[CPU] Support SHM based inference_all_reduce in TorchBackend #5391
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
delock
changed the title
25 tasks
|
Hi @loadams the formatting error had been fixed, thanks! |
tjruwase
reviewed
Apr 12, 2024
tjruwase
reviewed
Apr 12, 2024
tjruwase
approved these changes
Apr 15, 2024
rraminen
pushed a commit
to ROCm/DeepSpeed
that referenced
this pull request
May 9, 2024
…ft#5391) This PR adds SHM based `inference_all_reduce` kernel to `TorchBackend` communication backend. When inference on CPU server, this path replaces default `torch.distributed.all_reduce` which eventurally use gloo backend. This PR will improve inference performance with AutoTP when only stock PyTorch is installed without Intel Extension for PyTorch. Compared with gloo backend. SHM based inference_all_reduce kernel is a more directed path and perform much better on single node. | message size | gloo all_reduce(ms) | SHM all_reduce(ms) | | --- | --- | --- | | 32MB | 30.7 | 0.65 | | 64KB | 0.23 | 0.028 | In text generation of bloom-3b with AutoTP, average token latency improved 1.45x with this PR on 2S Xeon node. --------- Co-authored-by: Logan Adams <[email protected]> Co-authored-by: Olatunji Ruwase <[email protected]>
umchand
pushed a commit
to umchand/DeepSpeed
that referenced
this pull request
May 20, 2024
…ft#5391) This PR adds SHM based `inference_all_reduce` kernel to `TorchBackend` communication backend. When inference on CPU server, this path replaces default `torch.distributed.all_reduce` which eventurally use gloo backend. This PR will improve inference performance with AutoTP when only stock PyTorch is installed without Intel Extension for PyTorch. Compared with gloo backend. SHM based inference_all_reduce kernel is a more directed path and perform much better on single node. | message size | gloo all_reduce(ms) | SHM all_reduce(ms) | | --- | --- | --- | | 32MB | 30.7 | 0.65 | | 64KB | 0.23 | 0.028 | In text generation of bloom-3b with AutoTP, average token latency improved 1.45x with this PR on 2S Xeon node. --------- Co-authored-by: Logan Adams <[email protected]> Co-authored-by: Olatunji Ruwase <[email protected]>
dbyoung18
pushed a commit
to dbyoung18/DeepSpeed
that referenced
this pull request
Jun 11, 2024
…ft#5391) This PR adds SHM based `inference_all_reduce` kernel to `TorchBackend` communication backend. When inference on CPU server, this path replaces default `torch.distributed.all_reduce` which eventurally use gloo backend. This PR will improve inference performance with AutoTP when only stock PyTorch is installed without Intel Extension for PyTorch. Compared with gloo backend. SHM based inference_all_reduce kernel is a more directed path and perform much better on single node. | message size | gloo all_reduce(ms) | SHM all_reduce(ms) | | --- | --- | --- | | 32MB | 30.7 | 0.65 | | 64KB | 0.23 | 0.028 | In text generation of bloom-3b with AutoTP, average token latency improved 1.45x with this PR on 2S Xeon node. --------- Co-authored-by: Logan Adams <[email protected]> Co-authored-by: Olatunji Ruwase <[email protected]>
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.
This PR adds SHM based
inference_all_reducekernel toTorchBackendcommunication backend. When inference on CPU server, this path replaces defaulttorch.distributed.all_reducewhich eventurally use gloo backend. This PR will improve inference performance with AutoTP when only stock PyTorch is installed without Intel Extension for PyTorch.Compared with gloo backend. SHM based inference_all_reduce kernel is a more directed path and perform much better on single node.
In text generation of bloom-3b with AutoTP, average token latency improved 1.45x with this PR on 2S Xeon node.