jimver / cuda-toolkit

GitHub Action to install CUDA

cuda-toolkit

This action installs the NVIDIA® CUDA® Toolkit on the system. It adds the cuda install location as CUDA_PATH to GITHUB_ENV so you can access the CUDA install location in subsequent steps. CUDA_PATH/bin is added to GITHUB_PATH so you can use commands such as nvcc directly in subsequent steps. Right now both windows-2019 and ubuntu-20.04 runners have been tested to work successfully.

Inputs

cuda

Optional The CUDA version to install. View src/link/windows-links.ts and src/link/linux-links.ts for available versions.

Default: '12.5.0'.

sub-packages

NOTE: On Linux this only works with the 'network' method view details

Optional If set, only the specified CUDA subpackages will be installed. Only installs specified subpackages, must be in the form of a JSON array. For example, if you only want to install nvcc and visual studio integration: '["nvcc", "visual_studio_integration"]' (double quotes required)

Default: '[]'.

non-cuda-sub-packages

NOTE: This only works on Linux with the 'network' method view details

Optional If set, only the specified CUDA subpackages will be installed without prepending the "cuda-" prefix. Only installs specified subpackages without prepending the "cuda-" prefix, must be in the form of a JSON array. For example, if you only want to install libcublas and libcufft: '["libcublas", "libcufft"]' (double quotes required)

Default: '[]'.

method

Optional Installation method, can be either 'local' or 'network'.

• 'local' downloads the entire installer with all packages and runs that (you can still only install certain packages with sub-packages on Windows).
• 'network' downloads a smaller executable which only downloads necessary packages which you can define in sub-packages.

Default: 'local'.

linux-local-args

Optional (For Linux and 'local' method only) override arguments for the Linux .run installer. For example if you don't want samples use '["--toolkit"]' (double quotes required) See the Nvidia Docs for available options. Note that the --silent option is already always added by the action itself.

Default: '["--toolkit", "--samples"]'.

log-file-suffix

Required with matrix builds

Add suffix to the log file name which gets uploaded as an artifact. This has to be set when running a matrix build. The log file already contains the OS type (Linux/Windows) and install method (local/network) but it is not aware of other matrix variables, so add those here.

For example if you use multiple linux distros:

jobs:
  CI:
    strategy:
      matrix:
        os: [ubuntu-22.04, ubuntu-20.04]
    runs-on: ${{ matrix.os }}
    steps:
    - uses: Jimver/cuda-toolkit@master
      id: cuda-toolkit
      with:
        log-file-suffix: '${{matrix.os}}.txt'

Default: 'log.txt'

Outputs

cuda

The cuda version installed (same as cuda from input).

CUDA_PATH

The path where cuda is installed (same as CUDA_PATH in GITHUB_ENV).

Example usage

steps:
- uses: Jimver/[email protected]
  id: cuda-toolkit
  with:
    cuda: '12.5.0'

- run: echo "Installed cuda version is: ${{steps.cuda-toolkit.outputs.cuda}}"

- run: echo "Cuda install location: ${{steps.cuda-toolkit.outputs.CUDA_PATH}}"

- run: nvcc -V

cuda-toolkit's Issues

Add cache support

I noticed that it took about 3 min to install cuda toolkit everytime:

/usr/bin/sudo /opt/hostedtoolcache/cuda_installer-linux/11.2.2/x64/cuda_installer_11.2.2.run --silent --toolkit --samples

Is it possible to cache the installed files to speed up this process? Thanks!

Missing CUDA 11.8?

Hi, we use your github action in NNPOps, we are trying to write workflows for testing with CUDA 11.8 openmm/NNPOps#94 , would you consider adding it?
Thanks.

cuda 12.4.1 gives 404 error on ubuntu 24.04

  - uses: Jimver/[email protected]
    if: contains(matrix.args, 'cuda')
    id: cuda-toolkit
    with:
        cuda: '12.4.1`
        method: 'network'

/usr/bin/lsb_release -sr
24.04
/usr/bin/wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.0-1_all.deb -O cuda_keyring.deb
--2024-07-15 00:39:18--  https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/x86_64/cuda-keyring_1.0-1_all.deb
Resolving developer.download.nvidia.com (developer.download.nvidia.com)... 152.195.19.[14](https://github.com/thewh1teagle/vibe/actions/runs/9931974881/job/27432643855#step:5:15)2
Connecting to developer.download.nvidia.com (developer.download.nvidia.com)|[15](https://github.com/thewh1teagle/vibe/actions/runs/9931974881/job/27432643855#step:5:16)2.195.19.142|:443... connected.
HTTP request sent, awaiting response... 404 Not Found
2024-07-15 00:39:19 ERROR 404: Not Found.

