(g_s)
|
||||
call ga_zero(g_ks)
|
||||
call ga_zero(g_mks)
|
||||
call ga_zero(g_trials)
|
||||
|
||||
! W, K*W, and M*K*W are fixed to nroots columns
|
||||
g_w = ga_create_handle()
|
||||
call ga_set_data(g_w, 2, (/npoles, pars%nroots/), mt_dbl)
|
||||
call ga_set_array_name(g_w, 'w')
|
||||
call ga_set_tiled_irreg_proc_grid(g_w, int_mb(k_map),
|
||||
$ (/pars%nblocks, 1/), (/ga_nnodes(), 1/))
|
||||
stat = ga_allocate(g_w)
|
||||
stat = stat .and. ga_duplicate(g_w, g_kw, 'kw') .and.
|
||||
$ ga_duplicate(g_w, g_mkw, 'mkw')
|
||||
if(.not.stat) call errquit('Could not allocate g_w', 0, GA_ERR)
|
||||
call ga_zero(g_w)
|
||||
call ga_zero(g_kw)
|
||||
call ga_zero(g_mkw)
|
||||
|
||||
g_wia = ga_create_handle()
|
||||
call ga_set_data(g_wia,1,npoles,mt_dbl)
|
||||
call ga_set_array_name(g_wia,'wia')
|
||||
call ga_set_tiled_irreg_proc_grid(g_wia,int_mb(k_map),
|
||||
$ pars%nblocks,ga_nnodes())
|
||||
stat = stat.and.ga_allocate(g_wia)
|
||||
if(.not.stat) call errquit('ga create failed',110,GA_ERR)
|
||||
call ga_zero(g_wia)
|
||||
|
||||
|
||||
! obtain "old" eigenvalue difference to build W
|
||||
do isp=1,pars%ipol
|
||||
lSize = (isp-1)*maxpoles
|
||||
call bse_get_eia(dbl_mb(pars%k_mf_evals+(isp-1)*nbf_ao),
|
||||
& dbl_mb(k_dia+lSize),pars%nocc(isp),
|
||||
& pars%nvir(isp),pars%ovlo(isp),pars%ovhi(isp))
|
||||
enddo
|
||||
|
||||
! obtain screened Coulomb matrix
|
||||
call bse_buildw(pars,dbl_mb(k_dia),dbl_mb(kW),pars%nmo,nri,
|
||||
& maxpoles,pars%ipol)
|
||||
|
||||
! transform ERIs using screened Coulomb matrix
|
||||
call bse_wmn(pars,dbl_mb(kW),pars%nmo,nri,pars%ipol)
|
||||
if(.not.ma_chop_stack(lW))
|
||||
& call errquit(pname//'failed to chop stack',0,MA_ERR)
|
||||
|
||||
! obtain "GW" eigenvalue difference to build Casida matrix
|
||||
do isp=1,pars%ipol
|
||||
lSize = (isp-1)*maxpoles
|
||||
call bse_get_eia(dbl_mb(pars%k_gw_evals+(isp-1)*nbf_ao),
|
||||
& dbl_mb(k_dia+lSize),pars%nocc(isp),
|
||||
& pars%nvir(isp),pars%ovlo(isp),
|
||||
& pars%ovhi(isp))
|
||||
enddo
|
||||
|
||||
! Get diagonal of Casida matrix
|
||||
call bse_davidson_diagonal(pars,dbl_mb(k_dia),maxpoles)
|
||||
|
||||
! Get Eigenvalue differences
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
lSize = (isp-1)*maxpoles
|
||||
call nga_put(g_wia,pars%ovlo(isp)+off1,pars%ovhi(isp)+off1,
|
||||
$ dbl_mb(k_dia+lSize),pars%ovhi(isp)-pars%ovlo(isp)+1)
|
||||
enddo
|
||||
|
||||
! Get guess vectos
|
||||
if(pars%me.eq.0)
|
||||
$ call bse_davidson_guess(pars,dbl_mb(k_dia),g_trials,npoles,
|
||||
$ ntrials,pars%nroots)
|
||||
call ga_brdcst(mt_dbl,dbl_mb(k_dia),
|
||||
$ ma_sizeof(mt_dbl,npoles,mt_byte),0)
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
call dcopy(pars%mynpoles(isp),
|
||||
$ dbl_mb(k_dia+off1+pars%ovlo(isp)-1),
|
||||
$ 1,dbl_mb(k_dia+(isp-1)*pars%mynpoles(1)),1)
|
||||
enddo
|
||||
|
||||
|
||||
do isp=1,pars%ipol
|
||||
if(.not.ga_create(mt_dbl,nri*pars%nvir(isp),pars%nvir(isp),
|
||||
$ 'g_erivv',0,pars%nvir(isp),g_tmp))
|
||||
$ call errquit('could not create GA',0,GA_ERR)
|
||||
call ga_zero(g_tmp)
|
||||
call ga_copy_patch('n',pars%g_erivv(isp),1,nri,1,
|
||||
$ pars%nvir(isp)*pars%nvir(isp),
|
||||
$ g_tmp,1,nri*pars%nvir(isp),1,
|
||||
$ pars%nvir(isp))
|
||||
if(.not.ga_destroy(pars%g_erivv(isp)))
|
||||
$ call errquit('could not destroy GA',0,GA_ERR)
|
||||
pars%g_erivv(isp) = g_tmp
|
||||
enddo
|
||||
|
||||
! Start K-Davidson Algorithm
|
||||
idavidson = 0
|
||||
|
||||
! Restarts should land here
|
||||
1 continue
|
||||
|
||||
! K-orthogonalization
|
||||
call bse_davidson_kvec(pars,g_trials,g_wia,g_x,npoles,
|
||||
$ ntrials,pars%singlet,1d0)
|
||||
|
||||
call nga_access_block_grid(g_x, (/pars%me,0/), k_x, myld)
|
||||
call nga_access_block_grid(g_trials, (/pars%me,0/),k_trials,myld)
|
||||
if (pars%ipol.eq.1) then
|
||||
myld2 = 0
|
||||
call dcopy(myld*ntrials,dbl_mb(k_x),1,dbl_mb(k_tmp1),1)
|
||||
call dcopy(myld*ntrials,dbl_mb(k_trials),1,dbl_mb(k_tmp2),1)
|
||||
else
|
||||
call nga_access_block_grid(g_x,(/pars%me+nprocs,0/),k_x2,myld2)
|
||||
call nga_access_block_grid(g_trials,(/pars%me+nprocs,0/),
|
||||
$ k_trials2,myld2)
|
||||
do itrial=1,ntrials
|
||||
call dcopy(myld,dbl_mb(k_x+(itrial-1)*myld),1,
|
||||
$ dbl_mb(k_tmp1+(itrial-1)*(myld+myld2)),1)
|
||||
call dcopy(myld2,dbl_mb(k_x2+(itrial-1)*myld2),1,
|
||||
$ dbl_mb(k_tmp1+(itrial-1)*(myld+myld2)+myld),1)
|
||||
call dcopy(myld,dbl_mb(k_trials+(itrial-1)*myld),1,
|
||||
$ dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)),1)
|
||||
call dcopy(myld2,dbl_mb(k_trials2+(itrial-1)*myld2),1,
|
||||
$ dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)+myld),1)
|
||||
enddo
|
||||
endif
|
||||
|
||||
call bse_davidson_kortho(pars,dbl_mb(k_tmp2),dbl_mb(k_tmp1),
|
||||
$ dbl_mb(k_tmp1),dbl_mb(vkv),dbl_mb(k_tmp1),
|
||||
$ myld+myld2,ntrials,mdim,.true.)
|
||||
|
||||
if (pars%ipol.eq.1) then
|
||||
call dcopy(myld*ntrials,dbl_mb(k_tmp2),1,dbl_mb(k_trials),1)
|
||||
else
|
||||
do itrial=1,ntrials
|
||||
call dcopy(myld,dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)),1,
|
||||
$ dbl_mb(k_trials+(itrial-1)*myld),1)
|
||||
call dcopy(myld2,dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)+myld),
|
||||
$ 1,dbl_mb(k_trials2+(itrial-1)*myld2),1)
|
||||
enddo
|
||||
call nga_release_update_block_grid(g_trials,
|
||||
$ (/pars%me+nprocs,0/))
|
||||
call nga_release_block_grid(g_x,(/pars%me+nprocs,0/))
|
||||
endif
|
||||
call nga_release_update_block_grid(g_trials,(/pars%me,0/))
|
||||
call nga_release_block_grid(g_x,(/pars%me,0/))
|
||||
|
||||
! Starting vectors
|
||||
call ga_copy(g_trials,g_x)
|
||||
call bse_davidson_kvec(pars,g_x,g_wia,g_y,npoles,
|
||||
$ ntrials,pars%singlet,1d0)
|
||||
call bse_davidson_kvec(pars,g_y,g_wia,g_ym,npoles,
|
||||
$ ntrials,.false.,-1d0)
|
||||
call ga_copy(g_x,g_s)
|
||||
call ga_copy(g_y,g_ks)
|
||||
call ga_copy(g_ym,g_mks)
|
||||
|
||||
if (pars%me.eq.0 .and. idavidson.eq.0) then
|
||||
write(*,*)
|
||||
write(*,*) "--------------------------------"
|
||||
write(*,*) " Iter Ntrials Nconv Res "
|
||||
write(*,*) "--------------------------------"
|
||||
endif
|
||||
|
||||
2 idavidson = idavidson + 1
|
||||
if(idavidson.gt.pars%maxiter) goto 3
|
||||
|
||||
! Build projected matrix
|
||||
call dcopy(ntrials*ntrials,0d0,0,dbl_mb(vkv),1)
|
||||
do isp=1,pars%ipol
|
||||
subscript = (/pars%me+(isp-1)*ga_nnodes(),0/)
|
||||
call nga_access_block_grid(g_ks,subscript,k_ks,myld)
|
||||
call nga_access_block_grid(g_mks,subscript,k_mks,myld)
|
||||
call dgemm('t','n',ntrials,ntrials,myld,1d0,dbl_mb(k_ks),
|
||||
$ myld,dbl_mb(k_mks),myld,1d0,dbl_mb(vkv),ntrials)
|
||||
call nga_release_block_grid(g_ks,subscript)
|
||||
call nga_release_block_grid(g_mks,subscript)
|
||||
enddo
|
||||
call ga_dgop((/194/),dbl_mb(vkv),ntrials**2,'+')
|
||||
if (debug) call ma_print(dbl_mb(vkv),ntrials,ntrials,'Proj')
|
||||
|
||||
! Diagonalize projected matrix
|
||||
call dsyev('v','l',ntrials,dbl_mb(vkv),ntrials,dbl_mb(k_omega),
|
||||
$ rdum,-1,info)
|
||||
if(.not.ma_push_get(mt_dbl,int(rdum),'work',lwork,kwork))
|
||||
& call errquit(pname//'failed to allocate work',199,MA_ERR)
|
||||
call dsyev('v','l',ntrials,dbl_mb(vkv),ntrials,dbl_mb(k_omega),
|
||||
$ dbl_mb(kwork),int(rdum),info)
|
||||
if(.not.ma_chop_stack(lwork))
|
||||
& call errquit(pname//'failed to allocate work',192,MA_ERR)
|
||||
if (info.ne.0)
|
||||
$ call errquit(pname//'Diagonalization failed',206,0)
|
||||
if (debug) call ma_print(dbl_mb(vkv),ntrials,ntrials,'Eigen')
|
||||
if (debug) call ma_print(dbl_mb(k_omega),ntrials,1,'Omega')
|
||||
|
||||
|
||||
! Project back vectors
|
||||
do isp=1,pars%ipol
|
||||
subscript = (/pars%me+(isp-1)*nprocs,0/)
|
||||
call nga_access_block_grid(g_s,subscript,k_s,myld)
|
||||
call nga_access_block_grid(g_ks,subscript,k_ks,myld)
|
||||
call nga_access_block_grid(g_mks,subscript,k_mks,myld)
|
||||
call nga_access_block_grid(g_x,subscript,k_x,myld)
|
||||
call nga_access_block_grid(g_y,subscript,k_y,myld)
|
||||
call nga_access_block_grid(g_ym,subscript,k_ym,myld)
|
||||
|
||||
call dgemm('n','n',myld,ntrials,ntrials,1d0,dbl_mb(k_s),myld,
|
||||
$ dbl_mb(vkv),ntrials,0d0,dbl_mb(k_x),myld)
|
||||
call dgemm('n','n',myld,ntrials,ntrials,1d0,dbl_mb(k_ks),myld,
|
||||
$ dbl_mb(vkv),ntrials,0d0,dbl_mb(k_y),myld)
|
||||
call dgemm('n','n',myld,ntrials,ntrials,1d0,dbl_mb(k_mks),
|
||||
$ myld,dbl_mb(vkv),ntrials,0d0,dbl_mb(k_ym),myld)
|
||||
|
||||
call nga_release_block_grid(g_s,subscript)
|
||||
call nga_release_block_grid(g_ks,subscript)
|
||||
call nga_release_block_grid(g_mks,subscript)
|
||||
call nga_release_update_block_grid(g_x,subscript)
|
||||
call nga_release_update_block_grid(g_y,subscript)
|
||||
call nga_release_update_block_grid(g_ym,subscript)
|
||||
enddo
|
||||
|
||||
! Get residual
|
||||
subscript = (/pars%me,0/)
|
||||
call nga_access_block_grid(g_x,subscript,k_x,myld)
|
||||
call nga_access_block_grid(g_ym,subscript,k_trials,myld)
|
||||
if (pars%ipol.eq.1) then
|
||||
myld2 = 0
|
||||
call dcopy(myld*ntrials,dbl_mb(k_x),1,dbl_mb(k_tmp1),1)
|
||||
call dcopy(myld*ntrials,dbl_mb(k_trials),1,dbl_mb(k_tmp2),1)
|
||||
else
|
||||
subscript = (/pars%me+nprocs,0/)
|
||||
call nga_access_block_grid(g_x,subscript,k_x2,myld2)
|
||||
call nga_access_block_grid(g_ym,subscript,k_trials2,myld2)
|
||||
do itrial=1,pars%nroots
|
||||
call dcopy(myld,dbl_mb(k_x+(itrial-1)*myld),1,
|
||||
$ dbl_mb(k_tmp1+(itrial-1)*(myld+myld2)),1)
|
||||
call dcopy(myld2,dbl_mb(k_x2+(itrial-1)*myld2),1,
|
||||
$ dbl_mb(k_tmp1+(itrial-1)*(myld+myld2)+myld),1)
|
||||
call dcopy(myld,dbl_mb(k_trials+(itrial-1)*myld),1,
|
||||
$ dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)),1)
|
||||
call dcopy(myld2,dbl_mb(k_trials2+(itrial-1)*myld2),1,
|
||||
$ dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)+myld),1)
|
||||
enddo
|
||||
endif
|
||||
|
||||
call bse_davidson_residual(pars%nroots,ntrials,dbl_mb(k_omega),
|
||||
$ dbl_mb(k_tmp1),dbl_mb(k_tmp2),dbl_mb(k_r),
|
||||
$ dbl_mb(k_dia),dbl_mb(k_tmp3),myld+myld2,ndim,norm)
|
||||
|
||||
call nga_release_block_grid(g_x,(/pars%me,0/))
|
||||
call nga_release_block_grid(g_ym,(/pars%me,0/))
|
||||
call nga_access_block_grid(g_w,(/pars%me,0/),k_w,myld)
|
||||
if(pars%ipol.eq.1) then
|
||||
call dcopy(pars%nroots*myld,dbl_mb(k_tmp3),1,dbl_mb(k_w),1)
|
||||
else
|
||||
call nga_release_block_grid(g_x,subscript)
|
||||
call nga_release_block_grid(g_ym,subscript)
|
||||
call nga_access_block_grid(g_w,subscript,k_w2,myld2)
|
||||
do itrial=1,pars%nroots
|
||||
call dcopy(myld,dbl_mb(k_tmp3+(itrial-1)*(myld+myld2)),1,
|
||||
$ dbl_mb(k_w+(itrial-1)*myld),1)
|
||||
call dcopy(myld2,dbl_mb(k_tmp3+
|
||||
$ (itrial-1)*(myld+myld2)+myld),1,
|
||||
$ dbl_mb(k_w2+(itrial-1)*myld2),1)
|
||||
enddo
|
||||
call nga_release_update_block_grid(g_w,subscript)
|
||||
endif
|
||||
call nga_release_update_block_grid(g_w,(/pars%me,0/))
|
||||
|
||||
! Print info
|
||||
if (pars%me.eq.0) then
|
||||
write(*,9010) idavidson,ntrials,pars%nroots-ndim,maxval(norm)
|
||||
endif
|
||||
9010 format(I5,3X,I5,3X,I5,G16.5)
|
||||
|
||||
! All residuals are below the threshold
|
||||
if (ndim.eq.0) goto 3
|
||||
|
||||
! Restart calculation if dimensions exceeded
|
||||
if (ndim+ntrials.gt.pars%nmax) then
|
||||
nrestart = min(ntrials, pars%nroots*pars%nspace)
|
||||
call bse_davidson_restart(ntrials,nrestart,npoles,
|
||||
$ g_x,g_y,g_ym,g_s,g_ks,g_mks,g_trials,l_vkv,vkv,
|
||||
$ int_mb(k_map),pars%nblocks)
|
||||
goto 1
|
||||
endif
|
||||
|
||||
! Project-out current vectors
|
||||
if(.not.ma_push_get(mt_dbl,ndim*ntrials,'ksw',l_ksw,ksw))
|
||||
$ call errquit(pname//'ksw allocation failed',264,MA_ERR)
|
||||
call dcopy(ntrials*ndim,0d0,0,dbl_mb(ksw),1)
|
||||
do isp=1,pars%ipol
|
||||
subscript = (/pars%me+(isp-1)*nprocs,0/)
|
||||
call nga_access_block_grid(g_ks,subscript,k_ks,myld)
|
||||
call nga_access_block_grid(g_w,subscript,k_w,myld)
|
||||
call dgemm('t','n',ntrials,ndim,myld,1d0,dbl_mb(k_ks),myld,
|
||||
$ dbl_mb(k_w),myld,1d0,dbl_mb(ksw),ntrials)
|
||||
call nga_release_block_grid(g_ks,subscript)
|
||||
call nga_release_block_grid(g_w,subscript)
|
||||
enddo
|
||||
call ga_dgop((/268/),dbl_mb(ksw),ntrials*ndim,'+')
|
||||
|
||||
do isp=1,pars%ipol
|
||||
subscript = (/pars%me+(isp-1)*nprocs,0/)
|
||||
call nga_access_block_grid(g_s,subscript,k_s,myld)
|
||||
call nga_access_block_grid(g_w,subscript,k_w,myld)
|
||||
call dgemm('n','n',myld,ndim,ntrials,-1d0,dbl_mb(k_s),myld,
|
||||
$ dbl_mb(ksw),ntrials,1d0,dbl_mb(k_w),myld)
|
||||
call nga_release_update_block_grid(k_w,subscript)
|
||||
call nga_release_block_grid(k_s,subscript)
|
||||
enddo
|
||||
if(.not.ma_chop_stack(l_ksw))
|
||||
$ call errquit(pname//'ksw deallocation failed',272,MA_ERR)
|
||||
|
||||
|
||||
! K-orthogonalization
|
||||
call bse_davidson_kvec(pars,g_w,g_wia,g_kw,npoles,
|
||||
$ ndim,pars%singlet,1d0)
|
||||
call bse_davidson_kvec(pars,g_kw,g_wia,g_mkw,npoles,
|
||||
$ ndim,.false.,-1d0)
|
||||
|
||||
subscript = (/pars%me,0/)
|
||||
call nga_access_block_grid(g_w,subscript,k_w,myld)
|
||||
call nga_access_block_grid(g_kw,subscript,k_kw,myld)
|
||||
call nga_access_block_grid(g_mkw,subscript,k_mkw,myld)
|
||||
if (pars%ipol.eq.1) then
|
||||
myld2 = 0
|
||||
call dcopy(ndim*myld,dbl_mb(k_w),1,dbl_mb(k_tmp1),1)
|
||||
call dcopy(ndim*myld,dbl_mb(k_kw),1,dbl_mb(k_tmp2),1)
|
||||
call dcopy(ndim*myld,dbl_mb(k_mkw),1,dbl_mb(k_tmp3),1)
|
||||
else
|
||||
subscript = (/pars%me+nprocs,0/)
|
||||
call nga_access_block_grid(g_w,subscript,k_w2,myld2)
|
||||
call nga_access_block_grid(g_kw,subscript,k_kw2,myld2)
|
||||
call nga_access_block_grid(g_mkw,subscript,k_mkw2,myld2)
|
||||
do itrial=1,ndim
|
||||
call dcopy(myld,dbl_mb(k_w+(itrial-1)*myld),1,
|
||||
$ dbl_mb(k_tmp1+(itrial-1)*(myld+myld2)),1)
|
||||
call dcopy(myld2,dbl_mb(k_w2+(itrial-1)*myld2),1,
|
||||
$ dbl_mb(k_tmp1+(itrial-1)*(myld+myld2)+myld),1)
|
||||
call dcopy(myld,dbl_mb(k_kw+(itrial-1)*myld),1,
|
||||
$ dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)),1)
|
||||
call dcopy(myld2,dbl_mb(k_kw2+(itrial-1)*myld2),1,
|
||||
$ dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)+myld),1)
|
||||
call dcopy(myld,dbl_mb(k_mkw+(itrial-1)*myld),1,
|
||||
$ dbl_mb(k_tmp3+(itrial-1)*(myld+myld2)),1)
|
||||
call dcopy(myld2,dbl_mb(k_mkw2+(itrial-1)*myld2),1,
|
||||
$ dbl_mb(k_tmp3+(itrial-1)*(myld+myld2)+myld),1)
|
||||
enddo
|
||||
endif
|
||||
|
||||
call bse_davidson_kortho(pars,dbl_mb(k_tmp1),dbl_mb(k_tmp2),
|
||||
$ dbl_mb(k_tmp3),dbl_mb(vkv),dbl_mb(k_tmp4),myld+myld2,
|
||||
$ ndim,mdim,.false.)
|
||||
|
||||
if(pars%ipol.eq.1) then
|
||||
call dcopy(mdim*myld,dbl_mb(k_tmp1),1,dbl_mb(k_w),1)
|
||||
call dcopy(mdim*myld,dbl_mb(k_tmp2),1,dbl_mb(k_kw),1)
|
||||
call dcopy(mdim*myld,dbl_mb(k_tmp3),1,dbl_mb(k_mkw),1)
|
||||
else
|
||||
do itrial=1,mdim
|
||||
call dcopy(myld,dbl_mb(k_tmp1+(itrial-1)*(myld+myld2)),1,
|
||||
$ dbl_mb(k_w+(itrial-1)*myld),1)
|
||||
call dcopy(myld2,
|
||||
$ dbl_mb(k_tmp1+(itrial-1)*(myld+myld2)+myld),1,
|
||||
$ dbl_mb(k_w2+(itrial-1)*myld2),1)
|
||||
call dcopy(myld,dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)),1,
|
||||
$ dbl_mb(k_kw+(itrial-1)*myld),1)
|
||||
call dcopy(myld2,
|
||||
$ dbl_mb(k_tmp2+(itrial-1)*(myld+myld2)+myld),1,
|
||||
$ dbl_mb(k_kw2+(itrial-1)*myld2),1)
|
||||
call dcopy(myld,dbl_mb(k_tmp3+(itrial-1)*(myld+myld2)),1,
|
||||
$ dbl_mb(k_mkw+(itrial-1)*myld),1)
|
||||
call dcopy(myld2,
|
||||
$ dbl_mb(k_tmp3+(itrial-1)*(myld+myld2)+myld),1,
|
||||
$ dbl_mb(k_mkw2+(itrial-1)*myld2),1)
|
||||
enddo
|
||||
call nga_release_update_block_grid(g_w,subscript)
|
||||
call nga_release_update_block_grid(g_kw,subscript)
|
||||
call nga_release_update_block_grid(g_mkw,subscript)
|
||||
endif
|
||||
call nga_release_update_block_grid(g_w,(/pars%me,0/))
|
||||
call nga_release_update_block_grid(g_kw,(/pars%me,0/))
|
||||
call nga_release_update_block_grid(g_mkw,(/pars%me,0/))
|
||||
|
||||
! All trial vectors were linearly-dependent
|
||||
! restart calculation with current best guess
|
||||
if (mdim.lt.3) then
|
||||
nrestart = min(ntrials, pars%nroots*pars%nspace)
|
||||
call bse_davidson_restart(ntrials,nrestart,npoles,
|
||||
$ g_x,g_y,g_ym,g_s,g_ks,g_mks,g_trials,l_vkv,vkv,
|
||||
$ int_mb(k_map),pars%nblocks)
|
||||
goto 1
|
||||
endif
|
||||
|
||||
ndim = mdim
|
||||
|
||||
! Extend the size of the current vectors
|
||||
call bse_davidson_newvecs(ntrials,ndim,npoles,g_x,g_y,g_ym,
|
||||
$ g_s,g_ks,g_mks,g_w,g_kw,g_mkw,
|
||||
$ l_vkv,vkv,int_mb(k_map),pars%nblocks)
|
||||
goto 2
|
||||
|
||||
3 continue
|
||||
|
||||
if(pars%me.eq.0) write(*,*)
|
||||
|
||||
stat = ga_destroy(pars%g_apb) .and. ga_destroy(pars%g_amb)
|
||||
if(.not.stat) call errquit('deallocation failed',481,GA_ERR)
|
||||
|
||||
stat = ga_create(mt_dbl,npoles,pars%nroots,'X',0,pars%nroots,
|
||||
$ pars%g_apb)
|
||||
stat = ga_create(mt_dbl,npoles,pars%nroots,'Y',0,pars%nroots,
|
||||
$ pars%g_amb)
|
||||
call ga_zero(pars%g_apb)
|
||||
call ga_zero(pars%g_amb)
|
||||
if(.not.stat) call errquit('failed to creare GA',484,GA_ERR)
|
||||
call nga_copy_patch('n',g_x,(/1,1/),(/npoles,pars%nroots/),
|
||||
$ pars%g_apb,(/1,1/),(/npoles,pars%nroots/))
|
||||
call nga_copy_patch('n',g_y,(/1,1/),(/npoles,pars%nroots/),
|
||||
$ pars%g_amb,(/1,1/),(/npoles,pars%nroots/))
|
||||
|
||||
stat = ga_destroy(g_x).and.ga_destroy(g_y).and.ga_destroy(g_ym)
|
||||
stat = stat.and.ga_destroy(g_s).and.ga_destroy(g_ks)
|
||||
stat = stat.and.ga_destroy(g_mks).and.ga_destroy(g_w)
|
||||
stat = stat.and.ga_destroy(g_kw).and.ga_destroy(g_mkw)
|
||||
stat = stat.and.ga_destroy(g_trials).and.ga_destroy(g_wia)
|
||||
do isp=1,pars%ipol
|
||||
stat = stat.and.ga_destroy(pars%g_erioo(isp))
|
||||
stat = stat.and.ga_destroy(pars%g_eriov(isp))
|
||||
stat = stat.and.ga_destroy(pars%g_erivv(isp))
|
||||
stat = stat.and.ga_destroy(pars%g_wov(isp))
|
||||
enddo
|
||||
if(.not.stat) call errquit('deallocation failed',496,GA_ERR)
|
||||
|
||||
if (.not.pars%tda) then
|
||||
do ipole=0,pars%nroots-1
|
||||
dbl_mb(k_omega+ipole) = dsqrt(dbl_mb(k_omega+ipole))
|
||||
call ga_scale_patch(pars%g_apb,1,npoles,ipole+1,ipole+1,
|
||||
$ dsqrt(dbl_mb(k_omega+ipole)))
|
||||
call ga_scale_patch(pars%g_amb,1,npoles,ipole+1,ipole+1,
|
||||
$ 1.0/dsqrt(dbl_mb(k_omega+ipole)))
|
||||
enddo
|
||||
endif
|
||||
|
||||
call ga_add(0.5d0,pars%g_apb,0.5d0,pars%g_amb,pars%g_apb)
|
||||
call ga_add(1d0,pars%g_apb,-1d0,pars%g_amb,pars%g_amb)
|
||||
call bse_oscstr(pars,dbl_mb(k_omega),npoles)
|
||||
|
||||
stat = ma_chop_stack(l_dia)
|
||||
if(.not.stat) call errquit('deallocation failed',501,MA_ERR)
|
||||
|
||||
! Destroy GAs
|
||||
if (.not.ga_destroy(pars%g_apb))
|
||||
& call errquit(pname//'could not destroy temp GA',519,GA_ERR)
|
||||
if (.not.ga_destroy(pars%g_amb))
|
||||
& call errquit(pname//'could not destroy temp GA',519,GA_ERR)
|
||||
|
||||
end subroutine bse_davidson
|
||||
|
||||
|
||||
|
||||
|
||||
98
src/bsemol/bse_davidson_diagonal.F
Normal file
98
src/bsemol/bse_davidson_diagonal.F
Normal file
|
|
@ -0,0 +1,98 @@
|
|||
subroutine bse_davidson_diagonal(pars,wia,maxpoles)
|
||||
implicit none
|
||||
#include "mafdecls.fh"
|
||||
#include "errquit.fh"
|
||||
#include "global.fh"
|
||||
#include "bse.fh"
|
||||
#include "util.fh"
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
|
||||
integer maxpoles, npoles
|
||||
double precision wia(maxpoles,pars%ipol)
|
||||
|
||||
character(*),parameter :: pname = 'bse_casida: '
|
||||
|
||||
integer isp,jsp,nri,off1,off2
|
||||
integer ilo,ihi,jlo,jhi
|
||||
integer nocc2(2),nvir2(2)
|
||||
integer ipole,imo,jmo,amo,bmo
|
||||
integer k,l,kproc,lproc,klocal,llocal,kglobal,lglobal
|
||||
integer g_diag, g_tmp, g_amb, ktmp, ld, info, i, j
|
||||
integer ij,jb,llia,ulia,ltmp2,ltmp3,ktmp2,ktmp3
|
||||
double precision temp(maxpoles),factor
|
||||
|
||||
double precision, external :: ddot
|
||||
|
||||
#ifdef USE_OPENMP
|
||||
call util_blas_set_num_threads(pars%iMaxthreads)
|
||||
#endif
|
||||
|
||||
nri = pars%nri
|
||||
npoles = pars%npoles(1)
|
||||
if (pars%ipol.gt.1) npoles = npoles + pars%npoles(2)
|
||||
if (.not.nga_create(mt_dbl,1,npoles,'apb',0,pars%g_apb))
|
||||
$ call errquit(pname//'could not create wia',0,GA_ERR)
|
||||
if (.not.nga_create(mt_dbl,1,npoles,'amb',0,pars%g_amb))
|
||||
$ call errquit(pname//'could not create wia',0,GA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,nri,'dia',ltmp2,ktmp2))
|
||||
& call errquit(pname//'failed to allocate Delta_ia',0,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,nri,'dia',ltmp3,ktmp3))
|
||||
& call errquit(pname//'failed to allocate Delta_ia',0,MA_ERR)
|
||||
|
||||
factor = 2d0
|
||||
if ((pars%singlet) .and. (pars%ipol.eq.1)) factor = 4d0
|
||||
if ((pars%triplet) .and. (pars%ipol.eq.1)) factor = 0d0
|
||||
if (pars%tda) factor = factor/2d0
|
||||
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
|
||||
! Eigenvalue differences
|
||||
call ga_distribution(pars%g_eriov(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
call dcopy(jhi-jlo+1,wia(1,isp),1,temp,1)
|
||||
|
||||
! Screened Coulomb from A
|
||||
do ipole=jlo,jhi
|
||||
imo = (ipole-1)/pars%nvir(isp) + 1
|
||||
amo = ipole - (imo-1)*pars%nvir(isp)
|
||||
imo = imo + (imo-1)*pars%nocc(isp)
|
||||
amo = amo + (amo-1)*pars%nvir(isp)
|
||||
call ga_get(pars%g_erioo(isp),1,nri,imo,imo,dbl_mb(ktmp2),nri)
|
||||
call ga_get(pars%g_erivv(isp),1,nri,amo,amo,dbl_mb(ktmp3),nri)
|
||||
temp(ipole-jlo+1) = temp(ipole-jlo+1) -
|
||||
$ ddot(nri,dbl_mb(ktmp2),1,dbl_mb(ktmp3),1)
|
||||
enddo
|
||||
call nga_put(pars%g_apb,jlo+off1,jhi+off1,temp,jhi-jlo+1)
|
||||
call nga_put(pars%g_amb,jlo+off1,jhi+off1,temp,jhi-jlo+1)
|
||||
|
||||
! Screened Coulomb from B
|
||||
if (.not.pars%tda) then
|
||||
call ga_access(pars%g_wov(isp),ilo,ihi,jlo,jhi,ktmp,ld)
|
||||
do ipole=jlo,jhi
|
||||
temp(ipole-jlo+1) = ddot(nri,dbl_mb(ktmp+(ipole-jlo)*nri),1,
|
||||
$ dbl_mb(ktmp+(ipole-jlo)*nri),1)
|
||||
enddo
|
||||
call nga_acc(pars%g_apb,jlo+off1,jhi+off1,temp,jhi-jlo+1,-1d0)
|
||||
call nga_acc(pars%g_amb,jlo+off1,jhi+off1,temp,jhi-jlo+1,1d0)
|
||||
call ga_release(pars%g_wov(isp),ilo,ihi,jlo,jhi)
|
||||
endif
|
||||
|
||||
! Hartree term
|
||||
call ga_access(pars%g_eriov(isp),ilo,ihi,jlo,jhi,ktmp,ld)
|
||||
do ipole=jlo,jhi
|
||||
temp(ipole-jlo+1) =
|
||||
$ factor*ddot(nri,dbl_mb(ktmp+(ipole-jlo)*nri),1,
|
||||
$ dbl_mb(ktmp+(ipole-jlo)*nri),1)
|
||||
enddo
|
||||
call nga_acc(pars%g_apb,jlo+off1,jhi+off1,temp,jhi-jlo+1,1d0)
|
||||
call ga_release(pars%g_eriov(isp),ilo,ihi,jlo,jhi)
|
||||
|
||||
enddo
|
||||
|
||||
if(.not.ma_chop_stack(ltmp2))
|
||||
$ call errquit('bse_davidson_diagonal: chop stack',0,MA_ERR)
|
||||
|
||||
call ga_sync()
|
||||
|
||||
end subroutine
|
||||
141
src/bsemol/bse_davidson_guess.F
Normal file
141
src/bsemol/bse_davidson_guess.F
Normal file
|
|
@ -0,0 +1,141 @@
|
|||
subroutine bse_davidson_guess(pars,diagonal,g_trials,npoles,
|
||||
$ ntrials,nroots)
|
||||
|
||||
implicit none
|
||||
|
||||
#include "bse.fh"
|
||||
#include "stdio.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "cdft.fh"
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
integer npoles,ntrials,nroots,g_trials
|
||||
double precision diagonal(npoles)
|
||||
double precision amb(npoles)
|
||||
|
||||
character*4 oname,vname
|
||||
logical found
|
||||
integer amo,imo,itrial,ipole,jpole,indeces(ntrials)
|
||||
integer isym,jsym,i,j,pol,idum
|
||||
integer allindeces(npoles)
|
||||
double precision minimum
|
||||
logical swapped
|
||||
|
||||
call nga_get(pars%g_apb,1,npoles,diagonal,npoles)
|
||||
call nga_get(pars%g_amb,1,npoles,amb,npoles)
|
||||
|
||||
if (.not.pars%tda) then
|
||||
diagonal(:) = diagonal(:)*amb(:)
|
||||
endif
|
||||
indeces(:) = -1
|
||||
allindeces(:) = -1
|
||||
|
||||
do ipole=1,npoles
|
||||
allindeces(ipole) = ipole
|
||||
enddo
|
||||
|
||||
do ipole=npoles-1,1,-1
|
||||
swapped = .false.
|
||||
do jpole=1,ipole
|
||||
if (diagonal(allindeces(jpole)).gt.
|
||||
$ diagonal(allindeces(jpole+1))) then
|
||||
idum = allindeces(jpole)
|
||||
allindeces(jpole) = allindeces(jpole+1)
|
||||
allindeces(jpole+1) = idum
|
||||
swapped = .true.
|
||||
endif
|
||||
enddo
|
||||
if(.not.swapped) exit
|
||||
enddo
|
||||
|
||||
if (pars%dowindow) then
|
||||
do ipole=1,npoles
|
||||
if (dsqrt(diagonal(allindeces(ipole))).ge.pars%elower) then
|
||||
if (ipole+ntrials-1.gt.npoles) then
|
||||
indeces(1:ntrials) = allindeces(npoles-ntrials+1:npoles)
|
||||
else
|
||||
indeces(1:ntrials) = allindeces(ipole:ipole+ntrials-1)
|
||||
endif
|
||||
exit
|
||||
endif
|
||||
enddo
|
||||
else
|
||||
indeces(1:ntrials) = allindeces(1:ntrials)
|
||||
endif
|
||||
|
||||
! Get the indeces of the smallest diagonal elements
|
||||
! within energy window
|
||||
#if 0
|
||||
do itrial=1,ntrials
|
||||
minimum = 1d6
|
||||
do ipole=1,npoles
|
||||
if (abs(diagonal(ipole)).lt.minimum) then
|
||||
if (pars%dowindow) then
|
||||
else
|
||||
found = .true.
|
||||
endif
|
||||
do jpole=1,itrial
|
||||
if (ipole.eq.indeces(jpole)) then
|
||||
found = .false.
|
||||
exit
|
||||
endif
|
||||
enddo
|
||||
if (found) then
|
||||
indeces(itrial) = ipole
|
||||
minimum = abs(diagonal(ipole))
|
||||
endif
|
||||
endif
|
||||
enddo
|
||||
enddo
|
||||
#endif
|
||||
|
||||
! Get symmetry of ground state
|
||||
isym = 1
|
||||
if (pars%ipol.eq.2) then
|
||||
do i=1,pars%ipol
|
||||
do j=1,pars%nocc(i)
|
||||
isym = ieor(isym-1,int_mb(pars%k_irs(i)+j-1)-1)+1
|
||||
enddo
|
||||
enddo
|
||||
endif
|
||||
call sym_irrepname(geom, isym, oname)
|
||||
|
||||
|
||||
! Start the guess vector
|
||||
write(LuOut,*)
|
||||
write(LuOut,9000) nroots
|
||||
write(LuOut,9020)
|
||||
write(LuOut,9010)
|
||||
write(LuOut,9020)
|
||||
do itrial=1,ntrials
|
||||
ipole = indeces(itrial)
|
||||
call ga_put(g_trials,ipole,ipole,itrial,itrial,1d0,1)
|
||||
if(itrial.gt.nroots) cycle
|
||||
|
||||
! Print information
|
||||
if (ipole.le.pars%npoles(1)) then
|
||||
imo = (ipole-1)/pars%nvir(1) + 1
|
||||
amo = ipole - (imo-1)*pars%nvir(1) + pars%nocc(1)
|
||||
jsym = ieor(isym-1,int_mb(pars%k_irs(1)+imo-1)-1)+1
|
||||
jsym = ieor(jsym-1,int_mb(pars%k_irs(1)+amo-1)-1)+1
|
||||
pol = 1
|
||||
else
|
||||
imo = (ipole - pars%npoles(1) - 1)/pars%nvir(2) + 1
|
||||
amo = ipole - pars%npoles(1) - (imo-1)*pars%nvir(2) +
|
||||
$ pars%nocc(2)
|
||||
jsym = ieor(isym-1,int_mb(pars%k_irs(2)+imo-1)-1)+1
|
||||
jsym = ieor(jsym-1,int_mb(pars%k_irs(2)+amo-1)-1)+1
|
||||
pol = 2
|
||||
endif
|
||||
call sym_irrepname(geom, jsym, vname)
|
||||
write(luout,9030) itrial,pol,imo,amo,vname,
|
||||
$ dsqrt(diagonal(ipole))*27.211396d0
|
||||
enddo
|
||||
|
||||
9000 format(i5,' smallest diagonals (eV) ')
|
||||
9010 format(' No. Spin Occ Vir Irrep Diag')
|
||||
9020 format('------------------------------------')
|
||||
9030 format(4i5,3x,a4,1x,f8.3)
|
||||
|
||||
end subroutine
|
||||
|
||||
74
src/bsemol/bse_davidson_kortho.F
Normal file
74
src/bsemol/bse_davidson_kortho.F
Normal file
|
|
@ -0,0 +1,74 @@
|
|||
subroutine bse_davidson_kortho(pars,v,kv,mkv,vkv,work,npoles,n,m,
|
||||
$ first)
|
||||
|
||||
implicit none
|
||||
#include "bse.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "errquit.fh"
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
logical first
|
||||
integer npoles,n,m
|
||||
|
||||
integer l_work,k_work,info,i,j,k
|
||||
|
||||
double precision rdum
|
||||
double precision v(npoles,n),kv(npoles,n),mkv(npoles,n)
|
||||
double precision vkv(n,n),eig(n),work(npoles,n)
|
||||
double precision sqrtvkv(n,n)
|
||||
|
||||
! V**T K V
|
||||
call dgemm('t','n',n,n,npoles,1d0,v,npoles,kv,npoles,0d0,vkv,n)
|
||||
call ga_dgop((/14/),vkv,n*n,'+')
|
||||
|
||||
! Allocate work field
|
||||
call dsyev('v','l',n,vkv,n,eig,rdum,-1,info)
|
||||
if(.not.ma_push_get(mt_dbl,int(rdum),'work',l_work,k_work))
|
||||
$ call errquit('bse_kortho: failed to allocate work',199,MA_ERR)
|
||||
|
||||
! Diagonalize matrix
|
||||
call dsyev('v','l',n,vkv,n,eig,dbl_mb(k_work),int(rdum),info)
|
||||
if (info.ne.0)
|
||||
$ call errquit("bse_davidson_kortho: diago failed",0,0)
|
||||
|
||||
! Deallocate field
|
||||
if (.not.ma_chop_stack(l_work))
|
||||
$ call errquit('bse_kortho: failed to deallocate work',199,MA_ERR)
|
||||
|
||||
! Remove small eigenpairs
|
||||
m = n
|
||||
k = 1
|
||||
100 continue
|
||||
do i=k,m
|
||||
if (eig(i).lt.1d-9) then
|
||||
do j=i+1,n
|
||||
eig(j-1) = eig(j)
|
||||
vkv(:,j-1) = vkv(:,j)
|
||||
enddo
|
||||
m = m - 1
|
||||
k = i
|
||||
goto 100
|
||||
endif
|
||||
enddo
|
||||
|
||||
! Obtain VKV^(-1/2)
|
||||
do i=1,m
|
||||
vkv(:,i) = vkv(:,i)/dsqrt(eig(i))
|
||||
enddo
|
||||
|
||||
if (m.eq.0) return
|
||||
|
||||
! Obtain orthogonalized vectors
|
||||
call dgemm('n','n',npoles,m,n,1d0,v,npoles,vkv,n,0d0,work,npoles)
|
||||
call dcopy(npoles*m,work,1,v,1)
|
||||
|
||||
if (first) return
|
||||
|
||||
call dgemm('n','n',npoles,m,n,1d0,kv,npoles,vkv,n,0d0,work,npoles)
|
||||
call dcopy(npoles*m,work,1,kv,1)
|
||||
|
||||
call dgemm('n','n',npoles,m,n,1d0,mkv,npoles,vkv,n,0d0,work,
|
||||
$ npoles)
|
||||
call dcopy(npoles*m,work,1,mkv,1)
|
||||
|
||||
end subroutine
|
||||
159
src/bsemol/bse_davidson_kvec.F
Normal file
159
src/bsemol/bse_davidson_kvec.F
Normal file
|
|
@ -0,0 +1,159 @@
|
|||
subroutine bse_davidson_kvec(pars,vec,wia,kv,npoles,ntrials,
|
||||
$ dohartree,sgn)
|
||||
implicit none
|
||||
#include "bse.fh"
|
||||
#include "global.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "errquit.fh"
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
|
||||
integer npoles,ntrials
|
||||
integer vec,wia,kv,mv
|
||||
double precision sgn
|
||||
|
||||
logical dohartree
|
||||
|
||||
integer ipol,npoles1,npoles2,uli,lli,ipole,off1
|
||||
integer nocc(2),nvir(2),amo,imo,itrial,ula,lla,nri
|
||||
integer ilo,ihi,jlo,jhi,ld,ld2, myld
|
||||
|
||||
integer l_tmp1,k_tmp1,k_eriov,isp,k_vec,k_erivv
|
||||
integer g_tmp1, g_tmp2, g_tmp3, g_tmp4, k_tmp
|
||||
integer subscript(2), k_tmp3
|
||||
double precision temp(npoles),factor
|
||||
|
||||
ipol = pars%ipol
|
||||
nri = pars%nri
|
||||
nocc = pars%nocc
|
||||
nvir = pars%nvir
|
||||
npoles1 = nocc(1)*nvir(1)
|
||||
npoles2 = nocc(2)*nvir(2)
|
||||
|
||||
call ga_zero(kv)
|
||||
call ga_copy(vec,kv)
|
||||
|
||||
if (pars%tda .and. (sgn.eq.-1d0)) return
|
||||
|
||||
call ga_scale_rows(kv,wia)
|
||||
|
||||
! Screened Coulomb A
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
subscript = (/pars%me+(isp-1)*ga_nnodes(),0/)
|
||||
|
||||
if(.not.ga_create(mt_dbl,pars%nvir(isp),pars%nocc(isp),'g_tmp1',
|
||||
$ 0,pars%nocc(isp),g_tmp1))
|
||||
$ call errquit('bse_kvec: ga_create failed',42,GA_ERR)
|
||||
call ga_zero(g_tmp1)
|
||||
if(.not.ga_create(mt_dbl,nri,pars%nvir(isp),'g_tmp2',
|
||||
$ nri,0,g_tmp2))
|
||||
$ call errquit('bse_kvec: ga_create failed',45,GA_ERR)
|
||||
call ga_zero(g_tmp2)
|
||||
if(.not.ga_create(mt_dbl,nri*pars%nvir(isp),1,'g_tmp3',
|
||||
$ 0,1,g_tmp3))
|
||||
$ call errquit('bse_kvec: ga_create failed',48,GA_ERR)
|
||||
call ga_zero(g_tmp3)
|
||||
if(.not.ga_create(mt_dbl,pars%npoles(isp),1,'g_tmp4',
|
||||
$ 0,1,g_tmp4))
|
||||
$ call errquit('bse_kvec: ga_create failed',48,GA_ERR)
|
||||
call ga_zero(g_tmp4)
|
||||
|
||||
call nga_access_block_grid(vec,subscript,k_vec,ld2)
|
||||
call ga_distribution(pars%g_erivv(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
call ga_access(pars%g_erivv(isp),ilo,ihi,jlo,jhi,k_erivv,ld)
|
||||
call ga_access(g_tmp3,ilo,ihi,1,1,k_tmp3,ld)
|
||||
|
||||
do itrial=1,ntrials
|
||||
call nga_copy_patch('n',vec,(/off1+1,itrial/),
|
||||
$ (/off1+pars%npoles(isp),itrial/),g_tmp4,
|
||||
$ (/1,1/),(/pars%npoles(isp),1/))
|
||||
call ga_copy_patch('n',g_tmp4,1,pars%npoles(isp),1,1,
|
||||
$ g_tmp1,1,pars%nvir(isp),1,pars%nocc(isp))
|
||||
do imo=1,pars%nocc(isp)
|
||||
lli = (imo-1)*pars%nocc(isp) + 1
|
||||
uli = imo*pars%nocc(isp)
|
||||
call ga_matmul_patch('n','t',1d0,0d0,
|
||||
$ pars%g_erioo(isp),1,nri,lli,uli,
|
||||
$ g_tmp1,1,pars%nocc(isp),1,pars%nvir(isp),
|
||||
$ g_tmp2,1,nri,1,pars%nvir(isp))
|
||||
call ga_copy_patch('n',g_tmp2,1,nri,1,pars%nvir(isp),
|
||||
$ g_tmp3,1,nri*pars%nvir(isp),1,1)
|
||||
call dgemv('t',ld,pars%nvir(isp),-1d0,dbl_mb(k_erivv),ld,
|
||||
$ dbl_mb(k_tmp3),1,0d0,temp((imo-1)*pars%nvir(isp)+1),1)
|
||||
enddo
|
||||
call nga_acc(kv,(/1+off1,itrial/),
|
||||
$ (/pars%npoles(isp)+off1,itrial/),temp,npoles,1d0)
|
||||
enddo
|
||||
if(.not. (ga_destroy(g_tmp1).and.ga_destroy(g_tmp2).and.
|
||||
$ ga_destroy(g_tmp3).and.ga_destroy(g_tmp4)))
|
||||
$ call errquit('bse_davidson_kvec: ga_destroy_failed',68,GA_ERR)
|
||||
|
||||
call ga_release(pars%g_erivv(isp),ilo,ihi,jlo,jhi)
|
||||
call ga_release_update(g_tmp3,ilo,ihi,1,1)
|
||||
call nga_release_block_grid(vec,subscript)
|
||||
enddo
|
||||
call ga_sync()
|
||||
|
||||
if (.not.ma_push_get(mt_dbl,nri*ntrials,'tmp1',l_tmp1,k_tmp1))
|
||||
$ call errquit('bse_davidson_kvec: MA push failed',73,MA_ERR)
|
||||
|
||||
! Hartree
|
||||
if (dohartree) then
|
||||
factor = 4d0/pars%ipol
|
||||
if (pars%tda) factor = factor/2d0
|
||||
call dcopy(nri*ntrials,0d0,0,dbl_mb(k_tmp1),1)
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
subscript = (/pars%me+(isp-1)*ga_nnodes(),0/)
|
||||
call ga_distribution(pars%g_eriov(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
call ga_access(pars%g_eriov(isp),ilo,ihi,jlo,jhi,k_eriov,ld)
|
||||
call nga_access_block_grid(vec,subscript,k_vec,ld)
|
||||
call dgemm('n','n',nri,ntrials,ld,1d0,dbl_mb(k_eriov),nri,
|
||||
$ dbl_mb(k_vec),ld,1d0,dbl_mb(k_tmp1),nri)
|
||||
call ga_release(pars%g_eriov(isp),ilo,ihi,jlo,jhi)
|
||||
call nga_release_block_grid(vec,subscript)
|
||||
enddo
|
||||
call ga_dgop((/93/),dbl_mb(k_tmp1),nri*ntrials,'+')
|
||||
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
subscript = (/pars%me+(isp-1)*ga_nnodes(),0/)
|
||||
call ga_distribution(pars%g_eriov(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
call ga_access(pars%g_eriov(isp),ilo,ihi,jlo,jhi,k_eriov,ld)
|
||||
call nga_access_block_grid(kv,subscript,k_vec,ld)
|
||||
call dgemm('t','n',ld,ntrials,nri,factor,dbl_mb(k_eriov),nri,
|
||||
$ dbl_mb(k_tmp1),nri,1d0,dbl_mb(k_vec),ld)
|
||||
call ga_release(pars%g_eriov(isp),ilo,ihi,jlo,jhi)
|
||||
call nga_release_update_block_grid(kv,subscript)
|
||||
enddo
|
||||
endif
|
||||
|
||||
! Screened Coulomb B
|
||||
if(.not.pars%tda) then
|
||||
|
||||
do isp=1,pars%ipol
|
||||
off1 = (isp-1)*pars%npoles(1)
|
||||
subscript = (/pars%me+(isp-1)*ga_nnodes(),0/)
|
||||
call ga_distribution(pars%g_wov(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
call ga_access(pars%g_wov(isp),ilo,ihi,jlo,jhi,k_eriov,ld)
|
||||
call nga_access_block_grid(vec,subscript,k_vec,ld)
|
||||
call dgemm('n','n',nri,ntrials,ld,-1d0,dbl_mb(k_eriov),nri,
|
||||
$ dbl_mb(k_vec),ld,0d0,dbl_mb(k_tmp1),nri)
|
||||
call nga_release_block_grid(vec,subscript)
|
||||
|
||||
call ga_dgop((/115/),dbl_mb(k_tmp1),nri*ntrials,'+')
|
||||
|
||||
call nga_access_block_grid(kv,subscript,k_vec,ld)
|
||||
call dgemm('t','n',ld,ntrials,nri,sgn,dbl_mb(k_eriov),nri,
|
||||
$ dbl_mb(k_tmp1),nri,1d0,dbl_mb(k_vec),ld)
|
||||
call ga_release(pars%g_wov(isp),ilo,ihi,jlo,jhi)
|
||||
call nga_release_update_block_grid(kv,subscript)
|
||||
enddo
|
||||
|
||||
endif
|
||||
|
||||
if (.not.ma_chop_stack(l_tmp1))
|
||||
$ call errquit('bse_kvec: failed chop stack',122,MA_ERR)
|
||||
|
||||
end
|
||||
79
src/bsemol/bse_davidson_newvecs.F
Normal file
79
src/bsemol/bse_davidson_newvecs.F
Normal file
|
|
@ -0,0 +1,79 @@
|
|||
subroutine bse_davidson_newvecs(ntrials,ndim,npoles,g_x,g_y,
|
||||
$ g_ym,g_s,g_ks,g_mks,g_w,g_kw,g_mkw,lvkv,vkv,map,
|
||||
$ nblocks)
|
||||
|
||||
implicit none
|
||||
#include "global.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "errquit.fh"
|
||||
|
||||
integer ntrials,ndim,npoles,nblocks
|
||||
integer g_x,g_y,g_ym,g_s,g_ks,g_mks
|
||||
integer g_w,g_kw,g_mkw,lvkv,vkv
|
||||
integer map(nblocks),grid(2)
|
||||
|
||||
logical stat
|
||||
|
||||
grid = (/ga_nnodes(),1/)
|
||||
|
||||
! Save working vectors
|
||||
call ga_copy(g_s,g_x)
|
||||
call ga_copy(g_ks,g_y)
|
||||
call ga_copy(g_mks,g_ym)
|
||||
|
||||
! Destroy working vectors
|
||||
stat = ga_destroy(g_s).and.ga_destroy(g_ks).and.ga_destroy(g_mks)
|
||||
if(.not.stat)
|
||||
$ call errquit('bse_davidson_newvecs: destroy failed',1,GA_ERR)
|
||||
|
||||
! Create new working vectors
|
||||
g_s = ga_create_handle()
|
||||
call ga_set_data(g_s,2,(/npoles,ntrials+ndim/),mt_dbl)
|
||||
call ga_set_array_name(g_s,'s')
|
||||
call ga_set_tiled_irreg_proc_grid(g_s,map,(/nblocks,1/),grid)
|
||||
stat = ga_allocate(g_s)
|
||||
|
||||
stat = stat.and.ga_duplicate(g_s,g_ks,'ks').and.
