-
Notifications
You must be signed in to change notification settings - Fork 26.3k
[user triton] Ignore backend-specific args in the TTIR analysis #141062
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
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/141062
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ✅ No FailuresAs of commit 66bf59f with merge base 93aef68 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Fixes #140800. On AMD, backend-specific args like `matrix_instr_nonkdim`, `waves_per_eu` and `kpack` are passed either direclty to the kernel or via `triton.Config`, whereas they don't exist as kernel parameters. Native Triton code handles those excessive args [here](https://github.com/triton-lang/triton/blob/a6bb57d6285e723c58e87dd7cba263db6efff789/python/triton/runtime/jit.py#L594-L596). In this PR, we add similar handling to the TTIR analysis code to avoid bailing out. ghstack-source-id: 70037b8 Pull Request resolved: #141062
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
…rch#141062) Fixes pytorch#140800. On AMD, backend-specific args like `matrix_instr_nonkdim`, `waves_per_eu` and `kpack` are passed either direclty to the kernel or via `triton.Config`, whereas they don't exist as kernel parameters. Native Triton code handles those excessive args [here](https://github.com/triton-lang/triton/blob/a6bb57d6285e723c58e87dd7cba263db6efff789/python/triton/runtime/jit.py#L594-L596). In this PR, we add similar handling to the TTIR analysis code to avoid bailing out. Pull Request resolved: pytorch#141062 Approved by: https://github.com/oulgen (cherry picked from commit b740a1b)
…rch#141062) Fixes pytorch#140800. On AMD, backend-specific args like `matrix_instr_nonkdim`, `waves_per_eu` and `kpack` are passed either direclty to the kernel or via `triton.Config`, whereas they don't exist as kernel parameters. Native Triton code handles those excessive args [here](https://github.com/triton-lang/triton/blob/a6bb57d6285e723c58e87dd7cba263db6efff789/python/triton/runtime/jit.py#L594-L596). In this PR, we add similar handling to the TTIR analysis code to avoid bailing out. Pull Request resolved: pytorch#141062 Approved by: https://github.com/oulgen
… analysis (#141… (#1768) Fixes pytorch#140800. On AMD, backend-specific args like `matrix_instr_nonkdim`, `waves_per_eu` and `kpack` are passed either direclty to the kernel or via `triton.Config`, whereas they don't exist as kernel parameters. Native Triton code handles those excessive args [here](https://github.com/triton-lang/triton/blob/a6bb57d6285e723c58e87dd7cba263db6efff789/python/triton/runtime/jit.py#L594-L596). In this PR, we add similar handling to the TTIR analysis code to avoid bailing out. Pull Request resolved: pytorch#141062 Approved by: https://github.com/oulgen (cherry picked from commit b740a1b) Fixes #ISSUE_NUMBER Co-authored-by: Adnan Akhundov <[email protected]>
Stack from ghstack (oldest at bottom):
Fixes #140800.
On AMD, backend-specific args like
matrix_instr_nonkdim,waves_per_euandkpackare passed either direclty to the kernel or viatriton.Config, whereas they don't exist as kernel parameters. Native Triton code handles those excessive args here. In this PR, we add similar handling to the TTIR analysis code to avoid bailing out.