[WIP] Add JIT graph fuser for oneDNN Graph API #50256
Closed
Conversation
This file contains hidden or 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
facebook-github-bot
added
cla signed
oncall: jit
🔗 Helpful links
💊 CI failures summary and remediationsAs of commit fa56a64 (more details on the Dr. CI page): Commit fa56a64 was recently pushed. Waiting for builds... This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions to the (internal) Dr. CI Users group. |
facebook-github-bot
pushed a commit
that referenced
this pull request
Mar 21, 2022
Summary: ## Description Preview4 PR of this [RFC](#49444). On the basis of #50256, the below improvements are included: - The [preview4 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.4.1) of the oneDNN Graph API is used - The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: - SkyLake 8180 (1 socket of 28 cores):  - SkyLake 8180 (single thread):  \* By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) \** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code are placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is: ``` caffe2/CMakeLists.txt ``` ## Limitations - In this PR, we have only supported the optimization on Linux platform. The support on Windows and MacOS will be enabled as the next step. - We have only optimized the inference use case. Pull Request resolved: #68111 Reviewed By: eellison Differential Revision: D34584878 Pulled By: malfet fbshipit-source-id: ce817aa8cc9052ee9ed930c9cf66be83449e61a4
pytorchmergebot
pushed a commit
that referenced
this pull request
Mar 21, 2022
Summary: ## Description Preview4 PR of this [RFC](#49444). On the basis of #50256, the below improvements are included: - The [preview4 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.4.1) of the oneDNN Graph API is used - The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: - SkyLake 8180 (1 socket of 28 cores):  - SkyLake 8180 (single thread):  \* By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) \** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code are placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is: ``` caffe2/CMakeLists.txt ``` ## Limitations - In this PR, we have only supported the optimization on Linux platform. The support on Windows and MacOS will be enabled as the next step. - We have only optimized the inference use case. Pull Request resolved: #68111 Reviewed By: eellison Differential Revision: D34584878 Pulled By: malfet fbshipit-source-id: ce817aa8cc9052ee9ed930c9cf66be83449e61a4 (cherry picked from commit cd17683)
sanchitintel
pushed a commit
to sanchitintel/pytorch
that referenced
this pull request
Mar 22, 2022
Summary: ## Description Preview4 PR of this [RFC](pytorch#49444). On the basis of pytorch#50256, the below improvements are included: - The [preview4 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.4.1) of the oneDNN Graph API is used - The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: - SkyLake 8180 (1 socket of 28 cores):  - SkyLake 8180 (single thread):  \* By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) \** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code are placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is: ``` caffe2/CMakeLists.txt ``` ## Limitations - In this PR, we have only supported the optimization on Linux platform. The support on Windows and MacOS will be enabled as the next step. - We have only optimized the inference use case. Pull Request resolved: pytorch#68111 Reviewed By: eellison Differential Revision: D34584878 Pulled By: malfet fbshipit-source-id: ce817aa8cc9052ee9ed930c9cf66be83449e61a4
shahofblah
pushed a commit
that referenced
this pull request
Mar 25, 2022
Summary: ## Description Preview4 PR of this [RFC](#49444). On the basis of #50256, the below improvements are included: - The [preview4 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.4.1) of the oneDNN Graph API is used - The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: - SkyLake 8180 (1 socket of 28 cores):  - SkyLake 8180 (single thread):  \* By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) \** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code are placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is: ``` caffe2/CMakeLists.txt ``` ## Limitations - In this PR, we have only supported the optimization on Linux platform. The support on Windows and MacOS will be enabled as the next step. - We have only optimized the inference use case. Pull Request resolved: #68111 Reviewed By: eellison Differential Revision: D34584878 Pulled By: malfet fbshipit-source-id: ce817aa8cc9052ee9ed930c9cf66be83449e61a4 (cherry picked from commit cd17683)
facebook-github-bot
pushed a commit
that referenced
this pull request
Apr 28, 2022
….1) (#74596) Summary: Re-landing #68111 ## Description Preview4 PR of this [RFC](#49444). On the basis of #50256, the below improvements are included: - The [preview4 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.4.1) of the oneDNN Graph API is used - The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: - SkyLake 8180 (1 socket of 28 cores):  - SkyLake 8180 (single thread):  \* By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) \** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code are placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is: ``` caffe2/CMakeLists.txt ``` ## Limitations - In this PR, we have only supported the optimization on Linux platform. The support on Windows and MacOS will be enabled as the next step. - We have only optimized the inference use case. Pull Request resolved: #74596 Reviewed By: osalpekar Differential Revision: D35450045 Pulled By: eellison fbshipit-source-id: 151ec796c3eec785b7bda2b43f3532a4588aa36f
