OSS MPZCH CUDA kernel in FBGEMM#4214

Closed

lizhouyu wants to merge 1 commit intopytorch:mainfrom

lizhouyu:export-D75505020

Contributor

lizhouyu commented May 29, 2025

Summary:
Opensource FBGEMM CUDA Kernel for MPZCH feature

Major changes

Create a folder named faster_hash under the fbgemm/fbgemmgpu/src folder.
Copy the following files to the created folder from fbsource/fbcode/caffe2/torch/fb/retrieval
- faster_hash.cpp
- faster_hash.cu
- common_utils.cuh
Revise the faster_hash.cpp
- Change namespace fb to namespace fbgemm_gpu.
- Comment out using namespace torch::fb::turborec;
- Change TORCH_LIBRARY_IMPL(fb, ...) to TORCH_LIBRARY_IMPL(fbgemm, ...)
- Fix namespace calling issue due to the namespace change.
Revise the faster_hash.cu
- Change namespace fb to namespace fbgemm_gpu.
- Change TORCH_LIBRARY_IMPL(fb, ...) to TORCH_LIBRARY_IMPL(fbgemm, ...)
- Fix namespace calling issue due to the namespace change.
Revise the common_utils.cuh file
- Change namespace fb to namespace fbgemm_gpu.
Add a BUCK file to compile the cpp and cuda library.
Copy the faster_hash_test.py file to the fbgemm/fbgemm_gpu/test folder.
Add a section in the BUCK file under the test folder for python_unittest of faster_hash_test.
In the faster_hash_test.py file
- Load the faster_hash related libraries with torch.ops.load API.
- Replace all the torch.ops.fb to torch.ops.fbgemm.
- Following other test files to add opensource and gpu availability check.

Questions

After refactorying, the API calls torch.ops.create_zch_buffer, torch.ops.zero_collision_hash, torch.ops.fbgemm.zero_collision_hash, and torch.ops.fbgemm.create_zch_buffer are all valid, while torch.ops.create_zch_buffer and torch.ops.zero_collision_hash may incur certain parameter mismatches. How to resolve this issue and disable the API calls without fbgemm?
How to integrate the refactoryed library into fbgemm so the test can call something like from fbgemm_gpu import create_zch_buffer, zero_collision_hash?

Differential Revision: D75505020

netlify bot commented May 29, 2025 •

edited

Loading

✅ Deploy Preview for pytorch-fbgemm-docs ready!

Name	Link
🔨 Latest commit	`68c8d55`
🔍 Latest deploy log	https://app.netlify.com/projects/pytorch-fbgemm-docs/deploys/684299a5d6d4040008163e3f
😎 Deploy Preview	https://deploy-preview-4214--pytorch-fbgemm-docs.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

To edit notification comments on pull requests, go to your Netlify project configuration.

facebook-github-bot added the cla signed label

Contributor

facebook-github-bot commented May 29, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

facebook-github-bot added the fb-exported label

Contributor

facebook-github-bot commented May 29, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from 684b319 to a29cd2a Compare

May 29, 2025 20:46

lizhouyu added a commit to lizhouyu/FBGEMM-1 that referenced this pull request


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

a29cd2a

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.

### Questions
- After refactorying, the API calls `torch.ops.create_zch_buffer`, `torch.ops.zero_collision_hash`, `torch.ops.fbgemm.zero_collision_hash`, and `torch.ops.fbgemm.create_zch_buffer` are all valid, while `torch.ops.create_zch_buffer` and `torch.ops.zero_collision_hash` may incur certain parameter mismatches. How to resolve this issue and disable the API calls without `fbgemm`?
- How to integrate the refactoryed library into fbgemm so the test can call something like `from fbgemm_gpu import create_zch_buffer, zero_collision_hash`?

Differential Revision: D75505020

Contributor

facebook-github-bot commented May 29, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu added a commit to lizhouyu/FBGEMM-1 that referenced this pull request


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

2c3846e

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.

