CUDA backend integration for TFP #134
Closed
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
added 4 commits
June 9, 2020 07:53
Summary: Pull Request resolved: #131 This diff does the two following thing: 1. Implement a encoder for LongTensor that embed fixed-point arithmetic into floating-point arithmetic. It enables us to utilize existing CUDA kernel for floating points to compute matrix multiplication for LongTensor. 2. Add a new test file test_cuda.py for future CUDA related testing The encoder currently only supports matmul. Conv2d will be supported in future diffs. The API design is based on: D21860688 Differential Revision: D21882865 fbshipit-source-id: 1c46741a81ca5f386ecbb974dc937bbe759a1af2
Summary: Support CUDA for conv1d, conv2d, conv1d_transpose, conv2d_transpose Differential Revision: D21885098 fbshipit-source-id: 74f7b59ebc4e7e57519f0a8a38b81dca6d363c5c
Summary: Add comprehensive test case for TestCUDA that would enable us to check the correctness of cuda integration Differential Revision: D21911937 fbshipit-source-id: 0b65a38f0d838d91cb58f20cbf525ea3ac7f5504
Summary: This diff does the following thing: 1. Integrate torch.device into CrypTen. If a user provide a CUDA tensor as input, CrypTen will detect that and run the backend on a GPU. 2. Apply cuda_patches to beaver.py. As suggested by lvdmaaten, this design could make our code error prone if mpc.py and arithmetic.py is littered with if else statements. After running tests on the backend though, we discovered that only a few places related to arithmetic_function that we need to use cuda_patches. The rest of the codebase work smoothly without much changes. It is probably fine to use cuda_patches as a temporary hack. What do you think? lvdmaaten knottb This diff only integrate CUDA support for TFP. Future diff will extend it to TTP. Differential Revision: D21952814 fbshipit-source-id: 0811c4848ae9fd7a115df82ef998a9f5202938bf
facebook-github-bot
added
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D21952814 |
tanjuntao
pushed a commit
to tanjuntao/CrypTen
that referenced
this pull request
Nov 27, 2023
Summary: ## Types of changes - [ ] Bug fix (non-breaking change which fixes an issue) - [ ] New feature (non-breaking change which adds functionality) - [ ] Breaking change (fix or feature that would cause existing functionality to change) - [x] Docs change / refactoring / dependency upgrade ## Motivation and Context / Related issue Cleaning up and restructuring ## How Has This Been Tested (if it applies) Tested locally on macbook ## Checklist - [x] The documentation is up-to-date with the changes I made. - [x] I have read the **CONTRIBUTING** document and completed the CLA (see **CONTRIBUTING**). - [ ] All tests passed, and additional code has been covered with new tests. Pull Request resolved: fairinternal/CrypTen#134 Differential Revision: D17850045 Pulled By: vshobha fbshipit-source-id: 5f8e13d3031bc0212d6db10660df2f791b9aa114
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.
Summary:
This diff does the following thing:
Integrate torch.device into CrypTen. If a user provide a CUDA tensor as input, CrypTen will detect that and run the backend on a GPU.
Apply cuda_patches to beaver.py. As suggested by lvdmaaten, this design could make our code error prone if mpc.py and arithmetic.py is littered with if else statements. After running tests on the backend though, we discovered that only a few places related to arithmetic_function that we need to use cuda_patches. The rest of the codebase work smoothly without much changes. It is probably fine to use cuda_patches as a temporary hack. What do you think? lvdmaaten knottb
This diff only integrate CUDA support for TFP. Future diff will extend it to TTP.
Differential Revision: D21952814