|
||||
$ ga_duplicate(g_s,g_mks,'mks')
|
||||
if(.not.stat)
|
||||
$ call errquit('bse_davidson_newvecs: create failed',1,GA_ERR)
|
||||
|
||||
|
||||
! Copy back old working vectors plus new vectors
|
||||
call ga_copy_patch('n',g_x,1,npoles,1,ntrials,
|
||||
$ g_s,1,npoles,1,ntrials)
|
||||
call ga_copy_patch('n',g_w,1,npoles,1,ndim,
|
||||
$ g_s,1,npoles,ntrials+1,ntrials+ndim)
|
||||
|
||||
call ga_copy_patch('n',g_y,1,npoles,1,ntrials,
|
||||
$ g_ks,1,npoles,1,ntrials)
|
||||
call ga_copy_patch('n',g_kw,1,npoles,1,ndim,
|
||||
$ g_ks,1,npoles,ntrials+1,ntrials+ndim)
|
||||
|
||||
call ga_copy_patch('n',g_ym,1,npoles,1,ntrials,
|
||||
$ g_mks,1,npoles,1,ntrials)
|
||||
call ga_copy_patch('n',g_mkw,1,npoles,1,ndim,
|
||||
$ g_mks,1,npoles,ntrials+1,ntrials+ndim)
|
||||
|
||||
! Update ntrials
|
||||
ntrials = ntrials + ndim
|
||||
|
||||
! Destroy old solutions
|
||||
stat = ga_destroy(g_x).and.ga_destroy(g_y).and.ga_destroy(g_ym)
|
||||
if(.not.stat)
|
||||
$ call errquit('bse_davidson_newvecs: destroy failed',2,GA_ERR)
|
||||
|
||||
! Create new solutions
|
||||
stat = ga_duplicate(g_s,g_x,'x').and.
|
||||
$ ga_duplicate(g_s,g_y,'y').and.
|
||||
$ ga_duplicate(g_s,g_ym,'ym')
|
||||
if(.not.stat)
|
||||
$ call errquit('bse_davidson_newvecs: create failed',2,GA_ERR)
|
||||
|
||||
! Reallocate VKV
|
||||
if(.not.ma_chop_stack(lvkv))
|
||||
$ call errquit('bse_davidson_restart: alloc failed',3,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,ntrials*ntrials,'vkv',lvkv,vkv))
|
||||
$ call errquit('bse_davidson_restart: dealloc failed',3,MA_ERR)
|
||||
|
||||
end subroutine
|
||||
32
src/bsemol/bse_davidson_residual.F
Normal file
32
src/bsemol/bse_davidson_residual.F
Normal file
|
|
@ -0,0 +1,32 @@
|
|||
subroutine bse_davidson_residual(nroots,ntrials,omega,x,ym,r,dia,
|
||||
$ w,npoles,ndim,norm)
|
||||
implicit none
|
||||
|
||||
logical main
|
||||
integer nroots,npoles,ntrials,ndim
|
||||
double precision omega(nroots)
|
||||
double precision ym(npoles,nroots),r(npoles)
|
||||
double precision x(npoles,nroots),dia(npoles),w(npoles,nroots)
|
||||
|
||||
logical consecutive
|
||||
integer iroot
|
||||
double precision norm(nroots)
|
||||
|
||||
consecutive = .true.
|
||||
ndim = 0
|
||||
do iroot=1,nroots
|
||||
r(:) = ym(:,iroot) - omega(iroot)*x(:,iroot)
|
||||
norm(iroot) = maxval(abs(r))
|
||||
call ga_dgop((/1/),norm(iroot),1,"maxabs")
|
||||
if (consecutive .and. norm(iroot).gt.1d-4) consecutive = .false.
|
||||
if ((norm(iroot).gt.1d-4) .or. (.not.consecutive)) then
|
||||
ndim = ndim + 1
|
||||
w(:,ndim) = r(:)/(dia(:) - omega(iroot))
|
||||
endif
|
||||
enddo
|
||||
|
||||
end subroutine
|
||||
|
||||
|
||||
|
||||
|
||||
78
src/bsemol/bse_davidson_restart.F
Normal file
78
src/bsemol/bse_davidson_restart.F
Normal file
|
|
@ -0,0 +1,78 @@
|
|||
subroutine bse_davidson_restart(ntrials,nroots,npoles,g_x,g_y,
|
||||
$ g_ym,g_s,g_ks,g_mks,g_trials,
|
||||
$ lvkv,vkv,map,nblocks)
|
||||
|
||||
implicit none
|
||||
#include "global.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "errquit.fh"
|
||||
|
||||
integer ntrials,nroots,npoles,nblocks
|
||||
integer g_x,g_y,g_ym,g_s,g_ks,g_mks,lvkv,vkv
|
||||
integer map(nblocks),grid(2),g_trials
|
||||
|
||||
logical stat
|
||||
|
||||
grid = (/ga_nnodes(),1/)
|
||||
|
||||
! Reset the number of trials
|
||||
ntrials = nroots
|
||||
|
||||
! Copy current solutions
|
||||
call ga_copy_patch('n',g_x,1,npoles,1,ntrials,
|
||||
$ g_trials,1,npoles,1,ntrials)
|
||||
! call ga_copy_patch('n',g_y,1,npoles,1,ntrials,
|
||||
! $ g_ks,1,npoles,1,ntrials)
|
||||
! call ga_copy_patch('n',g_ym,1,npoles,1,ntrials,
|
||||
! $ g_mks,1,npoles,1,ntrials)
|
||||
|
||||
! Destroy current solutions
|
||||
stat = ga_destroy(g_x).and.ga_destroy(g_y).and.ga_destroy(g_ym)
|
||||
if (.not.stat)
|
||||
$ call errquit('bse_davidson_restar: failed destroy',1,GA_ERR)
|
||||
|
||||
! Re-create new solutions
|
||||
g_x = ga_create_handle()
|
||||
call ga_set_data(g_x,2,(/npoles,ntrials/),mt_dbl)
|
||||
call ga_set_array_name(g_x,'x')
|
||||
call ga_set_tiled_irreg_proc_grid(g_x,map,(/nblocks,1/),grid)
|
||||
stat = ga_allocate(g_x)
|
||||
|
||||
stat = stat.and.ga_duplicate(g_x,g_y,'y').and.
|
||||
$ ga_duplicate(g_x,g_ym,'ym')
|
||||
if (.not.stat)
|
||||
$ call errquit('bse_davidson_restar: failed create',1,GA_ERR)
|
||||
|
||||
! Copy back solutions
|
||||
! call ga_copy_patch('n',g_s,1,npoles,1,ntrials,
|
||||
! $ g_x,1,npoles,1,ntrials)
|
||||
! call ga_copy_patch('n',g_ks,1,npoles,1,ntrials,
|
||||
! $ g_y,1,npoles,1,ntrials)
|
||||
! call ga_copy_patch('n',g_mks,1,npoles,1,ntrials,
|
||||
! $ g_ym,1,npoles,1,ntrials)
|
||||
|
||||
! Destroy working vectors
|
||||
stat = ga_destroy(g_s).and.ga_destroy(g_ks).and.ga_destroy(g_mks)
|
||||
if (.not.stat)
|
||||
$ call errquit('bse_davidson_restar: failed destroy',2,GA_ERR)
|
||||
|
||||
! Re-create new solutions
|
||||
stat = ga_duplicate(g_x,g_s,'s').and.
|
||||
$ ga_duplicate(g_x,g_ks,'ks').and.
|
||||
$ ga_duplicate(g_x,g_mks,'mks')
|
||||
|
||||
if (.not.stat)
|
||||
$ call errquit('bse_davidson_restar: failed create',2,GA_ERR)
|
||||
|
||||
! Copy current solutions
|
||||
! call ga_copy(g_x,g_s)
|
||||
! call ga_copy(g_y,g_ks)
|
||||
! call ga_copy(g_ym,g_mks)
|
||||
|
||||
! Reallocate VKV
|
||||
if(.not.ma_chop_stack(lvkv))
|
||||
$ call errquit('bse_davidson_restart: alloc failed',3,MA_ERR)
|
||||
if(.not.ma_push_get(mt_dbl,ntrials*ntrials,'vkv',lvkv,vkv))
|
||||
$ call errquit('bse_davidson_restart: dealloc failed',3,MA_ERR)
|
||||
|
||||
end subroutine
|
||||
33
src/bsemol/bse_defaults.F
Normal file
33
src/bsemol/bse_defaults.F
Normal file
|
|
@ -0,0 +1,33 @@
|
|||
subroutine bse_defaults(rtdb)
|
||||
|
||||
implicit none
|
||||
|
||||
#include "errquit.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "rtdb.fh"
|
||||
|
||||
integer rtdb
|
||||
logical ok
|
||||
|
||||
ok = rtdb_put(rtdb,'bse:analytic',mt_log,1,.true.) .and.
|
||||
& rtdb_put(rtdb,'bse:davidson',mt_log,1,.false.) .and.
|
||||
& rtdb_put(rtdb,'bse:lanczos',mt_log,1,.false.) .and.
|
||||
& rtdb_put(rtdb,'bse:tda',mt_log,1,.false.) .and.
|
||||
& rtdb_put(rtdb,'bse:ncap',mt_log,1,.false.) .and.
|
||||
& rtdb_put(rtdb,'bse:ncapr',mt_log,1,.false.) .and.
|
||||
& rtdb_put(rtdb,'bse:singlet',mt_log,1,.true.) .and.
|
||||
& rtdb_put(rtdb,'bse:triplet',mt_log,1,.false.) .and.
|
||||
& rtdb_put(rtdb,'bse:dowindow',mt_log,1,.false.) .and.
|
||||
& rtdb_put(rtdb,'bse:elower',mt_dbl,1,0d0) .and.
|
||||
& rtdb_put(rtdb,'bse:upper',mt_dbl,1,0d0) .and.
|
||||
& rtdb_put(rtdb,'bse:nroots',mt_int,1,10) .and.
|
||||
& rtdb_put(rtdb,'bse:nmax',mt_int,1,200) .and.
|
||||
& rtdb_put(rtdb,'bse:skipgw',mt_log,1,.false.) .and.
|
||||
& rtdb_put(rtdb,'bse:maxiter',mt_int,1,100) .and.
|
||||
& rtdb_put(rtdb,'bse:nspace',mt_int,1,4) .and.
|
||||
& rtdb_put(rtdb,'bse:init',mt_log,1,.true.)
|
||||
|
||||
if (.not.ok)
|
||||
& call errquit('bse_defaults: could not put to rtdb',0,RTDB_ERR)
|
||||
|
||||
end subroutine bse_defaults
|
||||
38
src/bsemol/bse_finalize.F
Normal file
38
src/bsemol/bse_finalize.F
Normal file
|
|
@ -0,0 +1,38 @@
|
|||
subroutine bse_finalize(pars)
|
||||
implicit none
|
||||
#include "errquit.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "global.fh"
|
||||
#include "geom.fh"
|
||||
#include "bas.fh"
|
||||
#include "cdft.fh"
|
||||
#include "bse.fh"
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
character(*),parameter :: pname = 'bse_finalize: '
|
||||
|
||||
integer i
|
||||
logical ok
|
||||
|
||||
do i=1,ipol
|
||||
if (.not.ga_destroy(pars%g_movecs(i)))
|
||||
$ call errquit(pname//'failed to destroy g_movecs',0,GA_ERR)
|
||||
enddo
|
||||
|
||||
ok = ma_chop_stack(pars%l_mf_evals)
|
||||
if (.not.ok)
|
||||
& call errquit(pname//'could not chop stack',0,MA_ERR)
|
||||
|
||||
call schwarz_tidy()
|
||||
call int_terminate()
|
||||
call fock_2e_tidy(pars%rtdb)
|
||||
|
||||
if (.not.geom_destroy(geom))
|
||||
$ call errquit(pname//'geom_destroy failed',0,GEOM_ERR)
|
||||
if (.not.bas_destroy(ao_bas_han))
|
||||
$ call errquit(pname//'bas_destroy failed',0,BASIS_ERR)
|
||||
if (.not.bas_destroy(cd_bas_han))
|
||||
$ call errquit(pname//'bas_destroy failed',0,BASIS_ERR)
|
||||
|
||||
|
||||
end subroutine
|
||||
15
src/bsemol/bse_get_eia.F
Normal file
15
src/bsemol/bse_get_eia.F
Normal file
|
|
@ -0,0 +1,15 @@
|
|||
subroutine bse_get_eia(eig,eia,nocc,nvir,llpole,ulpole)
|
||||
implicit none
|
||||
integer,intent(in) :: nocc,nvir
|
||||
double precision,intent(in),dimension(nocc+nvir) :: eig
|
||||
double precision,intent(out),dimension(*) :: eia
|
||||
integer ipole,llpole,ulpole,imo,amo
|
||||
c
|
||||
do ipole=llpole,ulpole
|
||||
imo = (ipole-1)/nvir + 1
|
||||
amo = ipole - (imo-1)*nvir + nocc
|
||||
eia(ipole-llpole+1) = eig(amo) - eig(imo)
|
||||
enddo
|
||||
c
|
||||
return
|
||||
end subroutine
|
||||
322
src/bsemol/bse_init.F
Normal file
322
src/bsemol/bse_init.F
Normal file
|
|
@ -0,0 +1,322 @@
|
|||
subroutine bse_init(pars)
|
||||
|
||||
implicit none
|
||||
|
||||
#include "errquit.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "cdft.fh"
|
||||
#include "bas.fh"
|
||||
#include "case.fh"
|
||||
#include "geom.fh"
|
||||
#include "global.fh"
|
||||
#include "bse.fh"
|
||||
#include "util.fh"
|
||||
#include "stdio.fh"
|
||||
#include "rtdb.fh"
|
||||
|
||||
#ifdef SCALAPACK
|
||||
#include "scaleig.fh"
|
||||
#endif
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
character(*), parameter :: pname = 'bse_init: '
|
||||
|
||||
character(len=255) :: basisname, scftype, gwevals
|
||||
integer i, ilo, ihi, j, g_work
|
||||
integer nbf_temp, ipol_temp, nocc, nvir, nmo_temp(2)
|
||||
logical ldum, skipgw
|
||||
|
||||
logical int_normalize, int_norm_2c
|
||||
external int_normalize, int_norm_2c
|
||||
c
|
||||
logical movecs_read, movecs_read_header, atom_tag_check
|
||||
external movecs_read, movecs_read_header, atom_tag_check
|
||||
c
|
||||
integer ga_create_atom_blocked
|
||||
external ga_create_atom_blocked
|
||||
|
||||
integer resultlen,tid,cpu,ierr,iproc,itid
|
||||
character(len=80) names
|
||||
double precision t0, ehomo, vxddp, vxddm, aq2, zeta
|
||||
|
||||
integer,parameter :: unitno = 66
|
||||
integer :: nmix(2)
|
||||
|
||||
logical ncapr
|
||||
|
||||
#ifdef USE_OPENMP
|
||||
integer,external :: omp_get_max_threads, omp_get_thread_num
|
||||
integer,external :: omp_get_num_threads
|
||||
#endif
|
||||
|
||||
|
||||
! Initialize GA variables
|
||||
pars%me = ga_nodeid()
|
||||
pars%nprocs = ga_nnodes()
|
||||
|
||||
! Initialize OMP variables
|
||||
#ifdef USE_OPENMP
|
||||
pars%iMaxThreads = omp_get_max_threads()
|
||||
#else
|
||||
pars%iMaxThreads = 1
|
||||
#endif
|
||||
|
||||
pars%npoles(:) = 0
|
||||
|
||||
! Get defaults in case input block is not present
|
||||
if (.not.rtdb_get(pars%rtdb,'bse:init',mt_log,1,ldum))
|
||||
& call bse_defaults(pars%rtdb)
|
||||
|
||||
! Check method to solve BSE
|
||||
if (.not.rtdb_get(pars%rtdb,'bse:analytic',mt_log,1,
|
||||
& pars%analytic))
|
||||
& call errquit(pname//'could not read from rtdb',1,RTDB_ERR)
|
||||
if (.not.rtdb_get(pars%rtdb,'bse:davidson',mt_log,1,
|
||||
& pars%davidson))
|
||||
& call errquit(pname//'could not read from rtdb',2,RTDB_ERR)
|
||||
if (.not.rtdb_get(pars%rtdb,'bse:lanczos',mt_log,1,
|
||||
& pars%lanczos))
|
||||
& call errquit(pname//'could not read from rtdb',3,RTDB_ERR)
|
||||
|
||||
if ((pars%analytic.and.(pars%davidson.or.pars%lanczos)) .or.
|
||||
& (pars%davidson.and.(pars%analytic.or.pars%lanczos)) .or.
|
||||
& (pars%lanczos.and.(pars%analytic.or.pars%davidson)))
|
||||
& call errquit(pname//'more than one method',4,INPUT_ERR)
|
||||
|
||||
if (.not.rtdb_get(pars%rtdb, 'bse:skipgw', mt_log, 1, skipgw))
|
||||
& call errquit(pname//'could not read skipgw',5, RTDB_ERR)
|
||||
|
||||
! Check wether we are computing singlet or triplet excitations
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:singlet',mt_log,1,pars%singlet))
|
||||
& call errquit(pname//'could not read from rtdb',1,RTDB_ERR)
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:triplet',mt_log,1,pars%triplet))
|
||||
& call errquit(pname//'could not read from rtdb',1,RTDB_ERR)
|
||||
if((pars%singlet .and. pars%triplet).and.(pars%analytic)) then
|
||||
call errquit(pname//'select either triplet or singlet',0,
|
||||
& INPUT_ERR)
|
||||
endif
|
||||
|
||||
! Check if we are doing Tamm-Dancoff approximation
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:tda',mt_log,1,pars%tda))
|
||||
& call errquit(pname//'could not read from rtdb',101,RTDB_ERR)
|
||||
|
||||
! Check if Output is restricted to an energy window
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:dowindow',mt_log,1,pars%dowindow))
|
||||
& call errquit(pname//'could not read from rtdb',105,RTDB_ERR)
|
||||
if(pars%dowindow) then
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:elower',mt_dbl,1,pars%elower))
|
||||
& call errquit(pname//'could not read from rtdb',108,RTDB_ERR)
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:eupper',mt_dbl,1,pars%eupper))
|
||||
& call errquit(pname//'could not read from rtdb',110,RTDB_ERR)
|
||||
endif
|
||||
|
||||
! Number of roots
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:nroots',mt_int,1,pars%nroots))
|
||||
& call errquit(pname//'could not read from rtdb',115,RTDB_ERR)
|
||||
|
||||
! Starting space size
|
||||
if(.not.rtdb_get(pars%rtdb, 'bse:nspace', mt_int, 1, pars%nspace))
|
||||
& call errquit(pname//'could not read from rtdb', 119, RTDB_ERR)
|
||||
|
||||
! Maximum Number of roots
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:nmax',mt_int,1,pars%nmax))
|
||||
& call errquit(pname//'could not read from rtdb',123,RTDB_ERR)
|
||||
if ((pars%nmax.le.5*pars%nroots).and.pars%davidson) then
|
||||
call errquit(pname//'maxroots should be at least 5*nroots',
|
||||
& 126,INPUT_ERR)
|
||||
endif
|
||||
if((pars%nroots*pars%nspace.gt.pars%nmax).and.pars%davidson)
|
||||
& call errquit(pname//'maxroots should be at least nroots*nspace',
|
||||
& 130, INPUT_ERR)
|
||||
|
||||
! Maximum number of iterations
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:maxiter',mt_int,1,pars%maxiter))
|
||||
& call errquit(pname//'could not read from rtdb',134,RTDB_ERR)
|
||||
! Get some info from DFT module
|
||||
call dft_rdinput(pars%rtdb)
|
||||
if (cdfit) then
|
||||
if (.not.bas_destroy(cd_bas_han))
|
||||
$ call errquit(pname//'failed to destroy cd_bas',0,BASIS_ERR)
|
||||
endif
|
||||
if (.not.rtdb_get(pars%rtdb, 'dft:itol2e',mt_int,1,itol2e))
|
||||
$ call errquit(pname//'failed to read itol2e',0,RTDB_ERR)
|
||||
pars%tol2e = 10.0d0**(-itol2e-2)
|
||||
|
||||
! Fitting basis
|
||||
if (.not.bas_create(cd_bas_han,'ri basis'))
|
||||
$ call errquit(pname//'bas_create failed',0,BASIS_ERR)
|
||||
if (.not.bas_rtdb_load(pars%rtdb,geom,cd_bas_han,'ri basis'))then
|
||||
if (.not.bas_rtdb_load(pars%rtdb,geom,cd_bas_han,'cd basis'))
|
||||
$ call errquit(pname//'a "cd basis" or "ri basis" is needed',0,
|
||||
$ BASIS_ERR)
|
||||
endif
|
||||
call int_init(pars%rtdb, 1, cd_bas_han)
|
||||
if (.not.int_norm_2c(pars%rtdb, cd_bas_han))
|
||||
$ call errquit(pname//'int_norm_2c failed',0,INT_ERR)
|
||||
call int_terminate()
|
||||
if (.not. bas_numbf(cd_bas_han, nbf_cd))
|
||||
$ call errquit(pname//'basis set error',0,BASIS_ERR)
|
||||
if (.not. bas_nprim_cn_max(cd_bas_han,nbf_cd_mxprim))
|
||||
$ call errquit(pname//'basis set error:', 86, BASIS_ERR)
|
||||
if (.not. bas_high_angular(cd_bas_han,nbf_cd_mxang))
|
||||
$ call errquit(pname//'basis set error:', 86, BASIS_ERR)
|
||||
if (.not. bas_ncontr_cn_max(cd_bas_han,nbf_cd_mxcont))
|
||||
$ call errquit(pname//'basis set error:', 86, BASIS_ERR)
|
||||
if (.not. bas_nbf_cn_max(cd_bas_han,nbf_cd_mxnbf_cn))
|
||||
$ call errquit(pname//'basis set error:', 86, BASIS_ERR)
|
||||
if (.not. bas_nbf_ce_max(cd_bas_han,nbf_cd_mxnbf_ce))
|
||||
$ call errquit(pname//'basis set error:', 86, BASIS_ERR)
|
||||
if (.not. bas_numcont(cd_bas_han,nshells_cd))
|
||||
$ call errquit(pname//'basis set error:', 86, BASIS_ERR)
|
||||
if (bas_is_spherical(ao_bas_han).and.
|
||||
$ (.not.bas_is_spherical(cd_bas_han)))
|
||||
$ call int_app_set_no_texas(pars%rtdb)
|
||||
pars%nri = nbf_cd
|
||||
|
||||
!Spin multiplicity, occupations
|
||||
if (.not.rtdb_get(pars%rtdb,'dft:ipol',mt_int,1,ipol))
|
||||
$ call errquit(pname//'failed to read ipol',0,RTDB_ERR)
|
||||
if (.not.rtdb_get(pars%rtdb, 'dft:noc',mt_int,2,noc))
|
||||
$ call errquit(pname//'failed to read noc',0,RTDB_ERR)
|
||||
if (.not.rtdb_get(pars%rtdb,'dft:mult',mt_int,1,mult))
|
||||
$ call errquit(pname//'failed to read mult',0,RTDB_ERR)
|
||||
pars%ipol = ipol
|
||||
|
||||
! MO vectors
|
||||
if (.not. rtdb_cget(pars%rtdb, 'dft:output vectors',1,movecs_out))
|
||||
$ call errquit(pname//'no final MO found',0,RTDB_ERR)
|
||||
if (.not.ma_push_get(mt_dbl,nbf_ao*ipol,
|
||||
$ 'mf evals', pars%l_mf_evals, pars%k_mf_evals))
|
||||
$ call errquit(pname//'failed to allocate evals',0,MA_ERR)
|
||||
if (.not.ma_push_get(mt_dbl,nbf_ao*ipol,
|
||||
$ 'gw evals', pars%l_gw_evals, pars%k_gw_evals))
|
||||
$ call errquit(pname//'failed to allocate evals',0,MA_ERR)
|
||||
if (.not.ma_push_get(mt_dbl,nbf_ao*ipol,'occupancies', pars%l_occ,
|
||||
$ pars%k_occ))
|
||||
$ call errquit(pname//'failed to allocate occ',0,MA_ERR)
|
||||
if (.not.movecs_read_header(movecs_out,title,basisname,
|
||||
$ scftype,nbf_temp,ipol_temp,nmo_temp,2))
|
||||
$ call errquit(pname//'failed to read MO header',0,DISK_ERR)
|
||||
if (nbf_ao.ne.nbf_temp)
|
||||
$ call errquit(pname//'corrupted MO vectors',0,DISK_ERR)
|
||||
if (ipol.ne.ipol_temp)
|
||||
$ call errquit(pname//'corrupted MO vectors',0,DISK_ERR)
|
||||
|
||||
pars%nmo = nmo_temp(1)
|
||||
do i=1,ipol
|
||||
pars%g_movecs(i) = ga_create_atom_blocked(geom,
|
||||
$ ao_bas_han,'MO eigenvectors')
|
||||
if (.not.movecs_read(movecs_out,i,
|
||||
$ dbl_mb(pars%k_occ+(i-1)*nbf_ao),
|
||||
$ dbl_mb(pars%k_mf_evals+(i-1)*nbf_ao),
|
||||
$ pars%g_movecs(i)))
|
||||
$ call errquit(pname//'failed to read MO vectors',0,DISK_ERR)
|
||||
|
||||
nocc = 0
|
||||
nvir = 0
|
||||
do j=1,pars%nmo
|
||||
if (dbl_mb(pars%k_occ+(i-1)*nbf_ao+j-1).gt.0.d0) then
|
||||
nocc = nocc + 1
|
||||
cycle
|
||||
else
|
||||
nvir = nvir + 1
|
||||
endif
|
||||
enddo
|
||||
pars%nocc(i) = nocc
|
||||
pars%nvir(i) = nvir
|
||||
if (nocc+nvir.ne.pars%nmo)
|
||||
$ call errquit(pname//'something went wrong',0,0)
|
||||
pars%npoles(i) = pars%nocc(i)*pars%nvir(i)
|
||||
enddo
|
||||
|
||||
if ((pars%nroots.eq.-1).or.(pars%nroots.gt.sum(pars%npoles))) then
|
||||
pars%nroots = sum(pars%npoles)
|
||||
endif
|
||||
|
||||
if (pars%nmax.gt.sum(pars%npoles)) then
|
||||
pars%nmax = sum(pars%npoles)
|
||||
endif
|
||||
|
||||
|
||||
! Obtain NCAP DD shift
|
||||
1001 format(2X,A12,F8.3,A3)
|
||||
if (.not.rtdb_get(pars%rtdb,'bse:ncap',mt_log,1,pars%ncap))
|
||||
& call errquit(pname//'rtdb_get failed',176,RTDB_ERR)
|
||||
if (.not.rtdb_get(pars%rtdb,'bse:ncapr',mt_log,1,ncapr))
|
||||
& call errquit(pname//'rtdb_get failed',176,RTDB_ERR)
|
||||
|
||||
pars%ncap = ncapr .or. pars%ncap
|
||||
|
||||
if (pars%ncap) then
|
||||
zeta = 0.304121d0
|
||||
if (abs(xfac(87)).gt.1d-10) zeta=0.5d0
|
||||
aq2 = 2d0/9d0 * (1d0-zeta)**2
|
||||
do i=1,ipol
|
||||
ehomo = dbl_mb(pars%k_mf_evals+(i-1)*nbf_ao+pars%nocc(i)-1)
|
||||
vxddp = -0.5d0*aq2*(1d0 + sqrt(1d0-4d0*ehomo/aq2))!*0.85d0
|
||||
vxddm = -0.5d0*aq2*(1d0 - sqrt(1d0-4d0*ehomo/aq2))!*0.85d0
|
||||
pars%vxddp(i) = vxddp
|
||||
pars%vxddm(i) = vxddm
|
||||
if (pars%me.eq.0) then
|
||||
write(luout,*) ' Shifting eigvals by 100% NCAP DD'
|
||||
write(luout,1001) ' v_x^{DD-}: ',vxddp*ha2ev, ' eV'
|
||||
write(luout,1001) ' v_x^(DD+): ',vxddm*ha2ev, ' eV'
|
||||
endif
|
||||
call daxpy(pars%nocc(i),1d0,vxddp,0,
|
||||
& dbl_mb(pars%k_mf_evals+(i-1)*nbf_ao),1)
|
||||
call daxpy(pars%nvir(i),1d0,vxddm,0,
|
||||
& dbl_mb(pars%k_mf_evals+(i-1)*nbf_ao+pars%nocc(i)),1)
|
||||
enddo
|
||||
endif
|
||||
|
||||
! Read GW QP energies
|
||||
!if (.not.pars%ncap) then
|
||||
if(skipgw) then
|
||||
call dcopy(nbf_ao*pars%ipol,dbl_mb(pars%k_mf_evals),1,
|
||||
& dbl_mb(pars%k_gw_evals),1)
|
||||
else
|
||||
call util_file_name('gwevals',.false.,.false.,gwevals)
|
||||
call util_file_name_resolve(gwevals,.false.)
