Skip to content

[SYCL][libclc] Fix identity for multiplication #4337

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.

Sign up for GitHub

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

Jump to bottom
Merged
merged 2 commits into from
Aug 16, 2021
Merged

[SYCL][libclc] Fix identity for multiplication #4337

merged 2 commits into from
Aug 16, 2021

Conversation

Michoumichmich
Copy link
Contributor

@Michoumichmich Michoumichmich commented Aug 13, 2021 •
edited
Loading

This resolves #4336 issue which is a bug related to 0 being used as the identity in the CUDA back-end.

Signed-off-by: Michel Migdal [email protected]

@Michoumichmich Michoumichmich requested a review from bader as a code owner August 13, 2021 22:13
@bader bader requested a review from a team August 15, 2021 06:54
@bader bader added cuda CUDA back-end libclc libclc project related issues labels Aug 15, 2021
@bader bader changed the title [SYCL][libclc] fix identity for multiplication [SYCL][libclc] Fix identity for multiplication Aug 15, 2021
@bader
Copy link
Contributor

bader commented Aug 15, 2021

/summary:run

@bader
Copy link
Contributor

bader commented Aug 15, 2021

@Michoumichmich, thanks for working on this. Could you add the test from #4336 to https://github.com/intel/llvm-test-suite/tree/intel/SYCL/GroupAlgorithm, please? There are already tests for inclusive_scan there, which can be extended/updated.

@Michoumichmich
Copy link
Contributor Author

Michoumichmich commented Aug 15, 2021 •
edited
Loading

@Michoumichmich, thanks for working on this. Could you add the test from #4336 to https://github.com/intel/llvm-test-suite/tree/intel/SYCL/GroupAlgorithm, please? There are already tests for inclusive_scan there, which can be extended/updated.

Sure, I will open a PR there. The test missed that issue because they used std::iota(, , 0) which means that the expected result vector vas always full of zeros.

@Michoumichmich Michoumichmich mentioned this pull request Aug 15, 2021
Merged
@bader bader merged commit a6447ca into intel:sycl Aug 16, 2021
@Michoumichmich Michoumichmich deleted the fix_nvidiacl_group_multiplies branch August 17, 2021 14:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@bader bader bader approved these changes

Assignees
No one assigned
Labels
cuda CUDA back-end libclc libclc project related issues
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Joint_inclusive_scan returns the wrong result
2 participants
@Michoumichmich @bader