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

Improve performance of dpnp.matmul and dpnp.dot with out keyword #1694

Merged
merged 6 commits into from
Feb 8, 2024

Conversation

vtavana
Copy link
Collaborator

@vtavana vtavana commented Feb 6, 2024

In this PR, implementation of dpnp.matmul and dpnp.dot is updated to improve the performance when out keyword is present and its feature matches the appropriate shape, dtype, sycl_queue, and usm_type that is needed for performing the calculation in the corresponding OneMKL BLAS routines.

Sample timing including this update

>>> import dpnp
>>> a = dpnp.ones((21846, 3), device='cpu')
>>> b = dpnp.ones((21846, 3), device='cpu')
>>> c = dpnp.ones((21846, 21846), device='cpu')

>>> %timeit dpnp.matmul(a, b.T, c)
229 ms ± 22 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

Sample timing excluding this update

>>> import dpnp
>>> a = dpnp.ones((21846, 3), device='cpu')
>>> b = dpnp.ones((21846, 3), device='cpu')
>>> c = dpnp.ones((21846, 21846), device='cpu')

>>> %timeit dpnp.matmul(a, b.T, c)
798 ms ± 30.7 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
  • Have you provided a meaningful PR description?
  • Have you added a test, reproducer or referred to issue with a reproducer?
  • Have you tested your changes locally for CPU and GPU devices?
  • Have you made sure that new changes do not introduce compiler warnings?
  • Have you checked performance impact of proposed changes?
  • If this PR is a work in progress, are you filing the PR as a draft?

@vtavana vtavana self-assigned this Feb 6, 2024
Copy link
Contributor

github-actions bot commented Feb 6, 2024

View rendered docs @ https://intelpython.github.io/dpnp/pull//index.html

@vtavana vtavana marked this pull request as ready for review February 6, 2024 21:18
dpnp/dpnp_utils/dpnp_utils_linearalgebra.py Outdated Show resolved Hide resolved
dpnp/dpnp_utils/dpnp_utils_linearalgebra.py Outdated Show resolved Hide resolved
dpnp/dpnp_utils/dpnp_utils_linearalgebra.py Outdated Show resolved Hide resolved
Copy link
Collaborator

@npolina4 npolina4 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Performance improved.
Thanks @vtavana!

Copy link
Contributor

@antonwolfy antonwolfy left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you @vtavana for resolving the issue so quickly

dpnp/dpnp_utils/dpnp_utils_linearalgebra.py Outdated Show resolved Hide resolved
@vtavana vtavana merged commit d45bb24 into master Feb 8, 2024
45 checks passed
@vtavana vtavana deleted the improve_out_performance branch February 8, 2024 04:58
github-actions bot added a commit that referenced this pull request Feb 8, 2024
#1694)

* use out keyword for result

* fix strided or overlapping out

* address comments

* fix typo

* remove additional check d45bb24
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants