-
Notifications
You must be signed in to change notification settings - Fork 5.6k
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
[Reshard] Implement reshard from s to r with same process_mesh #56039
Merged
Merged
Changes from all commits
Commits
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
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
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
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
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
62 changes: 62 additions & 0 deletions
62
paddle/phi/core/distributed/auto_parallel/reshard_all_gather_functor.cc
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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,62 @@ | ||
| // Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved. | ||
| // | ||
| // Licensed under the Apache License, Version 2.0 (the "License"); | ||
| // you may not use this file except in compliance with the License. | ||
| // You may obtain a copy of the License at | ||
| // | ||
| // http://www.apache.org/licenses/LICENSE-2.0 | ||
| // | ||
| // Unless required by applicable law or agreed to in writing, software | ||
| // distributed under the License is distributed on an "AS IS" BASIS, | ||
| // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
| // See the License for the specific language governing permissions and | ||
| // limitations under the License. | ||
|
|
||
| #include "paddle/phi/core/distributed/auto_parallel/reshard_all_gather_functor.h" | ||
|
|
||
| #include "paddle/phi/backends/all_context.h" | ||
| #include "paddle/phi/core/distributed/auto_parallel/reshard_utils.h" | ||
| #include "paddle/phi/core/visit_type.h" | ||
| #include "paddle/phi/infermeta/unary.h" | ||
| #include "paddle/phi/kernels/all_gather_kernel.h" | ||
|
|
||
| namespace phi { | ||
| namespace distributed { | ||
|
|
||
| DenseTensor ReshardAllGatherFunctor(DeviceContext* dev_ctx, | ||
| const DenseTensor& input, | ||
| const std::vector<int64_t>& process_ids) { | ||
| DenseTensor out; | ||
|
|
||
| int64_t world_size = process_ids.size(); | ||
| auto* comm_context = CreateOrGetCommContext(*dev_ctx, process_ids); | ||
| dev_ctx->SetCommContext(comm_context); | ||
|
|
||
| if (phi::CPUContext::classof(dev_ctx)) { | ||
| PD_VISIT_FLOATING_AND_INTEGRAL_TYPES( | ||
| input.dtype(), "AllGather", ([&] { | ||
| AllGather<data_t>(static_cast<const CPUContext&>(*dev_ctx), | ||
| input, | ||
| world_size, | ||
| &out); | ||
| })); | ||
| return out; | ||
| } | ||
| #if defined(PADDLE_WITH_CUDA) || defined(PADDLE_WITH_HIP) | ||
| if (phi::GPUContext::classof(dev_ctx)) { | ||
| PD_VISIT_FLOATING_AND_INTEGRAL_TYPES( | ||
| input.dtype(), "AllGather", ([&] { | ||
| AllGather<data_t>(static_cast<const GPUContext&>(*dev_ctx), | ||
| input, | ||
| world_size, | ||
| &out); | ||
| })); | ||
| return out; | ||
| } | ||
| #endif | ||
| PADDLE_THROW(phi::errors::Unimplemented( | ||
| "The all_gather in reshard only supported on CPU and GPU for now.")); | ||
| } | ||
|
|
||
| } // namespace distributed | ||
| } // namespace phi |
31 changes: 31 additions & 0 deletions
31
paddle/phi/core/distributed/auto_parallel/reshard_all_gather_functor.h
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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,31 @@ | ||
| // Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved. | ||
| // | ||
| // Licensed under the Apache License, Version 2.0 (the "License"); | ||
| // you may not use this file except in compliance with the License. | ||
| // You may obtain a copy of the License at | ||
| // | ||
| // http://www.apache.org/licenses/LICENSE-2.0 | ||
| // | ||
| // Unless required by applicable law or agreed to in writing, software | ||
| // distributed under the License is distributed on an "AS IS" BASIS, | ||
| // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
| // See the License for the specific language governing permissions and | ||
| // limitations under the License. | ||
|
|
||
| #pragma once | ||
|
|
||
| #include <cstdint> | ||
| #include <vector> | ||
|
|
||
| namespace phi { | ||
| class DenseTensor; | ||
| class DeviceContext; | ||
|
|
||
| namespace distributed { | ||
|
|
||
| DenseTensor ReshardAllGatherFunctor(DeviceContext* dev_ctx, | ||
| const DenseTensor& input, | ||
| const std::vector<int64_t>& process_ids); | ||
|
|
||
| } // namespace distributed | ||
| } // namespace phi |
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
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
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
73 changes: 73 additions & 0 deletions
73
paddle/phi/core/distributed/auto_parallel/s_to_r_reshard_function.cc
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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,73 @@ | ||
| // Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved. | ||
| // | ||
| // Licensed under the Apache License, Version 2.0 (the "License"); | ||
| // you may not use this file except in compliance with the License. | ||
| // You may obtain a copy of the License at | ||
| // | ||
| // http://www.apache.org/licenses/LICENSE-2.0 | ||
| // | ||
| // Unless required by applicable law or agreed to in writing, software | ||
| // distributed under the License is distributed on an "AS IS" BASIS, | ||
| // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
| // See the License for the specific language governing permissions and | ||
| // limitations under the License. | ||
|
|
||
| #include "paddle/phi/core/distributed/auto_parallel/s_to_r_reshard_function.h" | ||
|
|
||
| #include "glog/logging.h" | ||
| #include "paddle/phi/core/distributed/auto_parallel/dist_attr.h" | ||
| #include "paddle/phi/core/distributed/auto_parallel/dist_tensor.h" | ||
| #include "paddle/phi/core/distributed/auto_parallel/reshard_all_gather_functor.h" | ||
| #include "paddle/phi/core/distributed/auto_parallel/reshard_utils.h" | ||
| #include "paddle/phi/core/distributed/comm_context_manager.h" | ||
| #include "paddle/phi/core/distributed/store/tcp_store.h" | ||
|
|
||
| namespace phi { | ||
| namespace distributed { | ||
|
|
||
| bool SToRReshardFunction::IsSuitable( | ||
| const DistTensor& in, | ||
| const std::shared_ptr<TensorDistAttr>& out_dist_attr) { | ||
| bool flag = true; | ||
| const auto& in_dist_attr = in.dist_attr(); | ||
|
|
||
| const auto& in_dims_mapping = in_dist_attr->dims_mapping(); | ||
| const auto& out_dims_mapping = out_dist_attr->dims_mapping(); | ||
|
|
||
| flag &= IsDimsMappingShard(in_dims_mapping); | ||
| flag &= IsDimsMappingReplicated(out_dims_mapping); | ||
|
|
||
| const auto& in_process_mesh = in_dist_attr->process_mesh(); | ||
| const auto& out_process_mesh = out_dist_attr->process_mesh(); | ||
|
|
||
| flag &= (in_process_mesh.ndim() == 1); | ||
| flag &= (out_process_mesh.ndim() == 1); | ||
| flag &= (in_process_mesh == out_process_mesh); | ||
|
|
||
| return flag; | ||
| } | ||
|
|
||
| std::shared_ptr<DistTensor> SToRReshardFunction::Eval( | ||
| DeviceContext* dev_ctx, | ||
| const DistTensor& in, | ||
| const std::shared_ptr<TensorDistAttr>& out_dist_attr) { | ||
| // TODO(liyurui): Only support transfer shard(0) to replicate for now. | ||
| // Concat is needed when transfer shard(x) to replicate, will be supported | ||
| // later. | ||
| const DenseTensor& in_physical_tensor_cur_rank = in.value(); | ||
| const auto& in_dist_attr = in.dist_attr(); | ||
| const auto& in_process_mesh = in_dist_attr->process_mesh(); | ||
| const auto& in_process_ids = in_process_mesh.process_ids(); | ||
|
|
||
| // Since the precondition ensure the out_process_ids is equal to the | ||
| // in_process_ids, so the participate process ids mush equal to either | ||
| // in_process_ids or out_process_ids. | ||
| DenseTensor out_all_gather = ReshardAllGatherFunctor( | ||
| dev_ctx, in_physical_tensor_cur_rank, in_process_ids); | ||
|
|
||
| return std::make_shared<DistTensor>( | ||
| std::make_shared<DenseTensor>(out_all_gather), out_dist_attr); | ||
| } | ||
|
|
||
| } // namespace distributed | ||
| } // namespace phi |
Oops, something went wrong.
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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
仅仅支持一维process mesh吗?如果高维和低维都是一个process id编号，key是否相同？
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
这个函数只负责把传入的process_ids的vector变成唯一的comm_key，有两种情况：
不管是高维还是低维，只要它们参与通信的进程相同，key就是相同的