Skip to content
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

[cudnn] Add library path #18005

Merged
merged 4 commits into from
May 24, 2021

Conversation

NancyLi1013
Copy link
Contributor

@NancyLi1013 NancyLi1013 commented May 19, 2021

Describe the pull request

  • What does your PR fix?

Fixes #17680

Update endif as endif() in FindCUDNN.cmake.


Install vcpkg-cmake-wrapper.cmake to share/${PORT} to enable cudnn found.

Additional fixes:
Currently, cuddn needs to download from official website manually. The libs in the latest version 8.2.0 are listed in cuda/lib/x64 folder. So I add cuda/lib/x64 to the library path to enable the library found.

image

Related #17377

Note: No feature needs to test.

@NancyLi1013 NancyLi1013 added category:port-bug The issue is with a library, which is something the port should already support info:internal This PR or Issue was filed by the vcpkg team. labels May 19, 2021
@cenit
Copy link
Contributor

cenit commented May 19, 2021

is it really necessary? I can find CUDNN even with current module 🤔
in any case it's no harm

@NancyLi1013
Copy link
Contributor Author

On my machine, it doesn't work without the additional path cuda/lib/x64.

Starting package 1/28: cudnn:x64-windows
Building package cudnn[core]:x64-windows...
-- Found CUDA V11.3.58
-- CUDNN_INCLUDE_DIR: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.3/cuda/include
-- CUDNN_LIBRARY: CUDNN_LIBRARY-NOTFOUND
CMake Error at ports/cudnn/portfile.cmake:69 (message):
  Please download CUDNN from official sources (such as
  https://developer.nvidia.com/rdp/cudnn-download ) and extract the zip into
  your CUDA_TOOLKIT_ROOT (C:/Program Files/NVIDIA GPU Computing
  Toolkit/CUDA/v11.3).  (For example: tar.exe -xvf
  cudnn-11.2-windows-x64-v8.1.1.33.zip --strip 1 --directory "C:/Program
  Files/NVIDIA GPU Computing Toolkit/CUDA/v11.3"
Call Stack (most recent call first):
  scripts/ports.cmake:141 (include)


Error: Building package cudnn:x64-windows failed with: BUILD_FAILED
Please ensure you're using the latest portfiles with `.\vcpkg update`, then
submit an issue at https://github.com/Microsoft/vcpkg/issues including:
  Package: cudnn:x64-windows
  Vcpkg version: 2021-05-05-9f849c4c43e50d1b16186ae76681c27b0c1be9d9

As you can see, it can only find headers in cuda/include, but library cannot be found.

@NancyLi1013
Copy link
Contributor Author

/azp run

@azure-pipelines
Copy link

Azure Pipelines successfully started running 1 pipeline(s).

@JackBoosY JackBoosY added the info:reviewed Pull Request changes follow basic guidelines label May 20, 2021
@ras0219-msft ras0219-msft merged commit fd5bc0c into microsoft:master May 24, 2021
@ras0219-msft
Copy link
Contributor

This LGTM, thanks!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
category:port-bug The issue is with a library, which is something the port should already support info:internal This PR or Issue was filed by the vcpkg team. info:reviewed Pull Request changes follow basic guidelines
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[<darknet>] build failure
4 participants