Add distributed PointToPoint op send/recv/reduce/all2all/gather/scatter #1204
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
| work->future_->markCompleted(at::IValue(*work->outputs_)); | ||
| return work; | ||
| } else { | ||
| at::xpu::OptionalXPUGuard gpuGuard(device); |
There was a problem hiding this comment.
Please add some comments to when work will be constructed in coalesced collective.
src/xccl/ProcessGroupXCCL.cpp
Outdated
| TORCH_CHECK(tensors.size() == 1, MULTI_DEVICE_ERROR_MSG); | ||
| // @lint-ignore CLANGTIDY | ||
| auto tensor = tensors.back(); | ||
| check_xpu_single_tensor(tensor, true); |
There was a problem hiding this comment.
I think the name is already refined in your last PR. Please update.
| (outputSplitsEqual ? outLen : outputSplitSizes[i] * outLen); | ||
| } | ||
| auto xcclDataType = getXcclDataType(output.scalar_type()); | ||
| ccl::alltoallv( |
There was a problem hiding this comment.
alltoallv will not exsit in new oneCCL API. Please use send/recv pair for alltoallv.
There was a problem hiding this comment.
I will do this as soon as the new API is ready.
src/xccl/ProcessGroupXCCL.cpp
Outdated
| auto device = outputTensors[0].device(); | ||
| int64_t total_numel = 0; | ||
| for (const auto r : c10::irange(outputTensors.size())) { | ||
| check_xpu_single_tensor(outputTensors[r], true); |
Chao1Han
changed the title
Chao1Han
changed the title
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Outdated suggestions cannot be applied.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:
implement pointtopoint op send/recv/reduce/all2all/gather/scatter