|
||||
if (pars%me.eq.0) then
|
||||
open(unit=unitno,status='old',form='unformatted',file=gwevals)
|
||||
do i=1,pars%ipol
|
||||
call sread(unitno, dbl_mb(pars%k_gw_evals+(i-1)*nbf_ao),
|
||||
& pars%nmo)
|
||||
enddo
|
||||
close(unit=unitno)
|
||||
endif
|
||||
call ga_brdcst(mt_dbl,dbl_mb(pars%k_gw_evals),
|
||||
& ma_sizeof(mt_dbl,nbf_ao*pars%ipol,mt_byte),0)
|
||||
endif
|
||||
|
||||
!3-center ERIs
|
||||
call int_init(pars%rtdb, 2, (/ao_bas_han, cd_bas_han/))
|
||||
call print_integrals((/ao_bas_han,cd_bas_han/),.false.)
|
||||
call schwarz_init(geom, ao_bas_han)
|
||||
call scf_get_fock_param(pars%rtdb, pars%tol2e)
|
||||
|
||||
! Obtain irreps information
|
||||
do i=1,ipol
|
||||
if(.not.ma_push_get(mt_int,nbf_ao,'irreps',pars%l_irs(i),
|
||||
$ pars%k_irs(i))) call errquit
|
||||
$ ('bse_init: failed to allocate irs',0,MA_ERR)
|
||||
if(.not.ga_create(mt_dbl,nbf_ao,pars%nmo,'Work',nbf_ao,-1,
|
||||
$ g_work))
|
||||
$ call errquit('bse_init: failed to allocate work',0,GA_ERR)
|
||||
call ga_copy_patch('n',pars%g_movecs(i),1,nbf_ao,1,pars%nmo,
|
||||
$ g_work,1,nbf_ao,1,pars%nmo)
|
||||
call sym_movecs_adapt(ao_bas_han, 1.0d-8, g_work,
|
||||
$ int_mb(pars%k_irs(i)),nmix(i))
|
||||
if(.not.ga_destroy(g_work)) call errquit
|
||||
$ ('bse_init: failed to destroy g_work',0, GA_ERR)
|
||||
enddo
|
||||
|
||||
!Clean previous grid
|
||||
call grid_cleanup(.false.)
|
||||
|
||||
!Print expected memory requirements
|
||||
|
||||
!Compute three-center integrals
|
||||
call bse_ri_init(pars)
|
||||
|
||||
end subroutine bse_init
|
||||
169
src/bsemol/bse_input.F
Normal file
169
src/bsemol/bse_input.F
Normal file
|
|
@ -0,0 +1,169 @@
|
|||
subroutine bse_input(rtdb)
|
||||
c
|
||||
c Input parser for molecular BSE module
|
||||
c
|
||||
c Daniel Mejia-Rodriguez
|
||||
c November 2022
|
||||
c
|
||||
implicit none
|
||||
c
|
||||
#include "errquit.fh"
|
||||
#include "util.fh"
|
||||
#include "inp.fh"
|
||||
#include "rtdb.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "stdio.fh"
|
||||
c
|
||||
integer rtdb
|
||||
character*255 test
|
||||
character(*), parameter :: pname = "bse_input: "
|
||||
c
|
||||
integer, parameter :: num_dirs = 13
|
||||
character*11 dirs(num_dirs)
|
||||
data dirs / 'method','tda','ncap','ncapr','singlet','triplet',
|
||||
& 'outwin', 'nroots', 'maxroots', 'skipgw', 'maxiter',
|
||||
& 'nstart', 'end' /
|
||||
c
|
||||
integer ind, itemp
|
||||
double precision threshold, eta, temp
|
||||
double precision, parameter :: ha2ev = 27.211386245988d0
|
||||
c
|
||||
c Set default values and put them in RTDB
|
||||
c
|
||||
call bse_defaults(rtdb)
|
||||
c
|
||||
c Parse the input file
|
||||
c
|
||||
10 if (.not. inp_read()) then
|
||||
call errquit(pname//'failed to read input',0,INPUT_ERR)
|
||||
endif
|
||||
c
|
||||
if (.not. inp_a(test)) then
|
||||
call errquit(pname//'failed to read keyword',0,INPUT_ERR)
|
||||
endif
|
||||
c
|
||||
if (.not.inp_match(num_dirs, .false., test, dirs, ind)) then
|
||||
call errquit(pname//'unknown directive', 0, INPUT_ERR)
|
||||
end if
|
||||
c
|
||||
goto (100, 200, 300, 301, 400, 500, 600, 700, 800, 900, 1000,
|
||||
& 1100, 9999) ind
|
||||
|
||||
call errquit(pname//'unknown directive',ind, INPUT_ERR)
|
||||
c
|
||||
c -----------------------------------------------------------------
|
||||
c
|
||||
c method
|
||||
c
|
||||
100 if (.not. inp_a(test))
|
||||
& call errquit(pname//'could not read method', 100, INPUT_ERR)
|
||||
|
||||
if (inp_compare(.false.,'analytic',test)) then
|
||||
if (.not.rtdb_put(rtdb, 'bse:analytic', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 101, RTDB_ERR)
|
||||
elseif (inp_compare(.false.,'davidson',test)) then
|
||||
if (.not.rtdb_put(rtdb, 'bse:davidson', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 102, RTDB_ERR)
|
||||
if (.not.rtdb_put(rtdb, 'bse:analytic', mt_log, 1, .false.))
|
||||
& call errquit(pname//'rtdb_put failed', 101, RTDB_ERR)
|
||||
elseif (inp_compare(.false.,'lanczos',test)) then
|
||||
if (.not.rtdb_put(rtdb, 'bse:lanczos', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 103, RTDB_ERR)
|
||||
if (.not.rtdb_put(rtdb, 'bse:analytic', mt_log, 1, .false.))
|
||||
& call errquit(pname//'rtdb_put failed', 101, RTDB_ERR)
|
||||
else
|
||||
call errquit(pname//'unknown BSE method',104,INPUT_ERR)
|
||||
endif
|
||||
goto 10
|
||||
c
|
||||
c use Tamm-Dancoff approximation
|
||||
c
|
||||
200 if (.not.rtdb_put(rtdb, 'bse:tda', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 200, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c use NCAP DD correction for eigenvalues
|
||||
c
|
||||
300 if (.not.rtdb_put(rtdb, 'bse:ncap', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 300, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c use NCAPR DD correction for eigenvalues
|
||||
c
|
||||
301 if (.not.rtdb_put(rtdb, 'bse:ncapr', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 301, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c compute singlet transitions
|
||||
c
|
||||
400 if (.not.rtdb_put(rtdb, 'bse:singlet', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 300, RTDB_ERR)
|
||||
if (.not.rtdb_put(rtdb, 'bse:triplet', mt_log, 1, .false.))
|
||||
& call errquit(pname//'rtdb_put failed', 300, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c compute triplet transitions
|
||||
c
|
||||
500 if (.not.rtdb_put(rtdb, 'bse:singlet', mt_log, 1, .false.))
|
||||
& call errquit(pname//'rtdb_put failed', 300, RTDB_ERR)
|
||||
if (.not.rtdb_put(rtdb, 'bse:triplet', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 300, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c output energy window
|
||||
c
|
||||
600 if (.not.rtdb_put(rtdb, 'bse:dowindow', mt_log, 1, .true.))
|
||||
& call errquit(pname//'rtdb_put failed', 600, RTDB_ERR)
|
||||
if(.not.inp_f(temp))
|
||||
& call errquit(pname//'could not read lower energy',600,INPUT_ERR)
|
||||
if(.not.rtdb_put(rtdb,'bse:elower',mt_dbl,1,temp))
|
||||
& call errquit(pname//'rtdb_put failed', 600, RTDB_ERR)
|
||||
if(.not.inp_f(temp))
|
||||
& call errquit(pname//'could not read upper energy',600,INPUT_ERR)
|
||||
if(.not.rtdb_put(rtdb,'bse:eupper',mt_dbl,1,temp))
|
||||
& call errquit(pname//'rtdb_put failed', 600, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c number of roots
|
||||
c
|
||||
700 if(.not.inp_i(itemp))
|
||||
& call errquit(pname//'could not read no. of roots',700,INPUT_ERR)
|
||||
if(.not.rtdb_put(rtdb,'bse:nroots',mt_int,1,itemp))
|
||||
& call errquit(pname//'rtdb_put failed', 700, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c maximum number of trial vectors
|
||||
c
|
||||
800 if(.not.inp_i(itemp))
|
||||
& call errquit(pname//'could not read maxroots',800,INPUT_ERR)
|
||||
if(.not.rtdb_put(rtdb,'bse:nmax',mt_int,1,itemp))
|
||||
& call errquit(pname//'rtdb_put failed', 800, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c Skip GW calculation
|
||||
c
|
||||
900 if(.not.rtdb_put(rtdb,'bse:skipgw',mt_log,1,.true.))
|
||||
& call errquit(pname//'rtdb_put failed', 900, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c Maximum number of iterations
|
||||
c
|
||||
1000 if(.not.inp_i(itemp))
|
||||
& call errquit(pname//'could not read maxiter', 1000, INPUT_ERR)
|
||||
if(.not.rtdb_put(rtdb,'bse:maxiter',mt_int,1,itemp))
|
||||
& call errquit(pname//'rtdb_put failed', 1000, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c Size of starting space
|
||||
c
|
||||
1100 if(.not.inp_i(itemp))
|
||||
& call errquit(pname//'could not read nstart', 1100, INPUT_ERR)
|
||||
if(.not.rtdb_put(rtdb,'bse:nspace', mt_int, 1, itemp))
|
||||
& call errquit(pname//'rtdb_put failed', 1100, RTDB_ERR)
|
||||
goto 10
|
||||
c
|
||||
c normal termination
|
||||
c
|
||||
9999 return
|
||||
c
|
||||
end subroutine bse_input
|
||||
9
src/bsemol/bse_lanczos.F
Normal file
9
src/bsemol/bse_lanczos.F
Normal file
|
|
@ -0,0 +1,9 @@
|
|||
subroutine bse_lanczos(pars)
|
||||
|
||||
implicit none
|
||||
|
||||
#include "bse.fh"
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
|
||||
end subroutine bse_lanczos
|
||||
515
src/bsemol/bse_oscstr.F
Normal file
515
src/bsemol/bse_oscstr.F
Normal file
|
|
@ -0,0 +1,515 @@
|
|||
subroutine bse_oscstr(pars,omega,totpoles)
|
||||
implicit none
|
||||
#include "errquit.fh"
|
||||
#include "stdio.fh"
|
||||
#include "global.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "bse.fh"
|
||||
#include "cdft.fh"
|
||||
#include "rtdb.fh"
|
||||
#include "sym.fh"
|
||||
#include "util.fh"
|
||||
#include "util_params.fh"
|
||||
#include "inp.fh"
|
||||
type(bse_params_t) :: pars
|
||||
|
||||
character(*), parameter :: pname = 'bse_oscstr: '
|
||||
character(len=7) :: key
|
||||
|
||||
integer totpoles,npoles
|
||||
double precision omega(*)
|
||||
double precision oscstr(3),oscstr_tot,tmp,moms(20),maxr,r(1)
|
||||
|
||||
integer llpole,ulpole,iroot,ixyz,ipole,mag_transdip(3)
|
||||
|
||||
integer isym,jsym,i,j,k,l,pol,occ,vir,ioff
|
||||
integer g_td(2),g_xy(2),g_dipmag, g_ovlp, g_work, g_corr
|
||||
integer g_tdtot(2)
|
||||
integer k_corr, l_corr, ulloop
|
||||
character*4 oname,vname
|
||||
|
||||
integer g_dipx, g_dipy, g_dipz
|
||||
integer g_qdrxx, g_qdrxy, g_qdrxz, g_qdryy, g_qdryz, g_qdrzz
|
||||
integer g_work2, g_work3, g_x(2), g_y(2)
|
||||
|
||||
logical ltransden,stat
|
||||
double precision origin(3)
|
||||
double precision na,nb,s2,s2_full
|
||||
double precision espread
|
||||
data origin/0d0,0d0,0d0/
|
||||
|
||||
character*(nw_max_path_len) fn_civecs, fn_transden
|
||||
logical,external :: dmat_to_file
|
||||
integer,external :: ga_create_atom_blocked
|
||||
|
||||
character*5 spin(2)
|
||||
double precision x(totpoles),y(totpoles)
|
||||
double precision thresh ! Threshold for printing
|
||||
parameter (thresh=5.0d-2)
|
||||
double precision xe(3), xh(3), xe2_exp, xh2_exp, xe_exp_2
|
||||
double precision xh_exp_2, xhxe_exp, dhtoe, dexc, sigma_h, sigma_e
|
||||
double precision omega_norm, cov, pcc, dcd
|
||||
spin(1)='alpha'
|
||||
spin(2)='beta '
|
||||
|
||||
if(.not.rtdb_get(pars%rtdb,'bse:ltransden',mt_log,1,ltransden))
|
||||
$ ltransden = .false.
|
||||
if(.not.nga_create(mt_dbl,3,(/nbf_ao,nbf_ao,3/),'mag_dipole',
|
||||
$ (/nbf_ao,-1,-1/),g_dipmag))
|
||||
$ call errquit('bse_oscstr: nga_create failed g_dipmag',0,GA_ERR)
|
||||
call ga_zero(g_dipmag)
|
||||
call int_giao_1ega(ao_bas_han,ao_bas_han,g_dipmag,'angmom',origin,
|
||||
$ 1,.false.)
|
||||
call ga_scale(g_dipmag,0.5d0)
|
||||
|
||||
g_ovlp = ga_create_atom_blocked(geom, ao_bas_han, 'Overlap')
|
||||
call ga_zero(g_ovlp)
|
||||
call int_1e_ga(ao_bas_han, ao_bas_han, g_ovlp, 'overlap', oskel)
|
||||
if(oskel) call sym_symmetrize(geom, ao_bas_han, .false., g_ovlp)
|
||||
|
||||
if (pars%ipol.eq.2) then
|
||||
if(.not.ma_push_get(mt_dbl,pars%nmo*pars%nmo,'corr',
|
||||
$ l_corr,k_corr))
|
||||
$ call errquit('bse_oscstr: failed to create MA',0,MA_ERR)
|
||||
if(.not.ga_create(mt_dbl,nbf_ao,pars%nmo,'work',0,0,g_work))
|
||||
$ call errquit('bse_oscstr: failed to create GA',0,GA_ERR)
|
||||
if(.not.ga_create(mt_dbl,pars%nmo,pars%nmo,'work',
|
||||
$ pars%nmo,0,g_corr))
|
||||
$ call errquit('bse_oscstr: failed to create GA',0,GA_ERR)
|
||||
call ga_dgemm('n','n',nbf_ao,pars%nmo,nbf_ao,1d0,g_ovlp,
|
||||
$ pars%g_movecs(2),0d0,g_work)
|
||||
call ga_dgemm('t','n',pars%nmo,pars%nmo,nbf_ao,1d0,
|
||||
$ pars%g_movecs(1),g_work,0d0,g_corr)
|
||||
call ga_get(g_corr,1,pars%nmo,1,pars%nmo,dbl_mb(k_corr),
|
||||
$ pars%nmo)
|
||||
stat = ga_destroy(g_work).and.ga_destroy(g_corr)
|
||||
if(.not.stat)
|
||||
$ call errquit('bse_osctr: failed to destroy GA',58,GA_ERR)
|
||||
endif
|
||||
|
||||
g_dipx = ga_create_atom_blocked(geom, ao_bas_han, 'X')
|
||||
g_dipy = ga_create_atom_blocked(geom, ao_bas_han, 'Y')
|
||||
g_dipz = ga_create_atom_blocked(geom, ao_bas_han, 'Z')
|
||||
g_qdrxx = ga_create_atom_blocked(geom, ao_bas_han, 'XX')
|
||||
g_qdrxy = ga_create_atom_blocked(geom, ao_bas_han, 'XY')
|
||||
g_qdrxz = ga_create_atom_blocked(geom, ao_bas_han, 'XZ')
|
||||
g_qdryy = ga_create_atom_blocked(geom, ao_bas_han, 'YY')
|
||||
g_qdryz = ga_create_atom_blocked(geom, ao_bas_han, 'YZ')
|
||||
g_qdrzz = ga_create_atom_blocked(geom, ao_bas_han, 'ZZ')
|
||||
g_work = ga_create_atom_blocked(geom, ao_bas_han, 'work1')
|
||||
g_work2 = ga_create_atom_blocked(geom, ao_bas_han, 'work2')
|
||||
g_work3 = ga_create_atom_blocked(geom, ao_bas_han, 'work3')
|
||||
|
||||
call int_dip_ga(ao_bas_han, ao_bas_han, g_dipx, g_dipy, g_dipz)
|
||||
call int_qdr_ga(ao_bas_han, ao_bas_han, g_qdrxx, g_qdrxy, g_qdrxz,
|
||||
$ g_qdryy, g_qdryz, g_qdrzz)
|
||||
|
||||
|
||||
|
||||
npoles = pars%npoles(1)
|
||||
if (pars%ipol.gt.1) npoles = npoles + pars%npoles(2)
|
||||
|
||||
|
||||
if (pars%singlet) then
|
||||
key = 'singlet'
|
||||
else
|
||||
key = 'triplet'
|
||||
endif
|
||||
|
||||
! Get the lower and upper limits
|
||||
llpole = -1
|
||||
if (pars%dowindow .and. (.not.pars%davidson)) then
|
||||
iroot = 0
|
||||
do ipole=1,npoles
|
||||
if (omega(ipole).lt.pars%elower) cycle
|
||||
if (llpole.eq.-1) llpole = ipole
|
||||
iroot = iroot + 1
|
||||
if (omega(ipole).gt.pars%eupper) then
|
||||
ulpole = ipole - 1
|
||||
exit
|
||||
endif
|
||||
if (iroot.eq.pars%nroots) then
|
||||
ulpole = ipole
|
||||
exit
|
||||
endif
|
||||
enddo
|
||||
else
|
||||
llpole = 1
|
||||
ulpole = pars%nroots
|
||||
endif
|
||||
|
||||
|
||||
! Get symmetry of ground state
|
||||
isym = 1
|
||||
if (pars%ipol.eq.2) then
|
||||
do i=1,pars%ipol
|
||||
do j=1,pars%nocc(i)
|
||||
isym = ieor(isym-1,int_mb(pars%k_irs(i)+j-1)-1)+1
|
||||
enddo
|
||||
enddo
|
||||
endif
|
||||
call sym_irrepname(geom, isym, oname)
|
||||
|
||||
|
||||
! Allocate transition density matrices
|
||||
do i=1,pars%ipol
|
||||
if(.not.ga_create(mt_dbl,nbf_ao,nbf_ao,'Transition density',
|
||||
$ -1,-1,g_td(i))) call errquit
|
||||
$ ('bse_oscstr: failed to create g_td',0,GA_ERR)
|
||||
if(.not.ga_create(mt_dbl,nbf_ao,nbf_ao,'Transition density',
|
||||
$ -1,-1,g_tdtot(i))) call errquit
|
||||
$ ('bse_oscstr: failed to create g_tdtot',0,GA_ERR)
|
||||
if(.not.ga_create(mt_dbl,pars%npoles(i),1,'X+Y vector',-1,1,
|
||||
$ g_xy(i))) call errquit
|
||||
$ ('bse_oscstr: failed to create g_xy',0,GA_ERR)
|
||||
if(.not.ga_create(mt_dbl,pars%npoles(i),1,'X vector',-1,1,
|
||||
$ g_x(i))) call errquit
|
||||
$ ('bse_oscstr: failed to create g_x',0,GA_ERR)
|
||||
if(.not.ga_create(mt_dbl,pars%npoles(i),1,'Y vector',-1,1,
|
||||
$ g_y(i))) call errquit
|
||||
$ ('bse_oscstr: failed to create g_y',0,GA_ERR)
|
||||
enddo
|
||||
|
||||
! Compute the oscillator strengths around an energy window
|
||||
iroot = 0
|
||||
do ipole=llpole,ulpole
|
||||
iroot = iroot + 1
|
||||
oscstr_tot = 0d0
|
||||
moms(:) = 0d0
|
||||
oscstr(:) = 0d0
|
||||
mag_transdip = 0d0
|
||||
|
||||
call ga_copy_patch('n',pars%g_apb,1,pars%npoles(1),ipole,ipole,
|
||||
& g_x(1),1,pars%npoles(1),1,1)
|
||||
if(pars%ipol.gt.1) then
|
||||
call ga_copy_patch('n',pars%g_apb,pars%npoles(1)+1,npoles,
|
||||
& ipole,ipole,g_x(2),1,pars%npoles(2),1,1)
|
||||
endif
|
||||
call tddft_transfm(1, g_x, pars%g_movecs, nbf_ao, pars%nocc,
|
||||
& (/pars%nmo, pars%nmo/), (/0,0/), (/0,0/),
|
||||
& pars%ipol, g_td, 1)
|
||||
call ga_transpose(g_td(1), g_tdtot(1))
|
||||
if(pars%ipol.gt.1) call ga_transpose(g_td(2), g_tdtot(2))
|
||||
|
||||
if(pars%davidson) then
|
||||
call ga_copy_patch('n',pars%g_amb,1,pars%npoles(1),
|
||||
I ipole,ipole,
|
||||
& g_y(1),1,pars%npoles(1),1,1)
|
||||
if(pars%ipol.gt.1) then
|
||||
call ga_copy_patch('n',pars%g_amb, pars%npoles(1)+1, npoles,
|
||||
& ipole, ipole, g_y(2), 1, pars%npoles(2),
|
||||
& 1, 1)
|
||||
endif
|
||||
call tddft_transfm(1, g_y, pars%g_movecs, nbf_ao, pars%nocc,
|
||||
& (/pars%nmo, pars%nmo/), (/0,0/), (/0,0/),
|
||||
& pars%ipol, g_td, 1)
|
||||
call ga_add(1.0d0, g_td(1), 1.0d0, g_tdtot(1), g_td(1))
|
||||
if(pars%ipol.gt.1)
|
||||
& call ga_add(1.0d0, g_td(2), 1.0d0, g_tdtot(2), g_td(2))
|
||||
else
|
||||
call ga_copy(g_tdtot(1), g_td(1))
|
||||
if (pars%ipol.gt.1) call ga_copy(g_tdtot(2), g_td(2))
|
||||
endif
|
||||
|
||||
do i=1,pars%ipol
|
||||
call multipole_density(ao_bas_han,origin,3,g_td(i),moms,20)
|
||||
do ixyz=1,3
|
||||
mag_transdip(ixyz) = mag_transdip(ixyz) +
|
||||
$ nga_ddot_patch(g_td(i),'n',(/1,1/),(/nbf_ao,nbf_ao/),
|
||||
$ g_dipmag,'n',(/1,1,ixyz/),
|
||||
$ (/nbf_ao,nbf_ao,ixyz/))
|
||||
enddo
|
||||
enddo
|
||||
|
||||
if(pars%ipol.gt.1) call ga_add(1d0,g_td(1),1d0,g_td(2),g_td(1))
|
||||
|
||||
! Indicators
|
||||
call ga_dgemm('t','n',nbf_ao,nbf_ao,nbf_ao,1d0,g_td(1),g_ovlp,
|
||||
$ 0d0,g_work)
|
||||
call ga_dgemm('n','n',nbf_ao,nbf_ao,nbf_ao,1d0,g_work,g_td(1),
|
||||
$ 0d0,g_work2)
|
||||
omega_norm = ga_ddot(g_work2,g_ovlp)
|
||||
xe(1) = ga_ddot(g_work2,g_dipx)/omega_norm
|
||||
xe(2) = ga_ddot(g_work2,g_dipy)/omega_norm
|
||||
xe(3) = ga_ddot(g_work2,g_dipz)/omega_norm
|
||||
xe_exp_2 = xe(1)**2 + xe(2)**2 + xe(3)**2
|
||||
xe2_exp = (ga_ddot(g_work2,g_qdrxx) +
|
||||
$ ga_ddot(g_work2,g_qdryy) +
|
||||
$ ga_ddot(g_work2,g_qdrzz))/omega_norm
|
||||
|
||||
|
||||
call ga_dgemm('n','n',nbf_ao,nbf_ao,nbf_ao,1d0,g_td(1),g_ovlp,
|
||||
$ 0d0,g_work)
|
||||
call ga_dgemm('n','t',nbf_ao,nbf_ao,nbf_ao,1d0,g_work,g_td(1),
|
||||
$ 0d0,g_work2)
|
||||
xh(1) = ga_ddot(g_work2,g_dipx)/omega_norm
|
||||
xh(2) = ga_ddot(g_work2,g_dipy)/omega_norm
|
||||
xh(3) = ga_ddot(g_work2,g_dipz)/omega_norm
|
||||
xh_exp_2 = xh(1)**2 + xh(2)**2 + xh(3)**2
|
||||
xh2_exp = (ga_ddot(g_work2,g_qdrxx) +
|
||||
$ ga_ddot(g_work2,g_qdryy) +
|
||||
$ ga_ddot(g_work2,g_qdrzz))/omega_norm
|
||||
|
||||
|
||||
xhxe_exp = 0d0
|
||||
call ga_dgemm('n','n',nbf_ao,nbf_ao,nbf_ao,1d0,g_td(1),g_dipx,
|
||||
$ 0d0,g_work)
|
||||
call ga_dgemm('n','t',nbf_ao,nbf_ao,nbf_ao,1d0,g_work,g_td(1),
|
||||
$ 0d0,g_work2)
|
||||
xhxe_exp = xhxe_exp + ga_ddot(g_work2,g_dipx)/omega_norm
|
||||
call ga_dgemm('n','n',nbf_ao,nbf_ao,nbf_ao,1d0,g_td(1),g_dipy,
|
||||
$ 0d0,g_work)
|
||||
call ga_dgemm('n','t',nbf_ao,nbf_ao,nbf_ao,1d0,g_work,g_td(1),
|
||||
$ 0d0,g_work2)
|
||||
xhxe_exp = xhxe_exp + ga_ddot(g_work2,g_dipy)/omega_norm
|
||||
call ga_dgemm('n','n',nbf_ao,nbf_ao,nbf_ao,1d0,g_td(1),g_dipz,
|
||||
$ 0d0,g_work)
|
||||
call ga_dgemm('n','t',nbf_ao,nbf_ao,nbf_ao,1d0,g_work,g_td(1),
|
||||
$ 0d0,g_work2)
|
||||
xhxe_exp = xhxe_exp + ga_ddot(g_work2,g_dipz)/omega_norm
|
||||
|
||||
|
||||
sigma_h = dsqrt(xh2_exp - xh_exp_2)
|
||||
sigma_e = dsqrt(xe2_exp - xe_exp_2)
|
||||
|
||||
dhtoe = dsqrt((xe(1)-xh(1))**2 +
|
||||
$ (xe(2)-xh(2))**2 +
|
||||
$ (xe(3)-xh(3))**2)
|
||||
dexc = dsqrt(xe2_exp + xh2_exp - 2d0*xhxe_exp)
|
||||
cov = xhxe_exp - xe(1)*xh(1) - xe(2)*xh(2) - xe(3)*xh(3)
|
||||
pcc = cov/sigma_e/sigma_h
|
||||
dcd = dhtoe - 0.5d0*(sigma_e + sigma_h)
|
||||
|
||||
if (ltransden) then
|
||||
call util_file_name('tdens',.false.,.false.,fn_transden)
|
||||
call createuniquefilename(fn_transden,ipole)
|
||||
call util_file_name_resolve(fn_transden,.false.)
