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

[AOT][BACKEND][AMD] Refactor Triton AOT function and add AMD GPU support #3751

Draft
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

jayfurmanek
Copy link
Contributor

@jayfurmanek jayfurmanek commented Apr 24, 2024

This commit refactors the Triton AOT function to conform to the new backend model. Backend specific files (compiler.c, link.py) have been moved the appropriate places in the third_party directory.

AMD GPU support has also been added to the AOT compiler, launcher and unit tests.

#include <stdint.h>
#include <stdio.h>

#endif
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this line placed here instead of end of a file intentionally?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think so. That is a bit odd but its the same in third_party/nvidia/backend/tools/compile.h which is how it was before. I just made it the same on the hip side.

@xinyazhang
Copy link

@jayfurmanek is there any plan to add cross-compiling support?
Also is there any chance to get the raw GPU binary blob instead of a combination of source/header files?

to conform to the new backend model. Backend specific
files (compiler.c, link.py) have been moved the appropriate places
in the third_party directory.

AMD GPU support has also been added to the AOT compiler, launcher and unit
tests.
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.

None yet

3 participants