Add host-side TMA support to AOTInductor #138878
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/138878
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit e417138 with merge base 72ea7ba ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
aakhundov
added
topic: not user facing
desertfire
reviewed
Oct 25, 2024
| autotune_configs=configs, | ||
| ) | ||
|
|
||
| def generate_tma_descriptor(self, desc): |
There was a problem hiding this comment.
Because AOTI still uses two-pass run at the moment, does the cpp wrapper codegen here need to read any information from the python run? I have added things like DeferredGpuGridLine for that purpose. Do you think it is necessary to do something like that here?
There was a problem hiding this comment.
I believe, the Python and C++ codegens for the TMA descriptor creation are independent.
desertfire
approved these changes
Oct 28, 2024
chenyang78
approved these changes
Oct 28, 2024
|
@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 |
rahulsingh-intel
pushed a commit
to rahulsingh-intel/pytorch
that referenced
this pull request
Oct 29, 2024
This adds host-side Triton TMA support to AOTInductor. Notes: - Two helper functions, `init1DTMADescriptor` and `init2DTMADescriptor` are added to the C++ wrapper codegen on GPU, conditioned on the model having user-defined Triton kernels with host-side TMA (CUDA-specific). - C++ wrapper codegen on GPU emits TMA descriptor initialization via the aforementioned helper functions. - Special handling added for the TMA descriptors (in the Python wrapper codegen) during the compile-time autotuning, as the underlying tensor can't be passed directly to the user-defined Triton kernel. TMA descriptors are generated in-between the source tensor's buffer and the kernel call, like in the full Python wrapper codegen. - This PR concludes the host-side Triton TMA support in PT2. Pull Request resolved: pytorch#138878 Approved by: https://github.com/desertfire, https://github.com/chenyang78 ghstack dependencies: pytorch#138759, pytorch#138877
rahulsingh-intel
pushed a commit
to rahulsingh-intel/pytorch
that referenced
this pull request
Nov 5, 2024
This adds host-side Triton TMA support to AOTInductor. Notes: - Two helper functions, `init1DTMADescriptor` and `init2DTMADescriptor` are added to the C++ wrapper codegen on GPU, conditioned on the model having user-defined Triton kernels with host-side TMA (CUDA-specific). - C++ wrapper codegen on GPU emits TMA descriptor initialization via the aforementioned helper functions. - Special handling added for the TMA descriptors (in the Python wrapper codegen) during the compile-time autotuning, as the underlying tensor can't be passed directly to the user-defined Triton kernel. TMA descriptors are generated in-between the source tensor's buffer and the kernel call, like in the full Python wrapper codegen. - This PR concludes the host-side Triton TMA support in PT2. Pull Request resolved: pytorch#138878 Approved by: https://github.com/desertfire, https://github.com/chenyang78 ghstack dependencies: pytorch#138759, pytorch#138877
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Stack from ghstack (oldest at bottom):
This adds host-side Triton TMA support to AOTInductor. Notes:
init1DTMADescriptorandinit2DTMADescriptorare added to the C++ wrapper codegen on GPU, conditioned on the model having user-defined Triton kernels with host-side TMA (CUDA-specific).cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @yf225 @chenyang78 @kadeng @muchulee8 @ColinPeppler @amjames @desertfire @chauhang