### Questions
- After refactorying, the API calls `torch.ops.create_zch_buffer`, `torch.ops.zero_collision_hash`, `torch.ops.fbgemm.zero_collision_hash`, and `torch.ops.fbgemm.create_zch_buffer` are all valid, while `torch.ops.create_zch_buffer` and `torch.ops.zero_collision_hash` may incur certain parameter mismatches. How to resolve this issue and disable the API calls without `fbgemm`?
- How to integrate the refactoryed library into fbgemm so the test can call something like `from fbgemm_gpu import create_zch_buffer, zero_collision_hash`?

Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from a29cd2a to 2c3846e Compare

May 29, 2025 20:56

Contributor

facebook-github-bot commented May 29, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from 2c3846e to 2b54f80 Compare

May 29, 2025 21:07

lizhouyu added a commit to lizhouyu/FBGEMM-1 that referenced this pull request


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

2b54f80

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.

### Questions
- After refactorying, the API calls `torch.ops.create_zch_buffer`, `torch.ops.zero_collision_hash`, `torch.ops.fbgemm.zero_collision_hash`, and `torch.ops.fbgemm.create_zch_buffer` are all valid, while `torch.ops.create_zch_buffer` and `torch.ops.zero_collision_hash` may incur certain parameter mismatches. How to resolve this issue and disable the API calls without `fbgemm`?
- How to integrate the refactoryed library into fbgemm so the test can call something like `from fbgemm_gpu import create_zch_buffer, zero_collision_hash`?

Differential Revision: D75505020

Contributor

facebook-github-bot commented May 29, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu added a commit to lizhouyu/FBGEMM-1 that referenced this pull request


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

d94acaa

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.

### Questions
- After refactorying, the API calls `torch.ops.create_zch_buffer`, `torch.ops.zero_collision_hash`, `torch.ops.fbgemm.zero_collision_hash`, and `torch.ops.fbgemm.create_zch_buffer` are all valid, while `torch.ops.create_zch_buffer` and `torch.ops.zero_collision_hash` may incur certain parameter mismatches. How to resolve this issue and disable the API calls without `fbgemm`?
- How to integrate the refactoryed library into fbgemm so the test can call something like `from fbgemm_gpu import create_zch_buffer, zero_collision_hash`?

Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from 2b54f80 to d94acaa Compare

May 29, 2025 21:15

Contributor

facebook-github-bot commented May 30, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from d94acaa to 3d36968 Compare

May 30, 2025 17:56

lizhouyu added a commit to lizhouyu/FBGEMM-1 that referenced this pull request


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

3d36968

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.

### Questions
- After refactorying, the API calls `torch.ops.create_zch_buffer`, `torch.ops.zero_collision_hash`, `torch.ops.fbgemm.zero_collision_hash`, and `torch.ops.fbgemm.create_zch_buffer` are all valid, while `torch.ops.create_zch_buffer` and `torch.ops.zero_collision_hash` may incur certain parameter mismatches. How to resolve this issue and disable the API calls without `fbgemm`?
- How to integrate the refactoryed library into fbgemm so the test can call something like `from fbgemm_gpu import create_zch_buffer, zero_collision_hash`?

Differential Revision: D75505020

Contributor

facebook-github-bot commented May 30, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu added a commit to lizhouyu/FBGEMM-1 that referenced this pull request


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

a273543

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.

### Questions
- After refactorying, the API calls `torch.ops.create_zch_buffer`, `torch.ops.zero_collision_hash`, `torch.ops.fbgemm.zero_collision_hash`, and `torch.ops.fbgemm.create_zch_buffer` are all valid, while `torch.ops.create_zch_buffer` and `torch.ops.zero_collision_hash` may incur certain parameter mismatches. How to resolve this issue and disable the API calls without `fbgemm`?
- How to integrate the refactoryed library into fbgemm so the test can call something like `from fbgemm_gpu import create_zch_buffer, zero_collision_hash`?

Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from 3d36968 to a273543 Compare

May 30, 2025 18:02

Contributor

facebook-github-bot commented May 30, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu added a commit to lizhouyu/FBGEMM-1 that referenced this pull request


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

d203230

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.

### Questions
- After refactorying, the API calls `torch.ops.create_zch_buffer`, `torch.ops.zero_collision_hash`, `torch.ops.fbgemm.zero_collision_hash`, and `torch.ops.fbgemm.create_zch_buffer` are all valid, while `torch.ops.create_zch_buffer` and `torch.ops.zero_collision_hash` may incur certain parameter mismatches. How to resolve this issue and disable the API calls without `fbgemm`?
- How to integrate the refactoryed library into fbgemm so the test can call something like `from fbgemm_gpu import create_zch_buffer, zero_collision_hash`?

Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from a273543 to d203230 Compare

May 30, 2025 18:07

Contributor

facebook-github-bot commented May 30, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from d203230 to 4490476 Compare

May 30, 2025 18:45

lizhouyu added a commit to lizhouyu/FBGEMM-1 that referenced this pull request


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.

### Questions
- After refactorying, the API calls `torch.ops.create_zch_buffer`, `torch.ops.zero_collision_hash`, `torch.ops.fbgemm.zero_collision_hash`, and `torch.ops.fbgemm.create_zch_buffer` are all valid, while `torch.ops.create_zch_buffer` and `torch.ops.zero_collision_hash` may incur certain parameter mismatches. How to resolve this issue and disable the API calls without `fbgemm`?
- How to integrate the refactoryed library into fbgemm so the test can call something like `from fbgemm_gpu import create_zch_buffer, zero_collision_hash`?

Differential Revision: D75505020

Contributor

facebook-github-bot commented May 30, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

lizhouyu force-pushed the export-D75505020 branch from 4490476 to e35f85d Compare

May 30, 2025 18:48

Contributor

facebook-github-bot commented Jun 3, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

26 similar comments

Contributor

facebook-github-bot commented Jun 4, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 4, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 4, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 4, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 4, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 4, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 4, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 5, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 5, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 5, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 5, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020


          OSS MPZCH CUDA kernel in FBGEMM (pytorch#4214)

68c8d55

Summary:
Pull Request resolved: pytorch#4214

X-link: facebookresearch/FBGEMM#1290

Opensource FBGEMM CUDA Kernel for MPZCH feature

### Major changes
- Create a folder named `faster_hash` under the `fbgemm/fbgemmgpu/src` folder.
- Copy the following files to the created folder from `fbsource/fbcode/caffe2/torch/fb/retrieval`
  - faster_hash.cpp
  - faster_hash.cu
  - common_utils.cuh
- Revise the `faster_hash.cpp`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Comment out `using namespace torch::fb::turborec;`
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `faster_hash.cu`
  - Change `namespace fb` to `namespace fbgemm_gpu`.
  - Change `TORCH_LIBRARY_IMPL(fb, ...)` to `TORCH_LIBRARY_IMPL(fbgemm, ...)`
  - Fix namespace calling issue due to the namespace change.
- Revise the `common_utils.cuh` file
  - Change `namespace fb` to `namespace fbgemm_gpu`.
- Add a BUCK file to compile the cpp and cuda library.
- Copy the `faster_hash_test.py` file to the `fbgemm/fbgemm_gpu/test` folder.
- Add a section in the BUCK file under the `test` folder for `python_unittest` of `faster_hash_test`.
- In the `faster_hash_test.py` file
  - Load the `faster_hash` related libraries with `torch.ops.load` API.
  - Replace all the `torch.ops.fb` to `torch.ops.fbgemm`.
  - Following other test files to add `opensource` and `gpu availability` check.
- OSS the `murmur_hash3` function
  - Write wrappers for the `murmur_hash3` function in `faster_hash.cu` and `faster_hash.cpp` files, register wrapper functions and expose to external calls.
  - Add a test for the `murmur_hash3` function to validate on `CPU` and `GPU` the hashed values are identical for the same input value.

Reviewed By: ionuthristodorescu, spcyppt

Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request was exported from Phabricator. Differential Revision: D75505020

Contributor

facebook-github-bot commented Jun 6, 2025

This pull request has been merged in cf90aac.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

category:new cla signed fb-exported feature:gemm feature:utils Merged