|
||||
if (pars%ipol.gt.1)
|
||||
& call ga_add(1d0,g_td(1),1d0,g_td(2),g_td(1))
|
||||
|
||||
call ga_symmetrize(g_td(1))
|
||||
|
||||
if(.not. dmat_to_file(g_td(1),fn_transden))
|
||||
& call errquit('bse_oscstr: dmat_to_file error',0,CALC_ERR)
|
||||
endif
|
||||
|
||||
if (pars%ipol.eq.1) then
|
||||
do i=1,20
|
||||
moms(i) = moms(i)*dsqrt(2.0d0)
|
||||
enddo
|
||||
endif
|
||||
|
||||
if (pars%davidson) then
|
||||
call ga_get(pars%g_apb,1,npoles,ipole,ipole,x,npoles)
|
||||
call ga_get(pars%g_amb,1,npoles,ipole,ipole,y,npoles)
|
||||
endif
|
||||
|
||||
if (pars%davidson.and.pars%ipol.eq.2) then
|
||||
call tddft_s2(pars%tda, (/0,0/), pars%nocc,
|
||||
& (/pars%nmo,pars%nmo/), (/0,0/),
|
||||
& pars%npoles,x,x(pars%npoles(1)+1),y,
|
||||
& y(pars%npoles(1)+1),dbl_mb(k_corr),s2)
|
||||
call bse_s2(pars%nocc,pars%nmo,pars%npoles,x,
|
||||
& x(pars%npoles(1)+1), y, y(pars%npoles(1)+1), dbl_mb(k_corr),
|
||||
& s2_full)
|
||||
if (abs(s2_full).lt.0.25d0) then
|
||||
key = 'singlet'
|
||||
elseif (abs(abs(s2_full)-2.0d0).lt.0.25) then
|
||||
key = 'triplet'
|
||||
else
|
||||
key = ' '
|
||||
endif
|
||||
endif
|
||||
|
||||
if (pars%me.eq.0) then
|
||||
call tddft_oscstr(moms, omega(ipole), mag_transdip,
|
||||
& oscstr, oscstr_tot)
|
||||
|
||||
l = 0
|
||||
maxr = 0.0d0
|
||||
pol = 1
|
||||
occ = 1
|
||||
vir = pars%nocc(1) + 1
|
||||
do i=1,pars%ipol
|
||||
if (pars%npoles(i).eq.0) cycle
|
||||
do j=1,pars%nocc(i)
|
||||
do k=pars%nocc(i)+1,pars%nmo
|
||||
l = l + 1
|
||||
call ga_get(pars%g_apb,l,l,ipole,ipole,r,1)
|
||||
if (dabs(r(1)).gt.maxr) then
|
||||
maxr = dabs(r(1))
|
||||
pol = i
|
||||
occ = j
|
||||
vir = k
|
||||
endif
|
||||
enddo
|
||||
enddo
|
||||
enddo
|
||||
jsym = ieor(isym-1,int_mb(pars%k_irs(pol)+occ-1)-1)+1
|
||||
jsym = ieor(jsym-1,int_mb(pars%k_irs(pol)+vir-1)-1)+1
|
||||
call sym_irrepname(geom, jsym, vname)
|
||||
|
||||
write(luout,9110) iroot,key,vname,omega(ipole),
|
||||
& omega(ipole)*ha2ev
|
||||
|
||||
if (pars%davidson.and.pars%ipol.eq.2) then
|
||||
write(luout,*)
|
||||
write(luout,9180) s2
|
||||
write(luout,9190) s2_full
|
||||
endif
|
||||
|
||||
if (pars%singlet) then
|
||||
write(luout,9200) moms(2),moms(3),moms(4),
|
||||
& moms(5),moms(6),moms(7),
|
||||
& moms(8),moms(9),moms(10),
|
||||
& oscstr(1),oscstr(2),oscstr(3),
|
||||
& oscstr_tot
|
||||
else
|
||||
write(luout,9210)
|
||||
endif
|
||||
|
||||
write(luout,*)
|
||||
write(luout,9340) omega_norm
|
||||
write(luout,9300) sigma_e*cau2ang
|
||||
write(luout,9310) sigma_h*cau2ang
|
||||
write(luout,9320) dhtoe*cau2ang
|
||||
write(luout,9330) dexc*cau2ang
|
||||
write(luout,9360) dcd*cau2ang
|
||||
write(luout,9350) pcc
|
||||
|
||||
|
||||
if (pars%davidson) then
|
||||
|
||||
write(luout,*)
|
||||
if (ipol.eq.1) then
|
||||
l=0
|
||||
do j=1,pars%nocc(1)
|
||||
do k=pars%nocc(1)+1,pars%nmo
|
||||
l=l+1
|
||||
if (pars%tda) then
|
||||
r=x(l)
|
||||
call sym_irrepname
|
||||
1 (geom,int_mb(pars%k_irs(1)+j-1),oname)
|
||||
call sym_irrepname
|
||||
1 (geom,int_mb(pars%k_irs(1)+k-1),vname)
|
||||
if ((dabs(r(1)).gt.thresh))
|
||||
2 write(LuOut,9120) j,oname,k,vname,r
|
||||
else
|
||||
r=x(l)
|
||||
call sym_irrepname
|
||||
1 (geom,int_mb(pars%k_irs(1)+j-1),oname)
|
||||
call sym_irrepname
|
||||
1 (geom,int_mb(pars%k_irs(1)+k-1),vname)
|
||||
if ((dabs(r(1)).gt.thresh))
|
||||
2 write(LuOut,9130) j,oname,k,vname,r
|
||||
r=y(l)
|
||||
if ((dabs(r(1)).gt.thresh))
|
||||
2 write(LuOut,9140) j,oname,k,vname,r
|
||||
endif
|
||||
enddo
|
||||
enddo
|
||||
else
|
||||
do i=1,ipol
|
||||
ioff = (i-1)*pars%npoles(1)
|
||||
l=0
|
||||
do j=1,pars%nocc(i)
|
||||
do k=pars%nocc(i)+1,pars%nmo
|
||||
l=l+1
|
||||
if (pars%tda) then
|
||||
r = x(ioff+l)
|
||||
call sym_irrepname
|
||||
1 (geom,int_mb(pars%k_irs(i)+j-1),oname)
|
||||
call sym_irrepname
|
||||
1 (geom,int_mb(pars%k_irs(i)+k-1),vname)
|
||||
if ((dabs(r(1)).gt.thresh))
|
||||
2 write(LuOut,9150) j,spin(i),oname,
|
||||
3 k,spin(i),vname,r
|
||||
else
|
||||
r = x(ioff+l)
|
||||
call sym_irrepname
|
||||
1 (geom,int_mb(pars%k_irs(i)+j-1),oname)
|
||||
call sym_irrepname
|
||||
1 (geom,int_mb(pars%k_irs(i)+k-1),vname)
|
||||
if ((dabs(r(1)).gt.thresh))
|
||||
2 write(LuOut,9160) j,spin(i),oname,
|
||||
3 k,spin(i),vname,r
|
||||
r=y(ioff+l)
|
||||
if ((dabs(r(1)).gt.thresh))
|
||||
2 write(LuOut,9170) j,spin(i),oname,
|
||||
3 k,spin(i),vname,r
|
||||
endif
|
||||
enddo
|
||||
enddo
|
||||
enddo
|
||||
endif
|
||||
call util_flush(LuOut)
|
||||
|
||||
endif
|
||||
|
||||
endif
|
||||
|
||||
call ga_sync()
|
||||
|
||||
enddo
|
||||
|
||||
if (.not.ga_destroy(g_dipmag))
|
||||
& call errquit('bse_oscstr: ga_destroy failed',0,GA_ERR)
|
||||
do i=1,pars%ipol
|
||||
if (.not.ga_destroy(g_td(i)))
|
||||
& call errquit('bse_oscstr: ga_destroy failed',0,GA_ERR)
|
||||
if (.not.ga_destroy(g_xy(i)))
|
||||
& call errquit('bse_oscstr: ga_destroy failed',0,GA_ERR)
|
||||
if (.not.ga_destroy(g_tdtot(i)))
|
||||
& call errquit('bse_oscstr: ga_destroy failed',0,GA_ERR)
|
||||
if (.not.ga_destroy(g_x(i)))
|
||||
& call errquit('bse_oscstr: ga_destroy failed',0,GA_ERR)
|
||||
if (.not.ga_destroy(g_y(i)))
|
||||
& call errquit('bse_oscstr: ga_destroy failed',0,GA_ERR)
|
||||
enddo
|
||||
if(.not.(ga_destroy(g_dipx).and.ga_destroy(g_dipy).and.
|
||||
$ ga_destroy(g_dipz).and.ga_destroy(g_qdrxx).and.
|
||||
$ ga_destroy(g_qdrxy).and.ga_destroy(g_qdrxz).and.
|
||||
$ ga_destroy(g_qdryy).and.ga_destroy(g_qdryz).and.
|
||||
$ ga_destroy(g_qdrzz).and.ga_destroy(g_ovlp).and.
|
||||
$ ga_destroy(g_work).and.ga_destroy(g_work2).and.
|
||||
$ ga_destroy(g_work3)))
|
||||
& call errquit('bse_oscstr: ga_destroy failed',0,GA_ERR)
|
||||
|
||||
|
||||
9110 format(2x,
|
||||
1 '---------------------------------------------------------------
|
||||
3-------------',
|
||||
2 /,2x,'Root',i4,1x,a7,1x,a4,f22.9,' a.u.',f22.4,' eV ',/,2x,
|
||||
1 '---------------------------------------------------------------
|
||||
3-------------')
|
||||
9200 format(5x,'Transition Moments X',f9.5,' Y',f9.5,' Z',f9.5
|
||||
1 ,/,5x,'Transition Moments XX',f9.5,' XY',f9.5,' XZ',f9.5
|
||||
2 ,/,5x,'Transition Moments YY',f9.5,' YZ',f9.5,' ZZ',f9.5
|
||||
5 ,/,5x,'Dipole Oscillator Strength',18x,f14.10
|
||||
5 ,/,5x,'Electric Quadrupole ',18x,f14.10
|
||||
5 ,/,5x,'Magnetic Dipole ',18x,f14.10
|
||||
5 ,/,5x,'Total Oscillator Strength ',18x,f14.10)
|
||||
9210 format(5x,'Transition Moments Spin forbidden'
|
||||
1 ,/,5x,'Oscillator Strength Spin forbidden')
|
||||
9120 format(5x,'Occ.',i5,2x,a4,'--- Virt.',i5,2x,a4,f12.5)
|
||||
9130 format(5x,'Occ.',i5,2x,a4,'--- Virt.',i5,2x,a4,f8.5,' X')
|
||||
9140 format(5x,'Occ.',i5,2x,a4,'--- Virt.',i5,2x,a4,f8.5,' Y')
|
||||
9150 format(5x,'Occ.',i5,1x,a5,1x,a4,'--- Virt.',
|
||||
1 i5,1x,a5,1x,a4,f12.5)
|
||||
9160 format(5x,'Occ.',i5,1x,a5,1x,a4,'--- Virt.',
|
||||
1 i5,1x,a5,1x,a4,f8.5,' X')
|
||||
9170 format(5x,'Occ.',i5,1x,a5,1x,a4,'--- Virt.',
|
||||
1 i5,1x,a5,1x,a4,f8.5,' Y')
|
||||
9180 format(2x,' <S2> = ',f8.4)
|
||||
9190 format(2x,' D<S2> = ',f8.4)
|
||||
9300 format(5x,' sigma_e = ',f8.4, ' Angstrom')
|
||||
9310 format(5x,' sigma_h = ',f8.4, ' Angstrom')
|
||||
9320 format(5x,' |d_h->e| = ',f8.4, ' Angstrom')
|
||||
9330 format(5x,' d_exc = ',f8.4, ' Angstrom')
|
||||
9360 format(5x,' d_CD = ',f8.4, ' Angstrom')
|
||||
9350 format(5x,'Pearson C.C. = ',f8.4)
|
||||
9340 format(5x,'Exciton norm = ',f8.4)
|
||||
|
||||
|
||||
|
||||
end
|
||||
90
src/bsemol/bse_puteris.F
Normal file
90
src/bsemol/bse_puteris.F
Normal file
|
|
@ -0,0 +1,90 @@
|
|||
subroutine bse_puteris(erimo,nmo,nstates,nri,nri_me,rilo,ilower,
|
||||
$ iupper,jlower,jupper,eri,blk,polelo,
|
||||
$ polehi,mynpoles)
|
||||
implicit none
|
||||
#include "mafdecls.fh"
|
||||
#include "errquit.fh"
|
||||
#include "global.fh"
|
||||
|
||||
character(len=2),intent(in) :: blk
|
||||
integer,intent(in) :: nmo, nstates, nri, nri_me, rilo
|
||||
integer,intent(in) :: ilower, iupper, jlower, jupper
|
||||
integer,intent(out) :: eri,polelo,polehi,mynpoles
|
||||
double precision,intent(in),dimension(nstates,nmo,nri_me) :: erimo
|
||||
|
||||
character(len=12),parameter :: pname = 'gw_puteris: '
|
||||
character(len=6) :: erilabel,idxlabel
|
||||
|
||||
integer ilo,ihi,jlo,jhi,imo,jmo,ipair,npairs,me,rihi
|
||||
integer ni,nj,l_tmp,k_tmp,k_loc,ld
|
||||
integer l_pair,k_pair,mynpairs,jpair,iproc,nprocs,ulpair,llpair
|
||||
integer ibatch,nbatch,myllpair,myulpair,ma_dbl
|
||||
integer lmap, kmap
|
||||
|
||||
logical ok
|
||||
integer,external :: gw_get_pair
|
||||
|
||||
integer l_oo,k_oo,l_ov,k_ov,l_vv,k_vv
|
||||
double precision temp(nri)
|
||||
|
||||
ni = iupper - ilower + 1
|
||||
if (ni.lt.1) return
|
||||
|
||||
nj = jupper - jlower + 1
|
||||
if (nj.lt.1) return
|
||||
|
||||
rihi = rilo + nri_me - 1
|
||||
me = ga_nodeid()
|
||||
nprocs = ga_nnodes()
|
||||
|
||||
erilabel = 'eri '//blk
|
||||
idxlabel = 'idx '//blk
|
||||
|
||||
npairs = ni*nj
|
||||
mynpairs = npairs/nprocs + 1
|
||||
|
||||
if (.not.ga_create(mt_dbl,nri,npairs,erilabel,nri,0,eri))
|
||||
$ call errquit(pname//'can''t create '//blk//' block',0,GA_ERR)
|
||||
|
||||
ma_dbl = ma_inquire_stack(mt_dbl)/nri_me
|
||||
nbatch = mynpairs/ma_dbl + 1
|
||||
mynpairs = npairs/nbatch
|
||||
if (mod(npairs,nbatch).ne.0) mynpairs = mynpairs + 1
|
||||
mynpairs = min(npairs,mynpairs)
|
||||
|
||||
if (.not.ma_push_get(mt_dbl,mynpairs*nri_me,'temp',l_pair,k_pair))
|
||||
$ call errquit(pname//'can''t create temporary array',0,MA_ERR)
|
||||
|
||||
ulpair = 0
|
||||
do iproc=0,nprocs-1
|
||||
llpair = ulpair + 1
|
||||
ulpair = ulpair + npairs/nprocs
|
||||
if (iproc.lt.mod(npairs,nprocs)) ulpair = ulpair + 1
|
||||
if (llpair.gt.ulpair) goto 101
|
||||
do ibatch=1,nbatch
|
||||
myllpair = (ibatch-1)*mynpairs + llpair
|
||||
myulpair = min(myllpair + mynpairs - 1,ulpair)
|
||||
do ipair=myllpair,myulpair
|
||||
jmo = (ipair-1)/ni + 1
|
||||
imo = ipair - (jmo-1)*ni + ilower - 1
|
||||
jmo = jmo + jlower - 1
|
||||
call dcopy(nri_me,erimo(imo,jmo,1),nstates*nmo,
|
||||
$ dbl_mb(k_pair+(ipair-myllpair)*nri_me),1)
|
||||
enddo
|
||||
call ga_put(eri,rilo,rihi,myllpair,myulpair,dbl_mb(k_pair),
|
||||
$ nri_me)
|
||||
enddo
|
||||
101 continue
|
||||
call ga_sync()
|
||||
enddo
|
||||
|
||||
ok = ma_chop_stack(l_pair)
|
||||
|
||||
call ga_distribution(eri,me,ld,ld,polelo,polehi)
|
||||
|
||||
if (blk .eq. 'ov') then
|
||||
mynpoles = polehi - polelo + 1
|
||||
endif
|
||||
|
||||
end subroutine
|
||||
|
||||
327
src/bsemol/bse_ri_init.F
Normal file
327
src/bsemol/bse_ri_init.F
Normal file
|
|
@ -0,0 +1,327 @@
|
|||
subroutine bse_ri_init(pars)
|
||||
c
|
||||
implicit none
|
||||
#include "errquit.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "cdft.fh"
|
||||
#include "bas.fh"
|
||||
#include "bse.fh"
|
||||
#include "stdio.fh"
|
||||
#include "util.fh"
|
||||
#include "global.fh"
|
||||
#ifdef SCALAPACK
|
||||
#include "scaleig.fh"
|
||||
#endif
|
||||
c
|
||||
type(bse_params_t) :: pars
|
||||
c
|
||||
character(len=*), parameter :: pname = "bse_ri_init: "
|
||||
c
|
||||
logical iolgc
|
||||
integer g_2ceris, g_2cinv, info, lmiss
|
||||
c
|
||||
integer ilo,ihi,jlo,jhi
|
||||
integer lbuf,lscr,ltmp,leri,lvecs
|
||||
integer l_buf,k_buf,l_scr,k_scr,l_tmp,k_tmp,k_eri,k_mo,l_mo
|
||||
integer l_vecs,k_vecs,l_idx,k_idx
|
||||
integer l_shlbuf, k_shlbuf, l_mobuf, k_mobuf
|
||||
integer me, isp, addr
|
||||
integer g_3ceris,shlo,shhi,offset,ld,g_tmp
|
||||
integer nmo,nvir,rihi,rilo,nri,nri_me,nocc,offk
|
||||
integer ish,jsh,ksh,ni,nj,nk,nij
|
||||
integer ifirst,jfirst,kfirst
|
||||
integer ilast,jlast,klast,imaxthreads
|
||||
integer aopairs,mopairs(2),idum,totmopairs
|
||||
logical IeqJ,ok,oactive
|
||||
double precision eri_est,timer,timing
|
||||
c
|
||||
double precision schwarz_shell
|
||||
external schwarz_shell
|
||||
c
|
||||
#ifdef USE_OPENMP
|
||||
integer,external :: omp_get_max_threads
|
||||
#endif
|
||||
|
||||
timer = util_wallsec()
|
||||
c
|
||||
c Create Global Arrays
|
||||
c
|
||||
if (.not.ga_create(mt_dbl,nbf_cd,nbf_cd,'2c ERI',
|
||||
$ 0, nbf_cd, g_2ceris))
|
||||
$ call errquit(pname//'Error creating 2c ERI',0,GA_ERR)
|
||||
if (.not.ga_create(mt_dbl,nbf_cd,nbf_cd,'2c INV',
|
||||
$ 0, nbf_cd, g_2cinv))
|
||||
$ call errquit(pname//'Error creating 2c ERI',0,GA_ERR)
|
||||
call ga_zero(g_2ceris)
|
||||
c
|
||||
c Compute 2-center ERIs
|
||||
c
|
||||
if (pars%me.eq.0) write(luout,9000)
|
||||
9000 format(/,10x,'Computing 2-center integrals')
|
||||
c
|
||||
call dft_get2eri(cd_bas_han, g_2ceris, .false.)
|
||||
c
|
||||
c Obtain inverse Cholesky factor
|
||||
c
|
||||
if (pars%me.eq.0) write(luout,9010)
|
||||
9010 format(10x,'Computing Inverse Cholesky factor')
|
||||
c
|
||||
call ga_sync()
|
||||
call ga_chol_inv_seq(g_2ceris, g_2cinv, info)
|
||||
if (info.ne.0) then
|
||||
if (pars%me.eq.0) then
|
||||
write(luout,*)
|
||||
write(luout,*)' Problem in performing Cholesky'
|
||||
write(luout,*)' Obtaining square root via truncated '
|
||||
write(luout,*)' eigenvalue decomposition '
|
||||
write(luout,*)
|
||||
endif
|
||||
call dft_invdiag(g_2ceris, g_2cinv, nbf_cd)
|
||||
endif
|
||||
c
|
||||
if (.not. ga_destroy(g_2ceris))
|
||||
$ call errquit(pname//'failed to destrou g_2ceri',0,GA_ERR)
|
||||
c
|
||||
c Allocate storage
|
||||
c
|
||||
nri = nbf_cd
|
||||
nmo = pars%nmo
|
||||
aopairs = (nbf_ao*(nbf_ao+1))/2
|
||||
|
||||
do isp=1,pars%ipol
|
||||
mopairs(isp) = pars%nmo*pars%nmo
|
||||
enddo
|
||||
totmopairs = sum(mopairs(1:pars%ipol))
|
||||
|
||||
if (.not.ga_create(mt_dbl,1,nri,'3c ERI',aopairs,0,
|
||||
$ g_3ceris))
|
||||
$ call errquit(pname//'can''t create 3c ERIs',0,GA_ERR)
|
||||
call ga_distribution(g_3ceris,pars%me,ilo,ihi,rilo,rihi)
|
||||
nri_me = rihi - rilo + 1
|
||||
ok = ga_destroy(g_3ceris)
|
||||
|
||||
c call ga_access(g_3ceris,ilo,ihi,rilo,rihi,k_eri,ld)
|
||||
c call dfill(aopairs*nri_me,0d0,dbl_mb(k_eri),1)
|
||||
|
||||
shlo = 0
|
||||
do ksh=1,nshells_cd
|
||||
ok = bas_cn2bfr(cd_bas_han,ksh,kfirst,klast)
|
||||
if (klast.lt.rilo) cycle
|
||||
if (shlo.eq.0) then
|
||||
offset = (rilo-kfirst)
|
||||
shlo = ksh
|
||||
endif
|
||||
shhi = ksh
|
||||
if (klast.ge.rihi) exit
|
||||
enddo
|
||||
|
||||
oactive = nri_me.gt.0
|
||||
|
||||
call int_mem_2e3c(lbuf, lscr)
|
||||
if (.not.ma_alloc_get(mt_dbl,lbuf, 'RI buffer',l_buf, k_buf))
|
||||
$ call errquit(pname//'can''t get buffer space',0,MA_ERR)
|
||||
lscr = max(lscr,nbf_cd*nbf_ao_mxnbf_cn**2)
|
||||
if (.not.ma_alloc_get(mt_dbl,lscr, 'RI scratch', l_scr, k_scr))
|
||||
$ call errquit(pname//'can''t get scratch space',0,MA_ERR)
|
||||
|
||||
if (.not.ma_alloc_get(mt_dbl,nri_me*nbf_ao_mxnbf_cn**2,
|
||||
$ 'Shell buffer',l_shlbuf, k_shlbuf))
|
||||
$ call errquit(pname//'can''t allocate shell buffer',0,MA_ERR)
|
||||
if (.not.ma_alloc_get(mt_dbl,
|
||||
$ nri_me*nbf_ao_mxnbf_cn*nmo*pars%ipol,
|
||||
$ 'MO buffer',l_mobuf, k_mobuf))
|
||||
$ call errquit(pname//'can''t allocate shell buffer',0,MA_ERR)
|
||||
c
|
||||
c Compute three-center ERIs in MO representation
|
||||
c
|
||||
call ga_sync()
|
||||
c
|
||||
if (pars%me.eq.0) write(luout,9020)
|
||||
9020 format(10x,'Computing 3-center integrals')
|
||||
|
||||
lvecs = nbf_ao*nmo*pars%ipol
|
||||
if (.not.ma_alloc_get(mt_dbl,lvecs,'MOVECS',l_vecs,k_vecs))
|
||||
$ call errquit(pname//'can''t get movecs space',0,MA_ERR)
|
||||
if (.not.ma_alloc_get(mt_dbl,totmopairs*nri_me,'MO Eris',
|
||||
$ l_mo,k_mo))
|
||||
$ call errquit(pname//'can''t get MO eris space',0,MA_ERR)
|
||||
call dfill(totmopairs*nri_me,0.0d0,dbl_mb(k_mo),1)
|
||||
|
||||
do isp=1,pars%ipol
|
||||
call ga_get(pars%g_movecs(isp),1,nbf_ao,1,nmo,
|
||||
$ dbl_mb(k_vecs+(isp-1)*nbf_ao*nmo),nbf_ao)
|
||||
enddo
|
||||
c
|
||||
#ifdef USE_OPENMP
|
||||
iMaxThreads = omp_get_max_threads()
|
||||
call util_blas_set_num_threads(iMaxThreads)
|
||||
#endif
|
||||
|
||||
do ish=1,nshells_ao
|
||||
ok = bas_cn2bfr(ao_bas_han,ish,ifirst,ilast)
|
||||
ni = ilast - ifirst + 1
|
||||
call dfill(ni*nmo*pars%ipol*nri_me,0d0,dbl_mb(k_mobuf),1)
|
||||
do jsh=1,nshells_ao
|
||||
eri_est = schwarz_shell(ish,jsh)
|
||||
if (eri_est.lt.pars%tol2e) cycle
|
||||
IeqJ = ish.eq.jsh
|
||||
ok = bas_cn2bfr(ao_bas_han,jsh,jfirst,jlast)
|
||||
nj = jlast - jfirst + 1
|
||||
nij = ni*nj
|
||||
do ksh=shlo,shhi
|
||||
ok = bas_cn2bfr(cd_bas_han, ksh, kfirst, klast)
|
||||
call int_2e3c(cd_bas_han,ksh,ao_bas_han,ish,jsh,lscr,
|
||||
$ dbl_mb(k_scr),lbuf,dbl_mb(k_buf))
|
||||
|
||||
if (ksh.eq.shlo) then
|
||||
kfirst = rilo
|
||||
offk = k_buf + offset*nij
|
||||
else
|
||||
offk = k_buf
|
||||
endif
|
||||
|
||||
if (ksh.eq.shhi) klast = rihi
|
||||
|
||||
nk = klast - kfirst + 1
|
||||
kfirst = kfirst - rilo + 1
|
||||
call dcopy(ni*nj*nk,dbl_mb(offk),1,
|
||||
$ dbl_mb(k_shlbuf+(kfirst-1)*ni*nj),1)
|
||||
enddo
|
||||
do isp=1,pars%ipol
|
||||
call gw_mo2(dbl_mb(k_shlbuf),
|
||||
$ dbl_mb(k_mobuf+(isp-1)*nmo*ni*nri_me),
|
||||
$ dbl_mb(k_vecs+(isp-1)*nbf_ao*nmo),nri_me,
|
||||
$ nbf_ao,ni,nj,nmo,nmo,ifirst,
|
||||
$ jfirst,ieqj)
|
||||
enddo
|
||||
enddo
|
||||
do isp=1,pars%ipol
|
||||
call gw_mo(dbl_mb(k_mobuf+(isp-1)*nmo*ni*nri_me),
|
||||
$ dbl_mb(k_mo+(isp-1)*mopairs(1)*nri_me),
|
||||
$ dbl_mb(k_vecs+(isp-1)*nbf_ao*nmo),nri_me,
|
||||
$ nbf_ao,ni,nmo,nmo,ifirst)
|
||||
enddo
|
||||
enddo
|
||||
c
|
||||
ok = ma_free_heap(l_buf) .and. ma_free_heap(l_scr) .and.
|
||||
$ ma_free_heap(l_shlbuf) .and.
|
||||
$ ma_free_heap(l_mobuf)
|
||||
if (.not.ok)
|
||||
$ call errquit(pname//'can''t free heap',165,MA_ERR)
|
||||
c
|
||||
c
|
||||
#ifdef GWDEBUG
|
||||
if (pars%me.eq.0) write(*,*) ' Transform ERIs'
|
||||
#endif
|
||||
|
||||
call ga_sync()
|
||||
|
||||
ok = ma_free_heap(l_vecs)
|
||||
if (.not.ok)
|
||||
$ call errquit(pname//'can''t free heap',176,MA_ERR)
|
||||
c
|
||||
call ga_sync()
|
||||
c
|
||||
c Distribute and orthonormalize ERIs
|
||||
c
|
||||
timer = util_wallsec()
|
||||
#ifdef GWDEBUG
|
||||
if (pars%me.eq.0) write(*,*) ' Distribute ERIs'
|
||||
#endif
|
||||
|
||||
addr = k_mo
|
||||
do isp=1,pars%ipol
|
||||
if (nmo.eq.0) cycle
|
||||
if (isp.eq.2) addr = k_mo + mopairs(1)*nri_me
|
||||
|
||||
! OO block
|
||||
call bse_puteris(dbl_mb(addr),nmo,nmo,nri,nri_me,
|
||||
$ rilo,1,pars%nocc(isp),1,pars%nocc(isp),
|
||||
$ pars%g_erioo(isp),'oo',pars%oolo(isp),
|
||||
$ pars%oohi(isp),idum)
|
||||
|
||||
! OV block
|
||||
call bse_puteris(dbl_mb(addr),nmo,nmo,nri,nri_me,
|
||||
$ rilo,pars%nocc(isp)+1,nmo,1,pars%nocc(isp),
|
||||
$ pars%g_eriov(isp),'ov',pars%ovlo(isp),
|
||||
$ pars%ovhi(isp),pars%mynpoles(isp))
|
||||
|
||||
! VV block
|
||||
call bse_puteris(dbl_mb(addr),nmo,nmo,nri,nri_me,
|
||||
$ rilo,pars%nocc(isp)+1,nmo,
|
||||
$ pars%nocc(isp)+1,nmo,pars%g_erivv(isp),
|
||||
$ 'vv',pars%vvlo(isp),pars%vvhi(isp),idum)
|
||||
|
||||
enddo
|
||||
|
||||
ok = ma_free_heap(l_mo)
|
||||
#ifdef GWDEBUG
|
||||
if (pars%me.eq.0) write(*,*) ' Read inverse'
|
||||
#endif
|
||||
c
|
||||
if (.not.ma_alloc_get(mt_dbl,nri**2,'RI temporary',l_tmp,k_tmp))
|
||||
$ call errquit(pname//'can''t get temp space',0,MA_ERR)
|
||||
if (pars%me.eq.0) then
|
||||
call ga_get(g_2cinv,1,nbf_cd,1,nbf_cd,dbl_mb(k_tmp),nbf_cd)
|
||||
endif
|
||||
call ga_sync()
|
||||
if (.not. ga_destroy(g_2cinv))
|
||||
$ call errquit(pname//'failed to destroy g_2cinv',0,GA_ERR)
|
||||
call ga_brdcst(1038,dbl_mb(k_tmp),ma_sizeof(mt_dbl,nri**2,
|
||||
$ mt_byte),0)
|
||||
|
||||
#ifdef GWDEBUG
|
||||
if (pars%me.eq.0) write(*,*) ' Orthogonalize ERIs'
|
||||
#endif
|
||||
do isp=1,pars%ipol
|
||||
c
|
||||
c Transform OO block
|
||||
c
|
||||
call ga_distribution(pars%g_erioo(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
if (jlo.gt.jhi) goto 101
|
||||
call ga_access(pars%g_erioo(isp),ilo,ihi,jlo,jhi,k_eri,ld)
|
||||
nij = jhi-jlo+1
|
||||
call dtrmm('l','l','n','n',nbf_cd,nij,1.0d0,dbl_mb(k_tmp),
|
||||
$ nbf_cd,dbl_mb(k_eri),nbf_cd)
|
||||
call ga_release_update(pars%g_erioo(isp),ilo,ihi,jlo,jhi)
|
||||
c
|
||||
c Transform OV block
|
||||
c
|
||||
101 continue
|
||||
call ga_distribution(pars%g_eriov(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
if (jlo.gt.jhi) goto 102
|
||||
call ga_access(pars%g_eriov(isp),ilo,ihi,jlo,jhi,k_eri,ld)
|
||||
nij = jhi-jlo+1
|
||||
call dtrmm('l','l','n','n',nbf_cd,nij,1.0d0,dbl_mb(k_tmp),
|
||||
$ nbf_cd,dbl_mb(k_eri),nbf_cd)
|
||||
call ga_release_update(pars%g_eriov(isp),ilo,ihi,jlo,jhi)
|
||||
c
|
||||
c Transform VV block
|
||||
c
|
||||
102 continue
|
||||
if (nmo.gt.pars%nocc(isp)) then
|
||||
call ga_distribution(pars%g_erivv(isp),pars%me,ilo,ihi,
|
||||
$ jlo,jhi)
|
||||
if (jlo.gt.jhi) goto 103
|
||||
call ga_access(pars%g_erivv(isp),ilo,ihi,jlo,jhi,k_eri,ld)
|
||||
nij = jhi-jlo+1
|
||||
call dtrmm('l','l','n','n',nbf_cd,nij,1.0d0,dbl_mb(k_tmp),
|
||||
$ nbf_cd,dbl_mb(k_eri),nbf_cd)
|
||||
call ga_release_update(pars%g_erivv(isp),ilo,ihi,jlo,jhi)
|
||||
endif
|
||||
|
||||
|
||||
103 continue
|
||||
enddo
|
||||
c
|
||||
c Free heap
|
||||
c
|
||||
ok = ma_free_heap(l_tmp)
|
||||
if (.not.ok)
|
||||
$ call errquit(pname//'failed to free heap',224,MA_ERR)
|
||||
c
|
||||
#ifdef USE_OPENMP
|
||||
call util_blas_set_num_threads(1)
|
||||
#endif
|
||||
return
|
||||
end subroutine
|
||||
99
src/bsemol/bse_s2.F
Normal file
99
src/bsemol/bse_s2.F
Normal file
|
|
@ -0,0 +1,99 @@
|
|||
subroutine bse_s2(nocc, nmo, npoles, xup, xdw, yup, ydw,
|
||||
$ ovlp, s2)
|
||||
implicit none
|
||||
|
||||
integer nocc(2), npoles(2), nmo
|
||||
double precision xup(npoles(1)), xdw(npoles(2))
|
||||
double precision yup(npoles(1)), ydw(npoles(2))
|
||||
double precision s2
|
||||
double precision ovlp(nmo,nmo)
|
||||
|
||||
double precision ov,xx,yy
|
||||
integer i,j,a,b,ipole,jpole,nvir(2)
|
||||
integer llipole,ulipole,lljpole,uljpole
|
||||
|
||||
double precision, external :: ddot
|
||||
|
||||
s2 = 0d0
|
||||
nvir(1) = nmo - nocc(1)
|
||||
nvir(2) = nmo - nocc(2)
|
||||
|
||||
! Up-Up
|
||||
do i=1,nocc(1)
|
||||
do j=1,nocc(1)
|
||||
llipole = (i-1)*nvir(1) + 1
|
||||
ulipole = i*nvir(1)
|
||||
lljpole = (j-1)*nvir(1) + 1
|
||||
uljpole = j*nvir(1)
|
||||
ov = dot_product(ovlp(i,1:nocc(2)),ovlp(j,1:nocc(2)))
|
||||
xx = dot_product(xup(llipole:ulipole),xup(lljpole:uljpole))
|
||||
|
||||
yy = dot_product(yup(llipole:ulipole),yup(lljpole:uljpole))
|
||||
s2 = s2 + ov*(xx+yy)
|
||||
enddo
|
||||
enddo
|
||||
do a=nocc(1)+1,nmo
|
||||
do b=nocc(1)+1,nmo
|
||||
ov = dot_product(ovlp(a,1:nocc(2)),ovlp(b,1:nocc(2)))
|
||||
llipole = a - nocc(1)
|
||||
lljpole = b - nocc(1)
|
||||
s2 = s2 - ov*ddot(nocc(1),xup(llipole),nvir(1),
|
||||
$ xup(lljpole),nvir(1))
|
||||
s2 = s2 - ov*ddot(nocc(1),yup(llipole),nvir(1),
|
||||
$ yup(lljpole),nvir(1))
|
||||
enddo
|
||||
enddo
|
||||
|
||||
! Dw-Dw
|
||||
do i=1,nocc(2)
|
||||
do j=1,nocc(2)
|
||||
llipole = (i-1)*nvir(2) + 1
|
||||
ulipole = i*nvir(2)
|
||||
lljpole = (j-1)*nvir(2) + 1
|
||||
uljpole = j*nvir(2)
|
||||
ov = dot_product(ovlp(1:nocc(1),i),ovlp(1:nocc(1),j))
|
||||
xx = dot_product(xdw(llipole:ulipole),xdw(lljpole:uljpole))
|
||||
yy = dot_product(ydw(llipole:ulipole),ydw(lljpole:uljpole))
|
||||
s2 = s2 + ov*(xx+yy)
|
||||
enddo
|
||||
enddo
|
||||
do a=nocc(2)+1,nmo
|
||||
do b=nocc(2)+1,nmo
|
||||
ov = dot_product(ovlp(1:nocc(1),a),ovlp(1:nocc(1),b))
|
||||
llipole = a - nocc(2)
|
||||
lljpole = b - nocc(2)
|
||||
s2 = s2 - ov*ddot(nocc(2),xdw(llipole),nvir(2),
|
||||
$ xdw(lljpole),nvir(2))
|
||||
s2 = s2 - ov*ddot(nocc(2),ydw(llipole),nvir(2),
|
||||
$ ydw(lljpole),nvir(2))
|
||||
enddo
|
||||
enddo
|
||||
|
||||
! Up-Dw
|
||||
do ipole=1,npoles(1)
|
||||
i = ipole/nvir(1) + 1
|
||||
a = ipole - (i-1)*nvir(1) + nocc(1)
|
||||
do jpole=1,npoles(2)
|
||||
j = jpole/nvir(2) + 1
|
||||
b = jpole - (j-1)*nvir(2) + nocc(2)
|
||||
s2 = s2 -
|
||||
$ 2d0*xup(ipole)*xdw(jpole)*ovlp(i,j)*ovlp(a,b) -
|
||||
$ 2d0*yup(ipole)*ydw(jpole)*ovlp(i,j)*ovlp(a,b) +
|
||||
$ 2d0*xup(ipole)*ydw(jpole)*ovlp(a,j)*ovlp(i,b)
|
||||
enddo
|
||||
enddo
|
||||
|
||||
! Dw-Up
|
||||
do ipole=1,npoles(2)
|
||||
i = ipole/nvir(2) + 1
|
||||
a = ipole - (i-1)*nvir(2) + nocc(2)
|
||||
do jpole=1,npoles(1)
|
||||
j = jpole/nvir(1) + 1
|
||||
b = jpole - (j-1)*nvir(1) + nocc(1)
|
||||
s2 = s2 +
|
||||
$ 2d0*xdw(ipole)*yup(jpole)*ovlp(j,a)*ovlp(b,i)
|
||||
enddo
|
||||
enddo
|
||||
|
||||
end subroutine
|
||||
|
||||
54
src/bsemol/bse_wmn.F
Normal file
54
src/bsemol/bse_wmn.F
Normal file
|
|
@ -0,0 +1,54 @@
|
|||
subroutine bse_wmn(pars,w,nmo,nri,ipol)
|
||||
implicit none
|
||||
#include "errquit.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "global.fh"
|
||||
#include "bse.fh"
|
||||
|
||||
type(bse_params_t) :: pars
|
||||
integer nmo, nri, ipol
|
||||
|
||||
character(*), parameter :: pname = 'bse_wmn: '
|
||||
|
||||
double precision w(*)
|
||||
|
||||
integer isp,ilo,ihi,jlo,jhi,keri,ld,g_tmp
|
||||
|
||||
do isp=1,ipol
|
||||
|
||||
if (.not.pars%tda) then
|
||||
! copy and transform occ-vir block
|
||||
if(.not.ga_duplicate(pars%g_eriov(isp),pars%g_wov(isp),'Wov'))
|
||||
& call errquit(pname//'could not duplicate ERI OV',0,GA_ERR)
|
||||
call ga_copy(pars%g_eriov(isp),pars%g_wov(isp))
|
||||
if(pars%mynpoles(isp).eq.0) cycle
|
||||
call ga_access(pars%g_wov(isp),1,nri,pars%ovlo(isp),
|
||||
& pars%ovhi(isp),keri,ld)
|
||||
call dtfsm('n','l','l','n','n',nri,pars%mynpoles(isp),1d0,
|
||||
& w,dbl_mb(keri),nri)
|
||||
call ga_release_update(pars%g_wov(isp),1,nri,pars%ovlo(isp),
|
||||
& pars%ovhi(isp))
|
||||
endif
|
||||
|
||||
! transform occ-occ and vir-vir blocks in place
|
||||
call ga_distribution(pars%g_erioo(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
if (jlo.gt.jhi) goto 101
|
||||
call ga_access(pars%g_erioo(isp),ilo,ihi,jlo,jhi,keri,ld)
|
||||
call dtfsm('n','l','l','n','n',nri,jhi-jlo+1,1.0d0,w,
|
||||
& dbl_mb(keri),nri)
|
||||
call ga_release_update(pars%g_erioo(isp),1,nri,jlo,jhi)
|
||||
|
||||
101 continue
|
||||
call ga_distribution(pars%g_erivv(isp),pars%me,ilo,ihi,jlo,jhi)
|
||||
if (jlo.gt.jhi) cycle
|
||||
call ga_access(pars%g_erivv(isp),ilo,ihi,jlo,jhi,keri,ld)
|
||||
call dtfsm('n','l','l','n','n',nri,jhi-jlo+1,1.0d0,w,
|
||||
& dbl_mb(keri),nri)
|
||||
call ga_release_update(pars%g_erivv(isp),1,nri,jlo,jhi)
|
||||
|
||||
|
||||
|
||||
enddo
|
||||
|
||||
|
||||
end
|
||||
75
src/bsemol/bsemol.F
Normal file
75
src/bsemol/bsemol.F
Normal file
|
|
@ -0,0 +1,75 @@
|
|||
logical function bsemol(rtdb)
|
||||
implicit none
|
||||
#include "rtdb.fh"
|
||||
integer rtdb
|
||||
c
|
||||
logical bse_doit
|
||||
external bse doit
|
||||
c
|
||||
logical status
|
||||
c
|
||||
status = rtdb_cput(rtdb,'dft:theory', 1, 'dft')
|
||||
if (status) status = bse_doit(rtdb)
|
||||
bsemol = status
|
||||
call grid_cleanup(.false.)
|
||||
c
|
||||
end function
|
||||
c
|
||||
c -----------------------------------------------------------------
|
||||
c
|
||||
logical function bse_doit(rtdb)
|
||||
implicit none
|
||||
#include "errquit.fh"
|
||||
#include "global.fh"
|
||||
#include "rtdb.fh"
|
||||
#include "stdio.fh"
|
||||
#include "util.fh"
|
||||
#include "bse.fh"
|
||||
|
||||
integer, intent(in) :: rtdb
|
||||
character(*), parameter :: pname = "bse_doit: "
|
||||
|
||||
type(bse_params_t) :: params
|
||||
character*80 theory
|
||||
integer me
|
||||
c
|
||||
c -----------------------------------------------------------------
|
||||
c
|
||||
c Initialize
|
||||
c
|
||||
bse_doit = .false.
|
||||
c
|
||||
c Determine the theory we are using
|
||||
c
|
||||
if (.not. rtdb_cget(rtdb, "dft:theory", 1, theory))
|
||||
$ call errquit(pname//"theory not specified",0,RTDB_ERR)
|
||||
c
|
||||
c Initialize GW calculation
|
||||
c
|
||||
params%rtdb = rtdb
|
||||
params%me = ga_nodeid()
|
||||
c
|
||||
call bse_init(params)
|
||||
c
|
||||
if (theory.eq."dft") then
|
||||
|
||||
if (params%davidson) then
|
||||
call bse_davidson(params)
|
||||
elseif (params%lanczos) then
|
||||
call bse_lanczos(params)
|
||||
else
|
||||
call bse_analytic(params)
|
||||
endif
|
||||
c
|
||||
elseif (theory.eq.'sodft') then
|
||||
call errquit(pname//"spin-orbit not implemented",0,0)
|
||||
else
|
||||
call errquit(pname//"invalid theory: "//trim(theory),0,0)
|
||||
endif
|
||||
c
|
||||
call bse_finalize(params)
|
||||
bse_doit = .true.
|
||||
c
|
||||
c -----------------------------------------------------------------
|
||||
c
|
||||
end function
|
||||
7
src/bsemol/scaleig.fh
Normal file
7
src/bsemol/scaleig.fh
Normal file
|
|
@ -0,0 +1,7 @@
|
|||
c Copy of DFT scaleig
|
||||
c
|
||||
c $Id$
|
||||
c
|
||||
character*1 yev
|
||||
integer nb
|
||||
common /dft_scaleigp/ nb, yev
|
||||
|
|
@ -21,7 +21,7 @@ KNOWN_MODULE_SUBDIRS = NWints atomscf ddscf develop gradients moints nwdft \
|
|||
property solvation nwpw fft rimp2_grad python argos analyz diana nwmd cafe space \
|
||||
drdy vscf uccsdt qmmm rism qmd etrans tce geninterface bq mm cons\
|
||||
perfm dntmc smd nbo dangchang leps ccca lucia dimqm rdmft fcidump\
|
||||
gwmol xtb nwc_columbus
|
||||
gwmol xtb nwc_columbus bsemol
|
||||
|
||||
# These are the libraries for the high-level modules. They should be
|
||||
# specified in an order that will link correctly, but that shouldn't
|
||||
|
|
@ -36,7 +36,7 @@ KNOWN_MODULE_LIBS = -ltest -lccsd -llucia -lmcscf -lselci -lmp2 -lmoints \
|
|||
-lcafe -lspace -lanalyze -ldiana -lqhop \
|
||||
-lpfft -ldplot -lnwpython -ldrdy -lvscf -lqmmm -lrism -lqmd -letrans\
|
||||
-lpspw -ltce -lgeninterface -lbq -lmm -lcons -lperfm -lnbo -ldntmc -lccca\
|
||||
-ldimqm -lrdmft -lfcidump -lgwmol -lnwxtb
|
||||
-ldimqm -lrdmft -lfcidump -lgwmol -lnwxtb -lbsemol
|
||||
|
||||
# This include file handles configuration of the NW_MODULE_SUBDIRS and
|
||||
# NW_MODULE_LIBS macros for what we actually want to build. It
|
||||
|
|
@ -77,22 +77,22 @@ EXP_alldev = develop $(EXP_all)
|
|||
EXP_pnnldev = develop $(EXP_pnnl)
|
||||
EXP_qm = NWints atomscf ddscf gradients moints nwdft rimp2 hessian bq cons \
|
||||
optim stepper driver ccsd vib mcscf property solvation dplot mp2_grad esp etrans tce vscf dimqm \
|
||||
gwmol
|
||||
gwmol bsemol
|
||||
EXP_smallqm = NWints atomscf ddscf gradients moints nwdft rimp2 hessian bq cons \
|
||||
optim stepper driver vib qmd mcscf property solvation dplot \
|
||||
mp2_grad esp etrans vscf nwmd dimqm gwmol
|
||||
mp2_grad esp etrans vscf nwmd dimqm gwmol bsemol
|
||||
EXP_smallqmmm = NWints atomscf ddscf gradients moints nwdft rimp2 hessian bq cons \
|
||||
optim stepper driver vib qmd mcscf property solvation dplot \
|
||||
mp2_grad esp etrans vscf nwmd qmmm prepar analyz dimqm gwmol
|
||||
mp2_grad esp etrans vscf nwmd qmmm prepar analyz dimqm gwmol bsemol
|
||||
EXP_notce = NWints atomscf ddscf gradients moints nwdft rimp2 hessian bq cons \
|
||||
optim stepper driver ccsd vib mcscf property solvation dplot mp2_grad esp etrans vscf dimqm \
|
||||
gwmol
|
||||
gwmol bsemol
|
||||
EXP_tinyqmpw = nwdft property solvation driver ccsd mp2_grad hessian vib nwpw
|
||||
EXP_ver2 = prepar gradients mp2_grad stepper vib mcscf \
|
||||
optim driver analyz diana
|
||||
#plane_wave ... broken? RJH
|
||||
EXP_qmandpw = nwpw NWints atomscf ddscf gradients moints nwdft rimp2 hessian bq cons \
|
||||
optim stepper driver ccsd vib mcscf property solvation dplot mp2_grad esp etrans vscf dimqm gwmol
|
||||
optim stepper driver ccsd vib mcscf property solvation dplot mp2_grad esp etrans vscf dimqm gwmol bsemol
|
||||
EXP_tps = prepar nwmd analyz diana
|
||||
EXP_md = prepar nwmd analyz ddscf nwdft gradients esp
|
||||
EXP_classical = prepar nwmd analyz
|
||||
|
|
@ -187,6 +187,7 @@ MODULES := $(patsubst nwpw,nwpw fft,$(MODULES) )
|
|||
MODULES := $(patsubst ddscf,ddscf atomscf nwdft NWints gradients,$(MODULES) )
|
||||
|
||||
MODULES := $(patsubst gwmol, gwmol nwdft ddscf gradients hessian atomscf NWints,$(MODULES) )
|
||||
MODULES := $(patsubst bsemol, bsemol gwmol nwdft ddscf gradients hessian atomscf NWints,$(MODULES) )
|
||||
|
||||
MODULES := $(patsubst lucia, lucia mp2_grad ddscf nwdft property cphf gradients hessian atomscf NWints moints,$(MODULES) )
|
||||
|
||||
|
|
@ -274,6 +275,7 @@ MOD_LIBS := $(patsubst dimqm,-ldimqm,$(MOD_LIBS) )
|
|||
MOD_LIBS := $(patsubst rdmft,-lrdmft,$(MOD_LIBS) )
|
||||
MOD_LIBS := $(patsubst fcidump,-lfcidump,$(MOD_LIBS) )
|
||||
MOD_LIBS := $(patsubst gwmol,-lgwmol,$(MOD_LIBS) )
|
||||
MOD_LIBS := $(patsubst bsemol,-lbsemol,$(MOD_LIBS) )
|
||||
MOD_LIBS := $(patsubst xtb,-lnwxtb,$(MOD_LIBS) )
|
||||
MOD_LIBS := $(patsubst nwc_columbus,-lnwc_columbus,$(MOD_LIBS) )
|
||||
MOD_LIBS := $(strip $(MOD_LIBS))
|
||||
|
|
@ -339,6 +341,7 @@ STUBROUTINES := $(patsubst solvation, solvation,$(STUBROUTINES))
|
|||
STUBROUTINES := $(patsubst dimqm,dimqm_addop dimqm_addop_uhf_damp dimqm_addop_uhf dimqm_tidy dimqm_input dimpar_input dimqm_main fock_dim dimqm_prep dim_grid_quadv0 dimqm_lclfld dimqm_polarizability dimqm_broadcast dimqm_Dens2Dipoles dimqm_rohf_calcDIM dimqm_rohf_wrap dimqm_used dimqm_setgo dimqm_getgo dimqm_getenergy dimqm_setlfirst dimqm_getlrsp dimqm_lrsp dimqm_getlclfld dimqm_getgdipeli dimqm_energy dimqm_init,$(STUBROUTINES) )
|
||||
#STUBROUTINES := $(patsubst cons, cons_end cons_input cons_add_egrad cons_add_energy cons_init cons_print_summary cons_active,$(STUBROUTINES) )
|
||||
STUBROUTINES := $(patsubst gwmol, gw_input gwmol, $(STUBROUTINES) )
|
||||
STUBROUTINES := $(patsubst bsemol, bse_input bsemol, $(STUBROUTINES) )
|
||||
STUBROUTINES := $(patsubst xtb, xtb_input xtb_energy xtb_energy_gradient, $(STUBROUTINES) )
|
||||
STUBROUTINES := $(patsubst nwc_columbus, sifcfg sifwh sodft_aoints, $(STUBROUTINES) )
|
||||
STUBROUTINES := $(strip $(STUBROUTINES))
|
||||
|
|
|
|||
|
|
@ -308,16 +308,7 @@ c
|
|||
write(LuOut,*) ' B matrix '
|
||||
call output(x, 1, ncart, 1, nzvar, ncart, nzvar, 1)
|
||||
endif
|
||||
call dgemm('t', 'n', nzvar, nzvar, ncart, 1d0,
|
||||
$ x, ncart, x, ncart, 0d0, y, nzvar) ! BT.B -> y ... should symmetrize
|
||||
do i = 1, nzvar
|
||||
do j = 1, nzvar
|
||||
ij = i + (j-1)*nzvar
|
||||
ji = j + (i-1)*nzvar
|
||||
y(ij) = 0.5d0*(y(ij) + y(ji))
|
||||
y(ji) = y(ij)
|
||||
end do
|
||||
end do
|
||||
call dsyrk('u', 't', nzvar, ncart, 1d0, x, ncart, 0d0, y, nzvar)
|
||||
call dsyev('v', 'u', nzvar, y, nzvar, e, z, ndim**2, info)
|
||||
if (info .ne. 0) call errquit('geom_binvr: dsyev failed', info,
|
||||
& MEM_ERR)
|
||||
|
|
|
|||
|
|
@ -12,6 +12,7 @@ c
|
|||
type(gw_params_t) :: pars
|
||||
c
|
||||
character(len=13), parameter :: pname = 'gw_analytic: '
|
||||
character*255 gwevals
|
||||
c
|
||||
integer,dimension(2) :: nocc,nvir,npoles,nstates,mynpoles
|
||||
|
||||
|
|
@ -45,6 +46,10 @@ c
|
|||
integer maxpoles,maxnpoles,isp,lSize
|
||||
|
||||
logical ok
|
||||
|
||||
integer,parameter :: unitno = 66
|
||||
|
||||
logical,external :: movecs_write
|
||||
c
|
||||
c -----------------------------------------------------------------
|
||||
c
|
||||
|
|
@ -89,6 +94,15 @@ c
|
|||
|
||||
call gw_vxc(dbl_mb(k_sigma),nmo,pars)
|
||||
|
||||
if (pars%ncap) then
|
||||
do isp=1,ipol
|
||||
call daxpy(pars%nocc(isp),1d0,pars%vxddp(isp),0,
|
||||
$ dbl_mb(k_sigma+(isp-1)*pars%nmo),1)
|
||||
call daxpy(pars%nvir(isp),1d0,pars%vxddm(isp),0,
|
||||
$ dbl_mb(k_sigma+(isp-1)*pars%nmo+pars%nocc(isp)),1)
|
||||
enddo
|
||||
endif
|
||||
|
||||
pars%timer%vxc = pars%timer%vxc + util_wallsec() - timer
|
||||
c
|
||||
c Obtain Sigma_x
|
||||
|
|
@ -135,6 +149,14 @@ c
|
|||
endif
|
||||
write(luout,*)
|
||||
endif
|
||||
|
||||
call gw_shift(dbl_mb(k_eold),dbl_mb(pars%k_evals),-efermi,
|
||||
$ pars%nmo,nbf_ao,pars%ipol)
|
||||
if(.not.movecs_write(pars%rtdb, ao_bas_han, movecs_out, 'dft',
|
||||
$ 'W vals',nbf_ao,ipol,(/pars%nmo,pars%nmo/),
|
||||
$ dbl_mb(pars%k_occ), nbf_ao,
|
||||
$ dbl_mb(pars%k_evals),nbf_ao,pars%g_movecs))
|
||||
$ call errquit('gw_analytic: could not save movecs',0,DISK_ERR)
|
||||
c
|
||||
c Obtain eigenvalue difference
|
||||
c
|
||||
|
|
@ -185,25 +207,26 @@ c
|
|||
endif
|
||||
|
||||
c
|
||||
if (pars%evgw0.or.pars%evgw) then
|
||||
call gw_scissor(dbl_mb(k_Enew),dbl_mb(k_Eold),pars%nmo,
|
||||
call gw_scissor(dbl_mb(k_Enew),dbl_mb(k_Eold),pars%nmo,
|
||||
$ pars%nocc,pars%nvir,pars%noqp,pars%nvqp,
|
||||
$ pars%ipol,pars%me)
|
||||
|
||||
call ga_brdcst(1038,dbl_mb(k_enew),
|
||||
$ ma_sizeof(mt_dbl,nmo,mt_byte),0)
|
||||
call ga_brdcst(1038,dbl_mb(k_sigma),
|
||||
$ ma_sizeof(mt_dbl,nmo,mt_byte),0)
|
||||
call ga_brdcst(1038,dbl_mb(k_enew),
|
||||
$ ma_sizeof(mt_dbl,pars%ipol*nmo,mt_byte),0)
|
||||
call ga_brdcst(1038,dbl_mb(k_sigma),
|
||||
$ ma_sizeof(mt_dbl,pars%ipol*nmo,mt_byte),0)
|
||||
|
||||
if (pars%evgw0.or.pars%evgw) then
|
||||
|
||||
!
