aparler/PyNorch
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fix cuda sum reduce

Browse source
This commit is contained in:
lucasdelimanogueira 2024-05-01 03:09:05 -03:00
parent a6da5bc9af
commit 9981bb8b6f
8 changed files with 12 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

6
norch/csrc/tensor.cpp
View file

@ -171,9 +171,9 @@ extern "C" {
if (strcmp(tensor->device, "cuda") == 0) {
float* result_data;
int num_of_blocks = (tensor->size + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK;
cudaMalloc((void **)&num_of_blocks, 1 * sizeof(float));
sum_tensor_cuda(tensor, result_data, int num_of_blocks);
int number_of_blocks = (tensor->size + THREADS_PER_BLOCK - 1) / THREADS_PER_BLOCK;
cudaMalloc((void**)&result_data, number_of_blocks * sizeof(float));
sum_tensor_cuda(tensor, result_data, number_of_blocks);
return create_tensor(result_data, shape, ndim, device);
}
else {