search

intel / intel-xpu-backend-for-triton

OpenAI Triton backend for Intel® GPUs
MIT License
109 stars 29 forks source link

Initiate a pull request to Triton that enables dispatching from Triton frontend to XPU backend #271

Open LiyangLingIntel opened 6 months ago

LiyangLingIntel commented 6 months ago

The latest Triton refactoring removed the Intel Triton backend from the third-party, llvm-target branch is a fork of openai/Triton with in-tree modifications. To upstream Intel XPU Triton backend to openai/triton, we need to extract and clean up these in-tree modifications and initiate a PR to Triton repo. The updated glue code should enable the dispatching from Triton frontend to XPU backend besides NV and AMD backends.

This task is related to https://github.com/intel/intel-xpu-backend-for-triton/issues/170.

LiyangLingIntel commented 6 months ago

There should be a mechanism that regularly updates intel-xpu-backend-for-triton as a third-party backend to OpenAI Triton repository.

etiotto commented 6 months ago

In order to upstream change to OpenAI we first need to have #170 complete. We also need to change the FE to use the SYCL RT.

etiotto commented 3 months ago

I do not think this issue is current. Putting it in the backlog for now.