Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Block Jacobi not parallelized for matrix-matrix system #1391

Open
blegouix opened this issue Aug 16, 2023 · 1 comment
Open

Block Jacobi not parallelized for matrix-matrix system #1391

blegouix opened this issue Aug 16, 2023 · 1 comment
Labels
is:affects-performance This is related to something which affects performance.

Comments

@blegouix
Copy link

Hello,

This issue follows #1381, in which the problem was not correctly identified.

I am using Bicgstab to solve a test problem with A a 1000x1000 matrix (this is a band matrice with 19 band width stored in Csr format whose non-zeros are like 1 2 3 4 5 6 7 8 9 10 9 8 7 6 5 4 3 2 1). B is a 1000x1000 dense matrix filled with ones.

The number of colums in B (or X) is the size n_batch=1000 of batch (but this is a batch where all individuals systems share the same A, that's why I dont use the new BatchDense or BatchCsr classes).

I compare the performance on GPU with and without Jacobi preconditionner (size 32):

  • Without Jacobi : 4700 iterations, 5s total execution time.
  • With Jacobi : 490 iterations, 20s total execution time.

So, the execution time of one iteration is much longer with preconditionner (~19ms vs ~1ms). This effect does not appears with n_batch=1 (matrix-vector system, total execution time 0.8s).

batch_compare_ginkgo2

The reason is n_batch gko::kernels::cuda::jacobi::kernel::apply are called sequentially. Could it be improved ?

Regards

@MarcelKoch
Copy link
Member

Perhaps to side step the issue with the block Jacobi, you could create the Jacobi preconditioner with .with_max_block_size(1u). Maybe that helps you already getting a shorter runtime.

But of course, we need to fix our implementation.

@MarcelKoch MarcelKoch added the is:affects-performance This is related to something which affects performance. label Nov 6, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
is:affects-performance This is related to something which affects performance.
Projects
None yet
Development

No branches or pull requests

2 participants