installing CUDA on ubuntu with cuda-toolkit broken due to NVIDIA signing key change :-(

cuda-toolkit has recently started failing with an error like:

Reading package lists...
[59](https://github.com/mmp/pbrt-v4/runs/6240711119?check_suite_focus=true#step:2:59)
W: GPG error: http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64  InRelease: The following signatures couldn't be verified because the public key is not available: NO_PUBKEY A4B469963BF863CC
[60](https://github.com/mmp/pbrt-v4/runs/6240711119?check_suite_focus=true#step:2:60)
E: The repository 'http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64  InRelease' is not signed.
[61](https://github.com/mmp/pbrt-v4/runs/6240711119?check_suite_focus=true#step:2:61)
Error: Error: The process '/usr/bin/sudo' failed with exit code 100

due to NVIDIA changing their signing keys. @jandom's PR #26 should fix this, but I thought I'd file a bug for visibility.

Some linux cuda subpackages don't have a cuda version suffix: particularly `drivers`

I cannot seem to be able to install a drivers package, e.g. cuda-drivers_520.61.05-1_amd64.deb, there is always a cuda version appended to the end.

E.g.

    - uses: Jimver/[email protected]
      id: cuda-toolkit
      with:
        method: network
        sub-packages: '[ "cudart", "cudart-dev", "nvrtc", "nvrtc-dev" ]'
        non-cuda-sub-packages: '[ "cuda-drivers_520.61.05" ]'

leads to: E: Unable to locate package cuda-drivers_520.61.05-12-4.

(I assumed non-cuda-sub-packages wouldn't post-process the package name.)

Is there any way to manually specify a version with its link?

I need to use CUDA 12.5.1 which includes an important fix (see https://forums.developer.nvidia.com/t/cudacompile-nvcc-error-cudafe-died-with-status-0xc0000409/260651/15). Is there any way to specify that in the action's params? Or would I have to fork it?

subPackages do not work on Ubuntu

Hi,

1. documentation seems to be wrong regarding the subPackages format. Valid format is subPackages: '["nvcc", "cublas_dev"]'
2. subPackages do not work as expected on Ubuntu (16.04, probably others too). Since Ubuntu .run installer does not support them (or at least that's what I understand from the documentation), using subPackages results in failed build:

/usr/bin/sudo /opt/hostedtoolcache/cuda_installer-linux/10.2.89/x64/cuda_installer_10.2.89.run --silent --toolkit --samples nvcc cublas_dev
Unknown option: nvcc

ARM versions

It'd be useful to add ARM versions. Maybe something like:

with:
  cuda: '12.1.1'
  arch: ARM
  os: Ubuntu-20.04

would fetch:
https://developer.download.nvidia.com/compute/cuda/12.1.1/local_installers/cuda_12.1.1_530.30.02_linux_sbsa.run

Remove or make silent flag configurable

Hello, often this action gets stuck on Windows, I would like to see what it is doing when it does.

Upgrade action to use Node.js 16

When I use this action, I get the following deprecation warnings from GitHub:

Node.js 12 actions are deprecated. For more information see: https://github.blog/changelog/2022-09-22-github-actions-all-actions-will-begin-running-on-node16-instead-of-node12/. Please update the following actions to use Node.js 16: Jimver/cuda-toolkit

I'm not sure if that's as simple as changing the version here: 🤷‍♂️

cuda-toolkit/package.json

Line 26 in ae49935

"node": "12.22.7"

Thanks for the great tool!

Support windows-2022

Github started switching Actions using windows-latest to use windows-2022 which it doesn't look like is supported by this action. My cuda build began failing unable to find the cuda toolkit here.

Note this could also potentially be an issue with how my project is configured, but since I have done manual builds using on my project using VS2022, I'm not sure.