pytorchmergebot
pushed a commit
that referenced
this pull request
Apr 29, 2022
Re-landing #68111 ## Description Preview4 PR of this [RFC](#49444). On the basis of #50256, the below improvements are included: - The [preview4 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.4.1) of the oneDNN Graph API is used - The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: - SkyLake 8180 (1 socket of 28 cores):  - SkyLake 8180 (single thread):  \* By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) \** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code are placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is: ``` caffe2/CMakeLists.txt ``` ## Limitations - In this PR, we have only supported the optimization on Linux platform. The support on Windows and MacOS will be enabled as the next step. - We have only optimized the inference use case. Pull Request resolved: #74596 Approved by: https://github.com/malfet
sanchitintel
pushed a commit
to sanchitintel/pytorch
that referenced
this pull request
Apr 30, 2022
Re-landing pytorch#68111 ## Description Preview4 PR of this [RFC](pytorch#49444). On the basis of pytorch#50256, the below improvements are included: - The [preview4 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.4.1) of the oneDNN Graph API is used - The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: - SkyLake 8180 (1 socket of 28 cores):  - SkyLake 8180 (single thread):  \* By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) \** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code are placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is: ``` caffe2/CMakeLists.txt ``` ## Limitations - In this PR, we have only supported the optimization on Linux platform. The support on Windows and MacOS will be enabled as the next step. - We have only optimized the inference use case. Pull Request resolved: pytorch#74596 Approved by: https://github.com/malfet
pytorchmergebot
pushed a commit
that referenced
this pull request
May 5, 2022
Re-landing #68111/#74596 ## Description v0.5 PR of this [RFC](#49444). On the basis of #50256, the below improvements are included: * The [v0.5 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.5) of the oneDNN Graph API is used * The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` `torch.jit.freeze` should be used after tracing (recommended) or scripting a model. ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: * SkyLake 8180 (1 socket of 28 cores):  * SkyLake 8180 (single thread):  * By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) ** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code is placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is in: ``` caffe2/CMakeLists.txt cmake/public/mkldnn.cmake cmake/Modules/FindMKLDNN.cmake ``` ## Limitations * In this PR, we only support Pytorch-oneDNN-Graph integration on Linux platform. Support on Windows and MacOS will be enabled as a next step. * We have only optimized the inference use-case. Pull Request resolved: #76622 Approved by: https://github.com/eellison
facebook-github-bot
pushed a commit
that referenced
this pull request
May 9, 2022
Summary: Re-landing #68111/#74596 ## Description v0.5 PR of this [RFC](#49444). On the basis of #50256, the below improvements are included: * The [v0.5 release branch](https://github.com/oneapi-src/oneDNN/releases/tag/graph-v0.5) of the oneDNN Graph API is used * The fuser now works with the profiling graph executor. We have inserted type check nodes to guard the profiled tensor properties. ### User API: The optimization pass is disabled by default. Users could enable it by: ``` torch.jit.enable_onednn_fusion(True) ``` `torch.jit.freeze` should be used after tracing (recommended) or scripting a model. ### Performance: [pytorch/benchmark](https://github.com/pytorch/benchmark) tool is used to compare the performance: * SkyLake 8180 (1 socket of 28 cores):  * SkyLake 8180 (single thread):  * By mapping hardswish to oneDNN Graph, it’s 8% faster than PyTorch JIT (NNC + OFI) ** We expect performance gain after mapping transpose, contiguous & view to oneDNN graph ops ### Directory structure of the integration code Fuser-related code is placed under: ``` torch/csrc/jit/codegen/onednn/ ``` Optimization pass registration is done in: ``` torch/csrc/jit/passes/onednn_graph_fuser.h ``` CMake for the integration code is in: ``` caffe2/CMakeLists.txt cmake/public/mkldnn.cmake cmake/Modules/FindMKLDNN.cmake ``` ## Limitations * In this PR, we only support Pytorch-oneDNN-Graph integration on Linux platform. Support on Windows and MacOS will be enabled as a next step. * We have only optimized the inference use-case. Pull Request resolved: #76622 Approved by: https://github.com/eellison Test Plan: contbuild & OSS CI, see https://hud.pytorch.org/commit/pytorch/pytorch/4ee29d6033f6ff4e1b75caa265d2646648b6636c Reviewed By: bdhirsh Differential Revision: D36135124 Pulled By: bdhirsh fbshipit-source-id: 37397924ae667669996f47469e98fdab218d71f5
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Since oneDNN graph is under development, this is only a preview of the PR for this RFC.