|
||||
! Update Fermi energy
|
||||
!
|
||||
do i=1,ipol
|
||||
lSize = k_enew + (i-1)*nbf_ao
|
||||
lSize = k_enew + (i-1)*nmo
|
||||
efermi_new(i) = gw_efermi(dbl_mb(lSize),pars%nocc(i))
|
||||
enddo
|
||||
call gw_shift(dbl_mb(k_enew),dbl_mb(k_eold),efermi_new,
|
||||
$ nbf_ao,pars%nmo,pars%ipol)
|
||||
$ pars%nmo,pars%nmo,pars%ipol)
|
||||
efermi(:) = efermi(:) + efermi_new(:)
|
||||
call dcopy(pars%ipol*nmo,dbl_mb(k_eold),1,dbl_mb(k_enew),1)
|
||||
|
||||
|
|
@ -214,6 +237,23 @@ c
|
|||
endif
|
||||
endif
|
||||
|
||||
!
|
||||
! Save eigenvalues
|
||||
!
|
||||
if (main) then
|
||||
|
||||
call gw_shift(dbl_mb(k_enew),dbl_mb(k_eold),-efermi,nmo,nmo,ipol)
|
||||
call util_file_name('gwevals',.false.,.false.,gwevals)
|
||||
call util_file_name_resolve(gwevals,.false.)
|
||||
open(unit=unitno,status='unknown',form='unformatted',file=gwevals)
|
||||
do i=1,ipol
|
||||
call swrite(unitno, dbl_mb(k_eold+(i-1)*nmo),nmo)
|
||||
enddo
|
||||
close(unit=unitno)
|
||||
|
||||
endif
|
||||
|
||||
|
||||
9050 format(10x,I3,6x,F8.3,6x,F6.3,6x,F8.3)
|
||||
c
|
||||
if (.not.ma_chop_stack(l_sigma))
|
||||
|
|
|
|||
|
|
@ -33,7 +33,7 @@
|
|||
integer nclusters, clusters(pars%nmo), myll, myul
|
||||
integer llqp, ulqp, icluster
|
||||
double precision upper, lower, resupper, reslower, const
|
||||
double precision values(20), errors(20), radius, maxradius
|
||||
double precision values(100), errors(100), radius, maxradius
|
||||
c
|
||||
!
|
||||
! Initialize variables
|
||||
|
|
@ -60,7 +60,7 @@ c !
|
|||
if (eviter.gt.4) then
|
||||
maxradius = 0.010d0
|
||||
else
|
||||
maxradius = 0.1d0
|
||||
maxradius = 0.030d0
|
||||
endif
|
||||
|
||||
c !
|
||||
|
|
@ -75,9 +75,6 @@ c !
|
|||
totalqp = pars%noqp(isp) + pars%nvqp(isp)
|
||||
if (totalqp.lt.1) cycle
|
||||
|
||||
totalqp = pars%noqp(isp) + pars%nvqp(isp)
|
||||
if (totalqp.lt.1) cycle
|
||||
|
||||
esterror(:) = 0d0
|
||||
|
||||
if ((isp.eq.1).and.main) then
|
||||
|
|
@ -139,16 +136,18 @@ c !
|
|||
|
||||
if (eviter.lt.10) then
|
||||
if (myul.lt.ulqp) then
|
||||
Eout = Enew(llmo+iqp,isp) + 0.002d0
|
||||
c Eout = Enew(llmo+iqp,isp) + 0.002d0
|
||||
Eout = Eout + 0.001d0
|
||||
elseif (myll.gt.llqp) then
|
||||
Eout = Enew(llmo+iqp-2,isp) - 0.002d0
|
||||
c Eout = Enew(llmo+iqp-2,isp) - 0.002d0
|
||||
Eout = Eout - 0.001d0
|
||||
endif
|
||||
endif
|
||||
|
||||
bracket = .false.
|
||||
const = Eold(llmo+iqp-1,isp) + Sigma_X - sigmaold
|
||||
lower = 1d100
|
||||
upper = -1d100
|
||||
lower = -1d100
|
||||
upper = 1d100
|
||||
|
||||
c Compute Screened-Coulomb matrix elements
|
||||
timer = util_wallsec()
|
||||
|
|
@ -164,7 +163,7 @@ c Compute Galiitski-Migdal correlation energy
|
|||
endif
|
||||
|
||||
c
|
||||
do iter=1,pars%qpiter+5
|
||||
do iter=1,pars%qpiter+15
|
||||
timer2 = util_wallsec()
|
||||
|
||||
Ein = Eout
|
||||
|
|
|
|||
|
|
@ -48,6 +48,10 @@ c
|
|||
|
||||
logical converged,ok
|
||||
double precision, external :: gw_efermi
|
||||
|
||||
logical,external :: movecs_write
|
||||
integer,parameter :: unitno = 66
|
||||
character*255 gwevals
|
||||
c
|
||||
c -----------------------------------------------------------------
|
||||
c
|
||||
|
|
@ -83,6 +87,15 @@ c
|
|||
|
||||
call gw_vxc(dbl_mb(kSigma),nmo,pars)
|
||||
|
||||
if (pars%ncap) then
|
||||
do isp=1,ipol
|
||||
call daxpy(pars%nocc(isp),1d0,pars%vxddp(isp),0,
|
||||
$ dbl_mb(ksigma+(isp-1)*pars%nmo),1)
|
||||
call daxpy(pars%nvir(isp),1d0,pars%vxddm(isp),0,
|
||||
$ dbl_mb(ksigma+(isp-1)*pars%nmo+pars%nocc(isp)),1)
|
||||
enddo
|
||||
endif
|
||||
|
||||
pars%timer%vxc = pars%timer%vxc + util_wallsec() - timer
|
||||
c
|
||||
c Obtain Sigma_x
|
||||
|
|
@ -127,7 +140,6 @@ c
|
|||
lSize = lSize*max(pars%eviter,1)
|
||||
if (.not.ma_push_get(mt_dbl,lSize,'values',lValues,kValues))
|
||||
$ call errquit(pname//'failed to allocate values',0,MA_ERR)
|
||||
lSize = lSize*max(pars%eviter,1)
|
||||
if (.not.ma_push_get(mt_dbl,lSize,'errors',lErrors,kErrors))
|
||||
$ call errquit(pname//'failed to allocate errors',0,MA_ERR)
|
||||
endif
|
||||
|
|
@ -169,19 +181,14 @@ c
|
|||
9020 format(/,2x,"*** G",I1,"W",I1)
|
||||
9021 format(/,2x,"*** G",I2,"W",I2)
|
||||
|
||||
if (eviter.eq.0 .and. pars%ncap .and. pars%evgw) then
|
||||
do i=1,ipol
|
||||
do j=0,pars%nmo-1
|
||||
if (j.lt.pars%nocc(i)) then
|
||||
dbl_mb(kEmod+(i-1)*nmo+j) =
|
||||
$ dbl_mb(kEmod+(i-1)*nmo+j) + pars%vxddp(i)
|
||||
else
|
||||
dbl_mb(kEmod+(i-1)*nmo+j) =
|
||||
$ dbl_mb(kEmod+(i-1)*nmo+j) + pars%vxddm(i)
|
||||
endif
|
||||
enddo
|
||||
enddo
|
||||
endif
|
||||
call gw_shift(dbl_mb(kemod),dbl_mb(pars%k_evals),-efermi,
|
||||
$ pars%nmo,nbf_ao,pars%ipol)
|
||||
if(.not.movecs_write(pars%rtdb, ao_bas_han, movecs_out, 'dft',
|
||||
$ 'W vals',nbf_ao,ipol,(/pars%nmo,pars%nmo/),
|
||||
$ dbl_mb(pars%k_occ), nbf_ao,
|
||||
$ dbl_mb(pars%k_evals),nbf_ao,pars%g_movecs))
|
||||
$ call errquit('gw_analytic: could not save movecs',0,DISK_ERR)
|
||||
|
||||
c
|
||||
c Obtain eigenvalue difference
|
||||
c
|
||||
|
|
@ -194,23 +201,25 @@ c
|
|||
c
|
||||
c Experimental bit, might be useful for BSE calculations
|
||||
c
|
||||
if (eviter.eq.0 .and. pars%ncap .and. (.not.pars%evgw)) then
|
||||
do i=1,ipol
|
||||
do j=0,pars%nmo-1
|
||||
if (j.lt.pars%nocc(i)) then
|
||||
dbl_mb(kEmod+(i-1)*nmo+j) =
|
||||
$ dbl_mb(kEmod+(i-1)*nmo+j) + pars%vxddp(i)
|
||||
else
|
||||
dbl_mb(kEmod+(i-1)*nmo+j) =
|
||||
$ dbl_mb(kEmod+(i-1)*nmo+j) + pars%vxddm(i)
|
||||
endif
|
||||
enddo
|
||||
enddo
|
||||
endif
|
||||
c if (eviter.eq.0 .and. pars%ncap .and. (.not.pars%evgw)) then
|
||||
c do i=1,ipol
|
||||
c do j=0,pars%nmo-1
|
||||
c if (j.lt.pars%nocc(i)) then
|
||||
c dbl_mb(kEmod+(i-1)*nmo+j) =
|
||||
c $ dbl_mb(kEmod+(i-1)*nmo+j) + pars%vxddp(i)
|
||||
c else
|
||||
c dbl_mb(kEmod+(i-1)*nmo+j) =
|
||||
c $ dbl_mb(kEmod+(i-1)*nmo+j) + pars%vxddm(i)
|
||||
c endif
|
||||
c enddo
|
||||
c enddo
|
||||
c endif
|
||||
c
|
||||
c if (eviter.eq.0 .and. pars%ncap) then
|
||||
c call dcopy(nmo*ipol,dbl_mb(kEmod),1,dbl_mb(kEref),1)
|
||||
c endif
|
||||
c
|
||||
call dcopy(nmo*ipol, dbl_mb(kEmod), 1, dbl_mb(kEref), 1)
|
||||
c
|
||||
c Obtain Screened Coulomb matrices on imaginary axis
|
||||
c
|
||||
|
|
@ -277,11 +286,11 @@ c
|
|||
! Update Fermi energy
|
||||
!
|
||||
do i=1,ipol
|
||||
lSize = kEnew + (i-1)*nbf_ao
|
||||
lSize = kEnew + (i-1)*nmo
|
||||
efermi_new(i) = gw_efermi(dbl_mb(lSize),pars%nocc(i))
|
||||
enddo
|
||||
call gw_shift(dbl_mb(kEnew),dbl_mb(kemod),efermi_new,
|
||||
$ nbf_ao,pars%nmo,pars%ipol)
|
||||
$ pars%nmo,pars%nmo,pars%ipol)
|
||||
efermi(:) = efermi(:) + efermi_new(:)
|
||||
call dcopy(pars%ipol*nmo,dbl_mb(kemod),1,dbl_mb(kEnew),1)
|
||||
|
||||
|
|
@ -293,12 +302,26 @@ c
|
|||
endif
|
||||
endif
|
||||
|
||||
if (main) then
|
||||
|
||||
call gw_shift(dbl_mb(kenew),dbl_mb(kemod),-efermi,nmo,nmo,ipol)
|
||||
call util_file_name('gwevals',.false.,.false.,gwevals)
|
||||
call util_file_name_resolve(gwevals,.false.)
|
||||
open(unit=unitno,status='unknown',form='unformatted',file=gwevals)
|
||||
do i=1,ipol
|
||||
call swrite(unitno, dbl_mb(kemod+(i-1)*nmo),nmo)
|
||||
enddo
|
||||
close(unit=unitno)
|
||||
|
||||
endif
|
||||
|
||||
if (.not.ga_destroy(pars%g_sols))
|
||||
$ call errquit('gw_cdgw_newton: cannot destroy GA',0,GA_ERR)
|
||||
|
||||
if (.not.(ma_free_heap(liW).and.ma_free_heap(liWidx)))
|
||||
$ call errquit(pname//'cannot destroy iW arrays',0,MA_ERR)
|
||||
|
||||
|
||||
c
|
||||
end subroutine
|
||||
|
||||
|
|
|
|||
|
|
@ -60,7 +60,7 @@ C
|
|||
logical main, ok, parallel_diag, skip
|
||||
logical, external :: is_near
|
||||
|
||||
double precision, parameter :: shift = 0.0001d0
|
||||
double precision, parameter :: shift = 0.00005d0
|
||||
double complex :: w, R1, dR1
|
||||
double complex :: ieta
|
||||
|
||||
|
|
|
|||
|
|
@ -40,7 +40,7 @@
|
|||
integer nclusters, clusters(pars%nmo),myll,myul
|
||||
integer llqp,ulqp,icluster
|
||||
double precision upper,lower,resupper,reslower,const
|
||||
double precision values(25),errors(25),radius
|
||||
double precision values(100),errors(100),radius
|
||||
|
||||
|
||||
main = pars%me .eq. 0
|
||||
|
|
@ -110,9 +110,11 @@
|
|||
|
||||
if (eviter.lt.3) then
|
||||
if (myul.lt.ulqp) then
|
||||
Eout = Enew(llmo+iqp,isp) + 0.001d0
|
||||
!Eout = Enew(llmo+iqp,isp) + 0.001d0
|
||||
Eout = Eout + 0.001d0
|
||||
elseif (myll.gt.llqp) then
|
||||
Eout = Enew(llmo+iqp-2,isp) - 0.001d0
|
||||
!Eout = Enew(llmo+iqp-2,isp) - 0.001d0
|
||||
Eout = Eout - 0.001d0
|
||||
endif
|
||||
endif
|
||||
|
||||
|
|
@ -207,6 +209,7 @@ c
|
|||
|
||||
! Save last energy
|
||||
Enew(llmo+iqp-1,isp) = Eout
|
||||
Sigma(llmo+iqp-1,isp) = SigmaC + Sigma_X
|
||||
|
||||
! check convergence info
|
||||
if (converged) then
|
||||
|
|
|
|||
|
|
@ -27,21 +27,17 @@ c
|
|||
& rtdb_put(rtdb,'gw:graph',mt_log,1,.false.).and.
|
||||
& rtdb_put(rtdb,'gw:linear',mt_log,1,.false.).and.
|
||||
& rtdb_put(rtdb,'gw:eviter',mt_int,1,4).and.
|
||||
& rtdb_put(rtdb,'gw:qpiter',mt_int,1,10).and.
|
||||
& rtdb_put(rtdb,'gw:qpiter',mt_int,1,20).and.
|
||||
& rtdb_put(rtdb,'gw:noqp_alpha',mt_int,1,1).and.
|
||||
& rtdb_put(rtdb,'gw:nvqp_alpha',mt_int,1,0).and.
|
||||
& rtdb_put(rtdb,'gw:noqp_beta',mt_int,1,1).and.
|
||||
& rtdb_put(rtdb,'gw:nvqp_beta',mt_int,1,0).and.
|
||||
& rtdb_put(rtdb,'gw:ngl',mt_int,1,200).and.
|
||||
& rtdb_put(rtdb,'gw:ngl',mt_int,1,160).and.
|
||||
& rtdb_put(rtdb,'gw:nomega',mt_int,1,100).and.
|
||||
& rtdb_put(rtdb,'gw:first',mt_int,1,1).and.
|
||||
& rtdb_put(rtdb,'gw:threshold',mt_dbl,1,0.005d0/ha2ev).and.
|
||||
& rtdb_put(rtdb,'gw:ncap',mt_log,1,.false.).and.
|
||||
#ifdef GWCMPLX
|
||||
& rtdb_put(rtdb,'gw:eta',mt_dbl,1,0.0001d0).and.
|
||||
#else
|
||||
& rtdb_put(rtdb,'gw:eta',mt_dbl,1,0.00001d0).and.
|
||||
#endif
|
||||
& rtdb_put(rtdb,'gw:eta',mt_dbl,1,0.01d0).and.
|
||||
& rtdb_put(rtdb,'gw:rpa',mt_log,1,.false.).and.
|
||||
& rtdb_put(rtdb,'gw:diag',mt_log,1,.false.).and.
|
||||
& rtdb_put(rtdb,'gw:core',mt_log,1,.false.).and.
|
||||
|
|
|
|||
|
|
@ -30,8 +30,12 @@ C
|
|||
!
|
||||
!take full newton step for large enough z
|
||||
!
|
||||
if (z.gt.0.4d0 .and. z.le.1d0 ) then
|
||||
xout = xin + step
|
||||
if (z.gt.0.6d0 .and. z.le.1d0 ) then
|
||||
if (iter.eq.1) then
|
||||
xout = xin + 0.2d0*step
|
||||
else
|
||||
xout = xin + 0.7d0*step
|
||||
endif
|
||||
|
||||
!
|
||||
! when we have bracketed the solution, the
|
||||
|
|
@ -51,7 +55,11 @@ C
|
|||
! Take an scaled Newton step for small Z
|
||||
!
|
||||
elseif (z.gt.0.1d0) then
|
||||
xout = xin + 0.60d0*step
|
||||
if (iter.eq.1) then
|
||||
xout = xin + 0.10d0*step
|
||||
else
|
||||
xout = xin + 0.30d0*step
|
||||
endif
|
||||
|
||||
!
|
||||
! If Z is greater than 1.0 or very small we have to suspect
|
||||
|
|
@ -59,7 +67,8 @@ C
|
|||
! steepest decent step
|
||||
!
|
||||
else
|
||||
xout = xin + radius*fx
|
||||
c xout = xin + radius*fx
|
||||
xout = xin + sign(0.002d0,fx)
|
||||
|
||||
endif
|
||||
|
||||
|
|
|
|||
|
|
@ -31,6 +31,8 @@ c
|
|||
c
|
||||
logical movecs_read, movecs_read_header, atom_tag_check
|
||||
external movecs_read, movecs_read_header, atom_tag_check
|
||||
|
||||
logical,external :: movecs_write
|
||||
c
|
||||
integer ga_create_atom_blocked
|
||||
external ga_create_atom_blocked
|
||||
|
|
@ -46,8 +48,6 @@ c
|
|||
c
|
||||
c -----------------------------------------------------------------
|
||||
c
|
||||
zeta = 0.304121d0
|
||||
aq2 = 2d0/9d0 * (1d0-zeta)**2
|
||||
|
||||
pars%timer%init = util_wallsec()
|
||||
pars%timer%total = pars%timer%init
|
||||
|
|
@ -71,6 +71,9 @@ c
|
|||
pars%me = ga_nodeid()
|
||||
pars%nprocs = ga_nnodes()
|
||||
|
||||
!oskel = .false.
|
||||
!oadapt = .false.
|
||||
|
||||
#ifdef USE_OPENMP
|
||||
pars%iMaxThreads = omp_get_max_threads()
|
||||
#else
|
||||
|
|
@ -135,9 +138,9 @@ c
|
|||
$ pars%nvqp(1)))
|
||||
$ call errquit(pname//'failed to read nvqp',0,RTDB_ERR)
|
||||
if (.not.rtdb_get(pars%rtdb,'gw:noqp_beta',mt_int,1,pars%noqp(2)))
|
||||
$ call errquit(pname//'failed to read noqp',0,RTDB_ERR)
|
||||
$ call errquit(pname//'failed to read noqp beta',0,RTDB_ERR)
|
||||
if (.not.rtdb_get(pars%rtdb,'gw:nvqp_beta',mt_int,1,pars%nvqp(2)))
|
||||
$ call errquit(pname//'failed to read nvqp',0,RTDB_ERR)
|
||||
$ call errquit(pname//'failed to read nvqp beta',0,RTDB_ERR)
|
||||
c
|
||||
if (.not.rtdb_get(pars%rtdb,'gw:analyticw',mt_log,1,pars%anaw))
|
||||
$ call errquit(pname//'failed to read analyticw',0,RTDB_ERR)
|
||||
|
|
@ -236,7 +239,6 @@ c
|
|||
do j=1,pars%nmo
|
||||
if (dbl_mb(pars%k_occ+(i-1)*nbf_ao+j-1).gt.0.d0) then
|
||||
nocc = nocc + 1
|
||||
cycle
|
||||
else
|
||||
nvir = nvir + 1
|
||||
endif
|
||||
|
|
@ -244,7 +246,7 @@ c
|
|||
pars%nocc(i) = nocc
|
||||
pars%nvir(i) = nvir
|
||||
if (nocc+nvir.ne.pars%nmo)
|
||||
$ call errquit(pname//'something went wrong',0,0)
|
||||
$ call errquit(pname//'Nocc + Nvir /= NMO ',0,0)
|
||||
|
||||
c
|
||||
if ((pars%noqp(i).gt.nocc).or.(pars%noqp(i).eq.-1))
|
||||
|
|
@ -253,12 +255,11 @@ c
|
|||
if (pars%noqp(i).lt.-1)
|
||||
$ call errquit(pname//'invalid number noqp specified',0,0)
|
||||
c
|
||||
if ((pars%nvqp(i).gt.nvir).or.(pars%nvqp(i).eq.-1))
|
||||
if ((pars%nvqp(i).gt.nvir).or.(pars%nvqp(i).le.-1))
|
||||
$ pars%nvqp(i) = nvir
|
||||
|
||||
if (pars%nvqp(i).lt.-1)
|
||||
$ call errquit(pname//'invalid number nvqp specified',0,0)
|
||||
|
||||
! If core option was passed, then start counting from the bottom
|
||||
! else, start counting from the top (Fermi level)
|
||||
if (pars%core) then
|
||||
pars%llmo(i) = pars%first
|
||||
else
|
||||
|
|
@ -284,16 +285,23 @@ c
|
|||
c
|
||||
1001 format(2X,A12,F8.3,A3)
|
||||
if (pars%ncap) then
|
||||
zeta = 0.304121d0
|
||||
if ((xfac(87)).gt.1d-10) zeta=0.5d0
|
||||
aq2 = 2d0/9d0 * (1d0-zeta)**2
|
||||
do i=1,ipol
|
||||
ehomo = dbl_mb(pars%k_evals+(i-1)*nbf_ao+pars%nocc(i)-1)
|
||||
vxddp = -0.5d0*aq2*(1d0 + sqrt(1d0-4d0*ehomo/aq2))
|
||||
vxddm = -0.5d0*aq2*(1d0 - sqrt(1d0-4d0*ehomo/aq2))
|
||||
vxddp = -0.5d0*aq2*(1d0 + sqrt(1d0-4d0*ehomo/aq2))!*0.15d0
|
||||
vxddm = -0.5d0*aq2*(1d0 - sqrt(1d0-4d0*ehomo/aq2))!*0.15d0
|
||||
pars%vxddp(i) = vxddp
|
||||
pars%vxddm(i) = vxddm
|
||||
if (pars%me.eq.0) then
|
||||
write(luout,1001) ' v_x^{DD+}: ',vxddp*ha2ev, ' eV'
|
||||
write(luout,1001) ' v_x^(DD-): ',vxddm*ha2ev, ' eV'
|
||||
endif
|
||||
call daxpy(pars%nocc(i),1d0,vxddp,0,
|
||||
$ dbl_mb(pars%k_evals+(i-1)*nbf_ao),1)
|
||||
call daxpy(pars%nvir(i),1d0,vxddm,0,
|
||||
$ dbl_mb(pars%k_evals+(i-1)*nbf_ao+pars%nocc(i)),1)
|
||||
enddo
|
||||
endif
|
||||
c
|
||||
|
|
|
|||
|
|
@ -484,6 +484,7 @@ c if (.not.ga_create(mt_dbl,nri,npairs,erilabel,nri,0,eri))
|
|||
enddo
|
||||
|
||||
ok = ma_chop_stack(l_tmp)
|
||||
ok = ma_chop_stack(lmap)
|
||||
|
||||
call ga_distribution(eri,me,ld,ld,polelo,polehi)
|
||||
|
||||
|
|
|
|||
|
|
@ -93,6 +93,7 @@ c
|
|||
$ pars%g_movecs(isp),pars%g_movecs(isp),0d0,
|
||||
$ g_dens(isp))
|
||||
if (odftps) call pstat_off(ps_dgemm)
|
||||
call ga_symmetrize(g_dens(isp))
|
||||
enddo
|
||||
c
|
||||
nscr = 2*nbf_ao_mxnbf_ce**2
|
||||
|
|
@ -131,7 +132,7 @@ c
|
|||
if (.not.ma_push_get(mt_dbl,2*nbf_ao,'vxc temp',l_tmp,k_tmp))
|
||||
$ call errquit(pname//'failed to allocate temp',0,MA_ERR)
|
||||
do isp=1,ipol
|
||||
if (oskel)
|
||||
if (oskel)
|
||||
$ call sym_symmetrize(geom, ao_bas_han, .false., g_vxc(isp))
|
||||
call ga_dgemm('n','n',nbf_ao,nmo,nbf_ao,1.0d0,g_vxc(isp),
|
||||
$ pars%g_movecs(isp),0.0d0,g_dens(isp))
|
||||
|
|
|
|||
|
|
@ -332,6 +332,8 @@ ckbn mrcc-r-1 -3
|
|||
#endif
|
||||
else if (inp_compare(.false.,test,'gw')) then
|
||||
call gw_input(rtdb)
|
||||
else if (inp_compare(.false.,test,'bse')) then
|
||||
call bse_input(rtdb)
|
||||
else if (inp_compare(.false.,test,'xtb')) then
|
||||
call xtb_input(rtdb)
|
||||
c
|
||||
|
|
|
|||
|
|
@ -703,6 +703,7 @@ c
|
|||
xname(83)=' MN15-L Exchange Functional'
|
||||
xname(84)=' MN15 Exchange Functional'
|
||||
xname(85)=' M06-SX Exchange Functional'
|
||||
xname(87)=' NCAPR GGA Exchange Functional'
|
||||
c
|
||||
c Exchange-Correlation combination functional name defaults.
