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.

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

atomic.add operation leads to incorrect results #774

Closed
diptorupd opened this issue Sep 21, 2022 · 2 comments
Closed

atomic.add operation leads to incorrect results #774

diptorupd opened this issue Sep 21, 2022 · 2 comments
Assignees
@diptorupd
Labels
bug Something isn't working

Comments

@diptorupd
Copy link
Collaborator

A test micro-benchmark implemented in dpbench (https://github.com/IntelPython/dpbench/blob/main/dpbench/benchmarks/l2_distance/l2_distance_numba_dpex_k.py) produces incorrect results. The most likely culprit is the usage of atomic.add.

The generated code for numba-dpex's atomic add function needs to be evaluated v/s dpcpp.
@diptorupd
Copy link
Collaborator Author

@mingjie-intel this might be related to #778. Maybe the llvm-spirv we are using is too archaic and generates incorrect code.

@diptorupd diptorupd added the bug Something isn't working label Mar 11, 2023
@diptorupd diptorupd mentioned this issue Jul 31, 2023
Merged
5 tasks
@diptorupd
Copy link
Collaborator Author

Closing the issue as the benchmark no longer exists and we cannot reproduce the problem. In any case, the fix in #1103 will fix the original issue.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@diptorupd diptorupd

Labels
bug Something isn't working
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@diptorupd @mingjie-intel