Add initial support for intel Gaudi accelerators #2121
Merged
Conversation
This file contains 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
ankurneog
requested review from
merrymercy,
Ying1123,
hnyls2002,
zhyncs,
ispobock and
ByronHsu
as code owners
|
@liangan1 : can you have a look at the initial changes. thanks |
liangan1
reviewed
Nov 22, 2024
| @@ -186,6 +186,9 @@ def init_torch_distributed(self): | |||
| elif self.device == "xpu": | |||
| torch.xpu.set_device(self.gpu_id) | |||
| backend = "gloo" | |||
| elif self.device == "hpu": | |||
| torch.get_device_module(self.device).set_device(self.gpu_id) | |||
| backend = "hccl" | |||
There was a problem hiding this comment.
Refine other device path with "torch.get_device_module(self.device).set_device(self.gpu_id)" may be a better choice to make the code simplifier.
There was a problem hiding this comment.
done , this will be further simplified once pytorch/pytorch#140536 is available, that way we dont have any device name references in the file.
ankurneog
force-pushed
the
intel_gaudi_support_1
branch
from
ba8495f to
91c7594
Compare
ankurneog
changed the title
37 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Motivation
Intel gaudi ( device name : hpu) has PyTorch support as an Out-of-tree device. In addition support for Triton is also available . With this initial PR we introduce intel gaudi ( device : hpu) to SGLang
More details on Intel Gaudi SW installation can be found here :
https://docs.habana.ai/en/latest/PyTorch/Getting_Started_with_PyTorch_and_Gaudi/Getting_Started_with_PyTorch.html#getting-started-pyt-model
Triton support
https://docs.habana.ai/en/latest/PyTorch/Inference_on_PyTorch/Triton_Inference.html