|
||||
c
|
||||
|
|
@ -785,6 +786,7 @@ c
|
|||
xcname(84)=' MN15 Method XC Functional'
|
||||
xcname(85)=' M06-SX Method XC Functional'
|
||||
xcname(86)=' r^2SCAN0 Method XC Functional'
|
||||
xcname(87)=' NCAPR Method XC Functional'
|
||||
c
|
||||
c place character XC parameters in rtdb
|
||||
c
|
||||
|
|
|
|||
|
|
@ -153,7 +153,7 @@ c
|
|||
c r2scan0 ==> 19
|
||||
c
|
||||
integer num_dirs, ind, mlen, iline, n, itemp
|
||||
parameter (num_dirs=168)
|
||||
parameter (num_dirs=169)
|
||||
c
|
||||
character*15 dirs(num_dirs)
|
||||
character*255 test
|
||||
|
|
@ -212,7 +212,7 @@ c
|
|||
& 'xscan', 'cscan',
|
||||
& 'hcth147@tz2p',
|
||||
W 'wb97','wb97x','wb97x-d3',
|
||||
& 'xr2scan', 'cr2scan','r2scan0' /
|
||||
& 'xr2scan', 'cr2scan','r2scan0', 'ncapr' /
|
||||
c becke97-d, ssb-d
|
||||
logical disp, dumdirect
|
||||
integer ivdw
|
||||
|
|
@ -298,7 +298,8 @@ c
|
|||
& 9710,9720,9650,9660,9615,9616,
|
||||
& 1947,
|
||||
& 9410,9420,9430,
|
||||
& 9661,9662,9670,
|
||||
& 9661,9662,9670,
|
||||
& 9730,
|
||||
& 1999 ) ind
|
||||
|
||||
|
||||
|
|
@ -1817,6 +1818,17 @@ c
|
|||
nlxfac(72) = .true.
|
||||
xfac(72) = 1.0d0
|
||||
goto 10
|
||||
|
||||
|
||||
9730 xccomb(87) = .true.
|
||||
cfac(3) = 1d0
|
||||
lcfac(3) = .true.
|
||||
cfac(4) = 1d0
|
||||
nlcfac(4) = .true.
|
||||
lxfac(87) = .true.
|
||||
nlxfac(87) = .true.
|
||||
xfac(87) = 1.0d0
|
||||
goto 10
|
||||
c
|
||||
c r2SCAN metaGGA
|
||||
c
|
||||
|
|
|
|||
|
|
@ -1646,6 +1646,7 @@ c Clean scratch space
|
|||
if (nodezero) call util_flush(LuOut)
|
||||
c
|
||||
endif ! gmh
|
||||
|
||||
c
|
||||
enddo ! outer loop over roots
|
||||
c
|
||||
|
|
|
|||
|
|
@ -1855,15 +1855,32 @@ c
|
|||
if (.not. do_2nd .and..not. do_3rd) then
|
||||
call xc_xncap(tol_rho, xfac(72), rho,
|
||||
& delrho, Amat, Cmat, nq, ipol,
|
||||
& Ex, qwght, ldew, func)
|
||||
& Ex, qwght, ldew, func, ' ')
|
||||
else if (.not. do_3rd) then
|
||||
call xc_xncap_d2(tol_rho, xfac(72), rho,
|
||||
& delrho, Amat, Amat2, Cmat, Cmat2, nq, ipol,
|
||||
& Ex, qwght, ldew, func)
|
||||
& Ex, qwght, ldew, func, ' ')
|
||||
else
|
||||
call xc_xncap_d3(tol_rho, xfac(72), rho,
|
||||
& delrho, Amat, Amat2, Amat3, Cmat, Cmat2, Cmat3,
|
||||
& nq, ipol, Ex, qwght, ldew, func)
|
||||
& nq, ipol, Ex, qwght, ldew, func, ' ')
|
||||
endif
|
||||
endif
|
||||
c
|
||||
|
||||
if (abs(xfac(87)).gt.eps)then
|
||||
if (.not. do_2nd .and..not. do_3rd) then
|
||||
call xc_xncap(tol_rho, xfac(87), rho,
|
||||
& delrho, Amat, Cmat, nq, ipol,
|
||||
& Ex, qwght, ldew, func, 'r')
|
||||
else if (.not. do_3rd) then
|
||||
call xc_xncap_d2(tol_rho, xfac(87), rho,
|
||||
& delrho, Amat, Amat2, Cmat, Cmat2, nq, ipol,
|
||||
& Ex, qwght, ldew, func, 'r')
|
||||
else
|
||||
call xc_xncap_d3(tol_rho, xfac(87), rho,
|
||||
& delrho, Amat, Amat2, Amat3, Cmat, Cmat2, Cmat3,
|
||||
& nq, ipol, Ex, qwght, ldew, func, 'r')
|
||||
endif
|
||||
endif
|
||||
c
|
||||
|
|
|
|||
|
|
@ -32,19 +32,23 @@ c
|
|||
c
|
||||
#if !defined SECOND_DERIV && !defined THIRD_DERIV
|
||||
Subroutine xc_xncap(tol_rho, fac, rho, delrho,
|
||||
& Amat, Cmat, nq, ipol, Ex, qwght,ldew,func)
|
||||
& Amat, Cmat, nq, ipol, Ex, qwght,ldew,func,
|
||||
& version)
|
||||
#elif defined(SECOND_DERIV) && !defined THIRD_DERIV
|
||||
Subroutine xc_xncap_d2(tol_rho, fac, rho, delrho,
|
||||
& Amat, Amat2, Cmat, Cmat2, nq, ipol, Ex,
|
||||
& qwght,ldew,func)
|
||||
& qwght,ldew,func,
|
||||
& version)
|
||||
#else
|
||||
Subroutine xc_xncap_d3(tol_rho, fac, rho, delrho,
|
||||
& Amat, Amat2, Amat3, Cmat, Cmat2, Cmat3,
|
||||
& nq, ipol, Ex, qwght,ldew,func)
|
||||
& nq, ipol, Ex, qwght,ldew,func,
|
||||
& version)
|
||||
#endif
|
||||
c
|
||||
implicit none
|
||||
c
|
||||
c
|
||||
character*1 version
|
||||
double precision fac, Ex
|
||||
integer nq, ipol
|
||||
logical ldew
|
||||
|
|
@ -77,8 +81,7 @@ c
|
|||
double precision ckf, Ax
|
||||
double precision F43, F13, F23, F49
|
||||
c
|
||||
parameter(mu=0.2195149727645171d0, beta=0.01808569669d0)
|
||||
parameter(zeta=0.30412141859531383d0)
|
||||
parameter(mu=0.2195149727645171d0)
|
||||
|
||||
parameter (F43=4.d0/3.d0, F13=1.d0/3.d0, F23=2.0d0/3.0d0)
|
||||
parameter (F49=4d0/9d0)
|
||||
|
|
@ -100,11 +103,22 @@ c
|
|||
#endif
|
||||
double precision ops, omz, logops
|
||||
|
||||
double precision s2overgamma,s2overgamma_rho43
|
||||
double precision dFxd1s,dFxd2s
|
||||
|
||||
arcsinh(s)=dlog(s+dsqrt(1d0+s*s))
|
||||
darcsinh(s)=1d0/dsqrt(1d0+s*s)
|
||||
d2arcsinh(s) = -s/dsqrt(1d0+s*s)**3
|
||||
d3arcsinh(s) = (2d0*s*s - 1d0)/dsqrt(1d0+s*s)**5
|
||||
c
|
||||
if (version.eq.'r') then
|
||||
zeta = 0.5d0
|
||||
beta = 0.017983d0
|
||||
else
|
||||
zeta = 0.30412141859531383d0
|
||||
beta = 0.01808569669d0
|
||||
endif
|
||||
|
||||
pi = acos(-1.d0)
|
||||
ckf = (3d0*pi*pi)**F13
|
||||
Ax = -3d0/(4d0*pi)*ckf
|
||||
|
|
@ -130,11 +144,21 @@ CDEC$ NOSWP
|
|||
s = dsqrt(gamma)/(2d0*ckf*rho43)
|
||||
s2 = s*s
|
||||
|
||||
s2overgamma = 1d0/(4d0*ckf*ckf*rho43*rho43)
|
||||
s2overgamma_rho43 = 1d0/(4d0*ckf*ckf*rho43)
|
||||
d1s = 0.5d0*s/gamma
|
||||
|
||||
if (s.lt.1d-8) then
|
||||
Fx = 1d0 + mu*s2 + alpha*zeta*mu*s2*s +
|
||||
& mu*(alpha - beta - alpha*zeta - 0.5d0)*s2*s2
|
||||
dFx = 2d0*mu*s + 3d0*alpha*zeta*mu*s2 +
|
||||
& 4d0*mu*(alpha - beta - alpha*zeta - 0.5d0)*s2*s
|
||||
dFxd1s = mu*s2overgamma*
|
||||
& (1d0 + 1.5d0*alpha*zeta*s +
|
||||
& 2d0*(alpha-beta-alpha*zeta-0.5d0)*s2)
|
||||
dFxd2s = -0.5d0*mu*
|
||||
& (1d0 + 1.5d0*alpha*zeta*s +
|
||||
& 2d0*(alpha-beta-alpha*zeta-0.5d0)*s2)
|
||||
d2Fx = 2d0*mu + 6d0*alpha*zeta*mu*s +
|
||||
& 12d0*mu*(alpha - beta - alpha*zeta - 0.5d0)*s2
|
||||
d3Fx = 6d0*alpha*zeta*mu +
|
||||
|
|
@ -147,7 +171,7 @@ CDEC$ NOSWP
|
|||
asymp = 1d0 + alpha*(omz*s*logops + zeta*s)
|
||||
dasymp = alpha*( (s + zeta)/ops + omz*logops )
|
||||
|
||||
if (s.gt.30d0) then
|
||||
if (s.gt.20d0) then
|
||||
tansin = arcsinh(s)
|
||||
dtansin = darcsinh(s)
|
||||
else
|
||||
|
|
@ -159,18 +183,19 @@ CDEC$ NOSWP
|
|||
ddenom = beta*dtansin
|
||||
|
||||
factor = tansin/denom
|
||||
dfactor = (dtansin - factor*ddenom)/denom
|
||||
dfactor = dtansin*(1d0 - beta*factor)/denom
|
||||
|
||||
Fx = 1d0 + mu*factor*asymp
|
||||
dFx = mu*(dfactor*asymp + factor*dasymp)
|
||||
dFxd1s = dFx*d1s
|
||||
dFxd2s = -0.5d0*dFxd1s/s2overgamma
|
||||
|
||||
#if defined(SECOND_DERIV) || defined(THIRD_DERIV)
|
||||
if (s.gt.30d0) then
|
||||
if (s.gt.24d0) then
|
||||
d2tansin = d2arcsinh(s)
|
||||
else
|
||||
d2tansin = 2d0*darcsinh(s)/dcosh(s)**2 +
|
||||
& dtanh(s)*d2arcsinh(s) -
|
||||
& 2d0*tansin/dcosh(s)**2
|
||||
d2tansin = (darcsinh(s)-tansin)*2d0/dcosh(s)**2 +
|
||||
& dtanh(s)*d2arcsinh(s)
|
||||
endif
|
||||
d2asymp = alpha*omz*(2d0 + s)/ops**2
|
||||
d2denom = beta*d2tansin
|
||||
|
|
@ -202,9 +227,8 @@ CDEC$ NOSWP
|
|||
Ex = Ex + Ax*rho43*Fx*qwght(n)*fac
|
||||
if (ldew) func(n) = func(n) + Ax*rho43*Fx*fac
|
||||
|
||||
d1s = 0.5d0*s/gamma
|
||||
Amat(n,1) = Amat(n,1) + F43*Ax*rho13*(Fx - s*dFx)*fac
|
||||
Cmat(n,D1_GAA) = Cmat(n,D1_GAA) + 2d0*Ax*rho43*dFx*d1s*fac
|
||||
Cmat(n,D1_GAA) = Cmat(n,D1_GAA) + 2d0*Ax*rho43*dFxd1s*fac
|
||||
|
||||
#if defined(SECOND_DERIV) || defined(THIRD_DERIV)
|
||||
|
||||
|
|
@ -214,10 +238,10 @@ CDEC$ NOSWP
|
|||
& (Fx - s*dFx + 4d0*s2*d2Fx)*fac
|
||||
|
||||
Cmat2(n,D2_RA_GAA) = Cmat2(n,D2_RA_GAA) -
|
||||
& (F43*Ax*rho13*d2Fx*d1s*s)*4d0*fac
|
||||
& (0.5d0*F43*Ax*rho13*d2Fx*s2overgamma)*4d0*fac
|
||||
|
||||
Cmat2(n,D2_GAA_GAA) = Cmat2(n,D2_GAA_GAA) + Ax*rho43*
|
||||
& (d2Fx*d1s**2 + dFx*d2s)*8d0*fac
|
||||
Cmat2(n,D2_GAA_GAA) = Cmat2(n,D2_GAA_GAA) + Ax*
|
||||
& s2overgamma_rho43/gamma*(0.25d0*d2Fx + dFxd2s)*8d0*fac
|
||||
#endif
|
||||
#ifdef THIRD_DERIV
|
||||
rhom53 = rhom23*rrho
|
||||
|
|
|
|||
|
|
@ -9,7 +9,7 @@
|
|||
task_save_state.o task.o task_input.o task_shell_input.o task_num_grad.o \
|
||||
task_vscf.o task_property.o task_dntmc.o task_bsse.o \
|
||||
task_jefftce.o task_ncc.o task_rfcf.o \
|
||||
task_gwmol.o \
|
||||
task_gwmol.o task_bsemol.o \
|
||||
task_aoints.o task_nadtrc.o
|
||||
|
||||
USES_BLAS = task_num_grad.F task_bsse.F
|
||||
|
|
|
|||
|
|
@ -163,6 +163,10 @@ c
|
|||
logical task_nadtrc
|
||||
external task_nadtrc
|
||||
|
||||
logical task_bsemol
|
||||
external task_bsemol
|
||||
|
||||
|
||||
c logical embed_optimize
|
||||
c external embed_optimize
|
||||
c
|
||||
|
|
@ -468,6 +472,8 @@ c status=embed_prepare(rtdb)
|
|||
status=task_dntmc(rtdb)
|
||||
else if (inp_compare(.false., 'gw', operation)) then
|
||||
status = task_gwmol(rtdb)
|
||||
else if (inp_compare(.false., 'bse', operation)) then
|
||||
status = task_bsemol(rtdb)
|
||||
c
|
||||
c **** check for theories that have specific operations ****
|
||||
c
|
||||
|
|
|
|||
138
src/task/task_bsemol.F
Normal file
138
src/task/task_bsemol.F
Normal file
|
|
@ -0,0 +1,138 @@
|
|||
c
|
||||
c Task handler for Molecular BSE calcualtions
|
||||
c Daniel Mejia-Rodriguez, 2022
|
||||
c
|
||||
logical function task_bsemol(rtdb)
|
||||
c
|
||||
implicit none
|
||||
c
|
||||
#include "errquit.fh"
|
||||
#include "rtdb.fh"
|
||||
#include "mafdecls.fh"
|
||||
#include "inp.fh"
|
||||
#include "stdio.fh"
|
||||
#include "util.fh"
|
||||
#include "global.fh"
|
||||
c
|
||||
integer rtdb
|
||||
c
|
||||
character*32 theory
|
||||
character*255 title
|
||||
integer iproc, ipol
|
||||
logical oprint, rodft, dftmp2, status, converged, ignore
|
||||
double precision cputime, walltime
|
||||
c
|
||||
logical bsemol, task_energy
|
||||
external bsemol, task_energy
|
||||
c
|
||||
c Initialization
|
||||
c
|
||||
task_bsemol = .false.
|
||||
c
|
||||
c Time at the start
|
||||
c
|
||||
cputime = util_cpusec()
|
||||
walltime = util_wallsec()
|
||||
c
|
||||
c Invalidate status
|
||||
c
|
||||
if (.not. rtdb_put( rtdb, 'task:status', mt_log, 1, .false.))
|
||||
& call errquit('task_bse: failed to invalidate status',0,RTDB_ERR)
|
||||
c
|
||||
c Print BSE header
|
||||
c
|
||||
oprint = util_print('information', print_low)
|
||||
if (oprint) then
|
||||
if (.not. rtdb_cget(rtdb, 'title', 1, title)) title = ' '
|
||||
if (ga_nodeid().eq.0) then
|
||||
write(LuOut,*)
|
||||
call util_print_centered(
|
||||
& LuOut, 'NWChem BSE Module', 40, .true.)
|
||||
write(LuOut,*)
|
||||
write(LuOut,*)
|
||||
if (title.ne.' ') then
|
||||
call util_print_centered(LuOut, title, 40, .false.)
|
||||
write(LuOut,*)
|
||||
write(LuOut,*)
|
||||
endif
|
||||
endif
|
||||
endif
|
||||
c
|
||||
c Check if theory has been specified
|
||||
c
|
||||
if (.not. rtdb_cget(rtdb, 'task:theory', 1, theory))
|
||||
& call errquit('task_bsemol: theory not specified',0, RTDB_ERR)
|
||||
c
|
||||
c Restrict to DFT for now (SCF could also be used)
|
||||
c
|
||||
if (.not.inp_compare(.false.,'dft',theory))
|
||||
& call errquit('task_bsemol: BSE is only implemented for DFT', 0,
|
||||
& INPUT_ERR)
|
||||
c
|
||||
if (rtdb_get(rtdb, 'dft:rodft', mt_log, 1, rodft))
|
||||
& call errquit('task_bsemol: BSE is not implemented for RODFT', 0,
|
||||
& INPUT_ERR)
|
||||
c
|
||||
if (rtdb_get(rtdb, 'dft:dftmp2', mt_log, 1, dftmp2))
|
||||
& call errquit('task_bsemol: BSE is not implemented for double'//
|
||||
& ' hybrid functionals',0,INPUT_ERR)
|
||||
c
|
||||
if (.not.rtdb_get(rtdb, 'dft:ipol', mt_int, 1, ipol))
|
||||
& call errquit('task_bsemol: failed to read ipol', 0, RTDB_ERR)
|
||||
|
||||
if (.not. rtdb_get(rtdb, 'task:ignore', mt_log, 1, ignore))
|
||||
& ignore = .true.
|
||||
c
|
||||
c Make sure a DFT solution is available
|
||||
c
|
||||
status = rtdb_get(rtdb, 'dft:converged', mt_log, 1, converged)
|
||||
if (.not.status) then
|
||||
status = task_energy(rtdb)
|
||||
if (.not.rtdb_get(rtdb, 'dft:converged', mt_log, 1, converged))
|
||||
& call errquit('task_bsemol: failed to read convergence info',0,
|
||||
& RTDB_ERR)
|
||||
endif
|
||||
|
||||
status = converged
|
||||
|
||||
|
||||
if (ga_nodeid().eq.0) then
|
||||
write(luout,*) " "
|
||||
write(luout,'(29X,A)') "***********************"
|
||||
write(luout,'(29X,A)') "*** BSE ***"
|
||||
write(luout,'(29X,A)') "***********************"
|
||||
write(luout,*) " "
|
||||
|
||||
if (.not.converged) then
|
||||
write(LuOut,*)
|
||||
write(LuOut,'(27X,A)') '!!! WARNING: !!!'
|
||||
write(LuOut,'(27X,A)') '!!! DFT DID NOT CONVERGE !!!'
|
||||
write(LuOut,*)
|
||||
endif
|
||||
|
||||
call util_flush(luout)
|
||||
endif
|
||||
|
||||
if ((.not.status) .and. (.not.ignore))
|
||||
& call errquit('task_bsemol: Energy failure',0,CALC_ERR)
|
||||
c
|
||||
c Actually compute GW
|
||||
c
|
||||
status = bsemol(rtdb)
|
||||
c
|
||||
cputime = util_cpusec() - cputime
|
||||
walltime = util_wallsec() - walltime
|
||||
c
|
||||
if (.not. rtdb_put(rtdb, 'task:status', mt_log, 1, .true.))
|
||||
& call errquit('task_bse: failed to set status',0,
|
||||
& RTDB_ERR)
|
||||
if (.not. rtdb_put(rtdb, 'task:cputime', mt_dbl, 1, cputime))
|
||||
& call errquit('task_bse: failed storing cputime',0,
|
||||
& RTDB_ERR)
|
||||
if (.not. rtdb_put(rtdb, 'task:walltime', mt_dbl, 1, walltime))
|
||||
& call errquit('task_bse: failed storing walltime',0,
|
||||
& RTDB_ERR)
|
||||
c
|
||||
task_bsemol = status
|
||||
c
|
||||
end function
|
||||
|
|
@ -37,7 +37,7 @@ c
|
|||
c task [qmmm] <string theory> [<string operation = energy>] [numerical || analytic]
|
||||
c [ignore]
|
||||
c
|
||||
logical numerical, ignore, oqmmm, analytic
|
||||
logical numerical, ignore, oqmmm, analytic, gwbse
|
||||
logical orism
|
||||
logical omm
|
||||
logical opes
|
||||
|
|
@ -61,6 +61,7 @@ c
|
|||
opes = .false.
|
||||
ofragment = .false.
|
||||
osmd = .false.
|
||||
gwbse = .false.
|
||||
c oembed = .false.
|
||||
c
|
||||
call rtdb_clean(rtdb, 'task:') ! Delete ALL previous task parameters
|
||||
|
|
@ -114,6 +115,10 @@ c
|
|||
operation = 'rt_tddft'
|
||||
else if (inp_compare(.false.,test,'gw')) then
|
||||
operation = 'gw'
|
||||
gwbse = .true.
|
||||
else if (inp_compare(.false.,test,'bse')) then
|
||||
operation = 'bse'
|
||||
gwbse = .true.
|
||||
else if (inp_compare(.false.,test,'fcidump')) then
|
||||
operation = 'fcidump'
|
||||
c else if (inp_compare(.false.,test,'prepare')) then
|
||||
|
|
@ -377,6 +382,7 @@ c
|
|||
c
|
||||
if (theory .eq. ' ') then
|
||||
if(osmd) theory="smd"
|
||||
if(gwbse) theory='dft'
|
||||
end if
|
||||
if (theory .ne. ' ') then
|
||||
if (.not. rtdb_cput(rtdb, 'task:theory', 1, theory))
|
||||
|
|
|
|||
|
|
@ -769,6 +769,7 @@ $(STAMP_DIR):
|
|||
|
||||
clean:
|
||||
@test -e $(BUILDDIR)/Makefile && { cd $(BUILDDIR) && $(MAKE) clean; } || echo "Not configured"
|
||||
rm -f build install
|
||||
rm -f ../mpi_include.txt ../ga*txt
|
||||
|
||||
realclean:
|
||||
|
|
|
|||
|
|
@ -16,7 +16,7 @@ c (c) Battelle, PNNL, 2002.
|
|||
write(LuOut,*)
|
||||
write(LuOut,*) title
|
||||
write(LuOut,9000)
|
||||
do k = 0,mdim/6
|
||||
do k = 0,(mdim-1)/6
|
||||
write(LuOut,*)
|
||||
write(LuOut,9010) (j,j=k*6+1,min(mdim,k*6+6))
|
||||
do i=1,ndim
|
||||
|
|
|
|||
|
|
@ -31,6 +31,7 @@ check_file () {
|
|||
grep -s 'l DFT energy' $file |tail
|
||||
echo ============================================================
|
||||
grep -s 'Total PSPW energy' $file |tail
|
||||
grep -s -B 3 "Root 1 singlet b1u" $file |grep E-0
|
||||
fi
|
||||
}
|
||||
|
||||
|
|
@ -46,6 +47,7 @@ check_file h2o2-response
|
|||
check_file pspw
|
||||
check_file pspw_md
|
||||
check_file aump2
|
||||
check_file bse_ethene
|
||||
# stuff for OpenBLAS issues on macos
|
||||
echo 'looking for cgemm_incopy.o'
|
||||
find $TRAVIS_BUILD_DIR/src/libext -name cgemm_incopy.o
|
||||
|
|
|
|||
|
|
@ -28,6 +28,7 @@ fi
|
|||
echo NWCHEM_TOP is $NWCHEM_TOP
|
||||
#TARBALL=https://github.com/nwchemgit/nwchem/releases/download/v7.0.0-beta1/nwchem-7.0.0-release.revision-5bcf0416-src.2019-11-01.tar.bz2
|
||||
export USE_MPI=y
|
||||
export MPICH_INTERFACE_HOSTNAME=localhost # fix for Mpich interface errors
|
||||
if [[ "$FC" == "flang" ]]; then
|
||||
export BUILD_MPICH=1
|
||||
fi
|
||||
|
|
@ -184,9 +185,9 @@ if [[ -z "$USE_INTERNALBLAS" ]]; then
|
|||
# if [[ ${FC} == gfortran ]] ; then
|
||||
if [[ "$MPI_IMPL" == "mpich" ]]; then
|
||||
export MPICH_FC=$FC
|
||||
if [[ "$arch" != "aarch64" ]]; then
|
||||
export BUILD_MPICH=1
|
||||
fi
|
||||
# if [[ "$arch" != "aarch64" ]]; then
|
||||
# export BUILD_MPICH=1
|
||||
# fi
|
||||
fi
|
||||
if [[ `${CC} -dM -E - < /dev/null 2> /dev/null | grep -c clang` == 0 ]] && [[ `${CC} -dM -E - < /dev/null 2> /dev/null | grep -c GNU` > 0 ]] && [[ "$(expr `${CC} -dumpversion | cut -f1 -d.` \> 7)" == 1 ]]; then
|
||||
if [[ "$os" == "Linux" ]] && [[ "$arch" == "x86_64" ]]; then
|
||||
|
|
|
|||
|
|
@ -35,7 +35,7 @@ if [[ ! -z "$USE_OPENMP" ]]; then
|
|||
export OMP_NUM_THREADS="$USE_OPENMP"
|
||||
export OMP_STACKSIZE=32M
|
||||
fi
|
||||
if [[ "$os" == "Darwin" && "$BUILD_MPICH" == 1 ]] || [[ "$os" == "Darwin" && "$ARMCI_NETWORK" == MPI-PT ]] ; then
|
||||
if [[ "$os" == "Darwin" && "$ARMCI_NETWORK" == MPI-PT ]] ; then
|
||||
nprocs=1
|
||||
fi
|
||||
env|egrep MP
|
||||
|
|
@ -88,6 +88,14 @@ env|egrep MP
|
|||
;;
|
||||
esac
|
||||
;;
|
||||
ARMCI)
|
||||
case "$os" in
|
||||
# fixes segfault for bse_ethene in ga_copy_patch...ARMCII_Iov_op_datatype
|
||||
Linux)
|
||||
ulimit -s unlimited
|
||||
;;
|
||||
esac
|
||||
;;
|
||||
esac
|
||||
if [[ "$MPI_IMPL" == "openmpi" ]]; then
|
||||
export MPIRUN_NPOPT=" --allow-run-as-root -mca mpi_yield_when_idle 0 --oversubscribe -np "
|
||||
|
|
@ -132,6 +140,9 @@ fi
|
|||
cd $TRAVIS_BUILD_DIR/QA && ./runtests.mpi.unix procs $nprocs ritddft_h2o ritddft_co
|
||||
cd $TRAVIS_BUILD_DIR/QA && NWCHEM_BASIS_LIBRARY=${NWCHEM_TOP}/src/basis/libraries.bse/ ./runtests.mpi.unix procs $nprocs gw_closedshell gw_openshell gw_symmetry
|
||||
fi
|
||||
if [[ ! $(grep -i bse $TRAVIS_BUILD_DIR/src/stubs.F| awk '/bse_input/') ]]; then
|
||||
cd $TRAVIS_BUILD_DIR/QA && ./runtests.mpi.unix procs $nprocs bse_ethene
|
||||
fi
|
||||
if [[ ! $(grep -i dplot $TRAVIS_BUILD_DIR/src/stubs.F| awk '/dplot_input/') ]]; then
|
||||
cd $TRAVIS_BUILD_DIR/QA && ./runtests.mpi.unix procs $nprocs rt_tddft_dplot
|
||||
fi
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue