Skip to content

Bump torch-npu version to 2.5.1.post1.dev20250528 #1242

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.

Sign up for GitHub

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

Jump to bottom
Closed
wants to merge 1 commit into from
Closed

Bump torch-npu version to 2.5.1.post1.dev20250528 #1242

wants to merge 1 commit into from

Conversation

@wxsIcey
Copy link
Collaborator

@wxsIcey wxsIcey commented Jun 16, 2025

What this PR does / why we need it?

Related to #1232

How was this patch tested?

CI passed with new added test.

@github-actions github-actions bot added the documentation Improvements or additions to documentation label Jun 16, 2025
Yikun
Yikun reviewed Jun 16, 2025
View reviewed changes
Copy link
Collaborator

@Yikun Yikun left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems Dockerfile should also be updated?
https://github.com/vllm-project/vllm-ascend/blob/main/Dockerfile#L52
https://github.com/vllm-project/vllm-ascend/blob/main/Dockerfile.openEuler#L49

docs/source/installation.md Outdated
# Install vLLM
git clone --depth 1 --branch |vllm_version| https://github.com/vllm-project/vllm
cd vllm
export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi
Copy link
Collaborator

@Yikun Yikun Jun 16, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
export PIP_EXTRA_INDEX_URL=https://mirrors.huaweicloud.com/ascend/repos/pypi

This should move to vLLM Ascend

Copy link
Collaborator Author

@wxsIcey wxsIcey Jun 16, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

modified

setup.py
pybind11_cmake_path = (subprocess.check_output(
[python_executable, "-m", "pybind11",
"--cmake"]).decode().strip())
"--cmakedir"]).decode().strip())
Copy link
Collaborator

@Yikun Yikun Jun 16, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure why this change is required? How did this work before?

Copy link
Collaborator Author

@wxsIcey wxsIcey Jun 16, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If this is not modified, the following CI exception will be triggered:
https://github.com/vllm-project/vllm-ascend/actions/runs/15676188937/job/44156927753

In the update of pybind11 2.6.0 version, the support for the --cmake parameter was explicitly removed. Instead, it is recommended to use --cmakedir to obtain the CMake configuration file path.
联想截图_20250616190019

@wangxiyuan
Copy link
Collaborator

wangxiyuan commented Jun 16, 2025

LGTM. the unit test is merged just now. Please rebase and make it work as well. 69b817e

@Yikun
Copy link
Collaborator

Yikun commented Jun 16, 2025

BTW, we can add

      - 'setup.py'
      - 'pyproject.toml'
      - 'requirements.txt'
      - 'cmake/**'
      - 'CMakeLists.txt'
      - 'csrc/**'

to
https://github.com/vllm-project/vllm-ascend/blob/main/.github/workflows/image_ubuntu.yml#L18
https://github.com/vllm-project/vllm-ascend/blob/main/.github/workflows/image_openeuler.yml#L18

@MengqingCao
Copy link
Collaborator

MengqingCao commented Jun 16, 2025

Plz backport this to branch 0.9.1-dev, too

@wxsIcey
Copy link
Collaborator Author

wxsIcey commented Jun 16, 2025

BTW, we can add

      - 'setup.py'
      - 'pyproject.toml'
      - 'requirements.txt'
      - 'cmake/**'
      - 'CMakeLists.txt'
      - 'csrc/**'

to https://github.com/vllm-project/vllm-ascend/blob/main/.github/workflows/image_ubuntu.yml#L18 https://github.com/vllm-project/vllm-ascend/blob/main/.github/workflows/image_openeuler.yml#L18

modified

@wxsIcey
Copy link
Collaborator Author

wxsIcey commented Jun 16, 2025

Plz backport this to branch 0.9.1-dev, too

ok

1 similar comment
@wxsIcey
Copy link
Collaborator Author

wxsIcey commented Jun 16, 2025

Plz backport this to branch 0.9.1-dev, too

ok

@github-actions github-actions bot added module:tests module:ops merge-conflicts and removed documentation Improvements or additions to documentation ci/build labels Jun 16, 2025
@github-actions
Copy link

github-actions bot commented Jun 16, 2025

This pull request has conflicts, please resolve those before we can evaluate the pull request.

@Yikun
Bump torch-npu version
8489e38 
Signed-off-by: Yikun Jiang <yikunkero@gmail.com>
@github-actions github-actions bot added documentation Improvements or additions to documentation ci/build and removed module:tests module:ops merge-conflicts labels Jun 16, 2025
@Yikun Yikun mentioned this pull request Jun 16, 2025
Merged
@MengqingCao MengqingCao mentioned this pull request Jun 16, 2025
Merged
@Yikun Yikun closed this in 96fa7ff Jun 16, 2025
ganyi1996ppo pushed a commit that referenced this pull request Jun 17, 2025
@MengqingCao @wxsIcey
[v0.9.1][DP][V1] Fix rank set in DP scenario & Bump torch-npu version…
d798125 
… to 2.5.1.post1.dev20250528 (#1247)

### What this PR does / why we need it?
Cherry-pick form #1235

1. Fix rank set in DP scenario. The new poc version of torch-npu support
setting `ASCEND_RT_VISIBLE_DEVICES` dynamically, thus we could use the
rank set in `DPEngineCoreProc` directly instead of calculating local
rank across dp by hand in the patched `_init_data_parallel`

Closes: #1170

2. Bump torch-npu version to 2.5.1.post1.dev20250528

Closes: #1242
Closes: #1232

### How was this patch tested?
CI passed with new added test.

---------

Signed-off-by: Icey <1790571317@qq.com>
Signed-off-by: MengqingCao <cmq0113@163.com>
Co-authored-by: Icey <1790571317@qq.com>
shiyuan680 pushed a commit to raindaywhu/vllm-ascend that referenced this pull request Jul 7, 2025
@MengqingCao @wxsIcey
[DP][V1] Fix rank set in DP scenario & Bump torch-npu version to 2.5.…
cacde11 
…1.post1.dev20250528 (vllm-project#1235)

1. Fix rank set in DP scenario. The new poc version of torch-npu support
setting `ASCEND_RT_VISIBLE_DEVICES` dynamically, thus we could use the
rank set in `DPEngineCoreProc` directly instead of calculating local
rank across dp by hand in the patched `_init_data_parallel`

Closes: vllm-project#1170

2. Bump torch-npu version to 2.5.1.post1.dev20250528

Closes: vllm-project#1242
Closes: vllm-project#1232

CI passed with new added test.

---------

Signed-off-by: MengqingCao <cmq0113@163.com>
Signed-off-by: Icey <1790571317@qq.com>
Co-authored-by: Icey <1790571317@qq.com>
@Yikun Yikun mentioned this pull request Jul 28, 2025
Closed
chopper0126 pushed a commit to chopper0126/vllm-ascend that referenced this pull request Oct 16, 2025
@MengqingCao @wxsIcey
[DP][V1] Fix rank set in DP scenario & Bump torch-npu version to 2.5.…
14ebc06 
…1.post1.dev20250528 (vllm-project#1235)

### What this PR does / why we need it?
1. Fix rank set in DP scenario. The new poc version of torch-npu support
setting `ASCEND_RT_VISIBLE_DEVICES` dynamically, thus we could use the
rank set in `DPEngineCoreProc` directly instead of calculating local
rank across dp by hand in the patched `_init_data_parallel`

Closes: vllm-project#1170

2. Bump torch-npu version to 2.5.1.post1.dev20250528

Closes: vllm-project#1242
Closes: vllm-project#1232


### How was this patch tested?
CI passed with new added test.

---------

Signed-off-by: MengqingCao <cmq0113@163.com>
Signed-off-by: Icey <1790571317@qq.com>
Co-authored-by: Icey <1790571317@qq.com>
Angazenn pushed a commit to Angazenn/vllm-ascend that referenced this pull request Oct 21, 2025
@MengqingCao @wxsIcey @Angazenn
[DP][V1] Fix rank set in DP scenario & Bump torch-npu version to 2.5.…
ff15514 
…1.post1.dev20250528 (vllm-project#1235)

### What this PR does / why we need it?
1. Fix rank set in DP scenario. The new poc version of torch-npu support
setting `ASCEND_RT_VISIBLE_DEVICES` dynamically, thus we could use the
rank set in `DPEngineCoreProc` directly instead of calculating local
rank across dp by hand in the patched `_init_data_parallel`

Closes: vllm-project#1170

2. Bump torch-npu version to 2.5.1.post1.dev20250528

Closes: vllm-project#1242
Closes: vllm-project#1232


### How was this patch tested?
CI passed with new added test.

---------

Signed-off-by: MengqingCao <cmq0113@163.com>
Signed-off-by: Icey <1790571317@qq.com>
Co-authored-by: Icey <1790571317@qq.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@Yikun Yikun Yikun left review comments

@wangxiyuan wangxiyuan wangxiyuan approved these changes

Assignees

No one assigned

Labels

ci/build documentation Improvements or additions to documentation

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@wxsIcey @wangxiyuan @Yikun @MengqingCao