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

[FEEDSTOCK REQUEST] Triton #1078

Open
ce107 opened this issue Apr 14, 2024 · 0 comments
Open

[FEEDSTOCK REQUEST] Triton #1078

ce107 opened this issue Apr 14, 2024 · 0 comments

Comments

@ce107
Copy link

ce107 commented Apr 14, 2024

Describe the package you'd like added
OpenAI Triton for PPC64LE especially as there is no way to use it currently on Summit and other remaining AC922 systems.

Describe how this package fits in with the project
Triton is necessary for PyTorch compile to work - any of my attempts to use PyTorch compile on PPC64LE fails.

Additional context
Currently Triton will not build on PPC64LE because the build script downloads a preselected version of LLVM that is x86_64 and more recently ARM64 specific. Triton could also be useful moving forward with P10 MMA if it can learn to create MMA specialized kernels.

The code in question is in python/setup.py (get_llvm_package_info and friends) - and it can be made to point to an alternative location of a prebuilt PPC64LE version of LLVM - of the right level as explained in the Triton build instructions for an alternative LLVM build.

There are a few other places in the code where architecture/os checks happen and a PPC64LE entry would be necessary.

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

No branches or pull requests

1 participant