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

Recommit [OPTIMIZATION] Fix addptr combine pattern (#3148) #3289

Merged
merged 1 commit into from
Mar 11, 2024

Conversation

junparser
Copy link
Contributor

Add addptr(addptr(%ptr, %idx0), %idx1) pattern again with constraint which makes sure that both element type of offset should be same.

@junparser junparser requested a review from ptillet as a code owner March 6, 2024 06:01
@junparser
Copy link
Contributor Author

@ThomasRaoux, this will fix the crash.

@junparser
Copy link
Contributor Author

Seems the error log "RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)" is irrelevant with this patch

@ThomasRaoux
Copy link
Collaborator

Seems the error log "RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)" is irrelevant with this patch

GPU errors are asynchronous so in general this poont to a crash that may be due to code generated by triton. I'll kick off a rerun but it is most likely a legetimate error

@junparser junparser force-pushed the addptr-combine branch 2 times, most recently from f6e3d34 to c739377 Compare March 8, 2024 01:50
@junparser
Copy link
Contributor Author

@ThomasRaoux would you help merge this? Thanks!

Add addptr(addptr(%ptr, %idx0), %idx1) pattern again with constraint
which makes sure that both element type of offset should be same.
@ThomasRaoux ThomasRaoux enabled auto-merge (squash) March 11, 2024 04:18
@ThomasRaoux ThomasRaoux merged commit 91fcb98 into triton-lang:main Mar 11, 2024
4 checks passed
@junparser junparser deleted the addptr-combine branch March 13, 2024 06:03
htyu pushed a commit to htyu/triton that referenced this pull request Mar 20, 2024
…triton-lang#3289)

Add addptr(addptr(%ptr, %idx0), %idx1) pattern again with constraint
which makes sure that both element type of offset should be same.
karupayun pushed a commit to openxla/triton that referenced this pull request Apr 3, 2024
…triton-lang#3289)

Add addptr(addptr(%ptr, %idx0), %idx1) pattern again with constraint
which makes sure that both element type of offset should be same.
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.

2 participants