Dependency error installing library that uses cudatoolkit

Trying to install NewsSentiment[cuda] on Ubuntu in a fresh mamba environment with Python 3.8. This yields an error

newssentiment[cuda] 1.1.25 depends on cudatoolkit==10.1; extra == "cuda"

which repeats through versions down to newssentiment[cuda] 1.0.0. If I install cudatoolkit 10.1 using conda, I get the same errors.

I reported the issue over there, asking the dev if he's certain the cuda support is working. He said "Yep, cudatoolkit should be working fine. looks like an error related to cudatoolkit, so ill suggest to open an issue over there."

So that's what I'm doing. Any idea how to fix?

Add MacOS support

Please release a Mac version.

Some linux subpackages dont have a `cuda-` prefix

I want to install [libcublas-dev] which doesn't have a cuda- prefix. See here.

Could the action be changed to not always include the cuda- prefix on windows?

Getting a warning because of Node.js version used

When using the action, I'm getting a warning in my runner:

build
Node.js 16 actions are deprecated. Please update the following actions to use Node.js 20: Jimver/[email protected]. For more information see: https://github.blog/changelog/2023-09-22-github-actions-transitioning-from-node-16-to-node-20/.

It might just be a question of using actions/checkout@v4 instead of actions/checkout@v3/

Doesn't work on Rocky Linux 8 or CentOS 7

I get this error:
Error: Unable to locate executable file: lsb_release
when it's trying to do this:

/usr/bin/docker exec  e293caa31043d6a32d7423f647ae3185b4671dc6abaf24070a41bf7a4e0c77cb sh -c "cat /etc/*release | grep ^ID"

Maybe check for /etc/os-release and if it exists, source it to get its variables instead there?

Bug for 12.1

Hey,

since I changed to CUDA 12.1 I sometimes get the following error on some runs. It could be unrelated to 12.1 but I noticed it now first.

Run Jimver/[email protected]
  with:
    cuda: 1[2](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:2).1.0
    sub-packages: []
    method: local
    linux-local-args: ["--toolkit", "--samples"]
    use-github-cache: true
  env:
    INPUT_RUN_POST: true
    CONDA: /usr/share/miniconda[3](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:3)
    CONDA_PKGS_DIR: /home/runner/conda_pkgs_dir
/usr/bin/tar --posix -cf cache.tzst --exclude cache.tzst -P -C /home/runner/work/SplinePSF/SplinePSF --files-from manifest.txt --use-compress-program zstdmt
Failed to save: Unable to reserve cache with key cuda_installer-linux-5.15.0-10[4](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:4)0-azure-12.1.0, another job may be creating this cache. More details: Cache already exists. Scope: refs/heads/dev_fix_hopper_ci, Key: cuda_installer-linux-[5](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:5).15.0-1040-azure-12.1.0, Version: 4bfd4[6](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:6)a3233f39e[7](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:7)afb92a41e0e6a5d43d677cf4e3f9feca[8](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:8)11a22308a54230c
/usr/bin/sudo /opt/hostedtoolcache/cuda_installer-linux/12.1.0/x64/cuda_installer-linux-5.15.0-1040-azure_12.1.0.run --silent --toolkit --samples
terminate called after throwing an instance of 'boost::filesystem::filesystem_error'
  what():  boost::filesystem::copy_file: No such file or directory: "./builds/cuda_cupti/extras/CUPTI/doc/Cupti/structCUpti__ActivityMemcpy3.html", "/usr/local/cuda-12.1/extras/CUPTI/doc/Cupti/structCUpti__ActivityMemcpy3.html"
Aborted (core dumped)
Warning: Error during installation: Error: The process '/usr/bin/sudo' failed with exit code 134
Starting artifact upload
For more detailed logs during the artifact upload process, enable step-debugging: https://docs.github.com/actions/monitoring-and-troubleshooting-workflows/enabling-debug-logging#enabling-step-debug-logging
Artifact name is valid!
Container for artifact "install-log" successfully created. Starting upload of file(s)
Total size of all the files uploaded is 10[9](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:9)13 bytes
File upload process has finished. Finalizing the artifact upload
Artifact has been finalized. All files have been successfully uploaded!

The raw size of all the files that were specified for upload is 172032 bytes
The size of all the files that were uploaded is [10](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:10)9[13](https://github.com/Haydnspass/SplinePSF/actions/runs/5344931866/jobs/9689924701#step:4:14) bytes. This takes into account any gzip compression used to reduce the upload size, time and storage

Note: The size of downloaded zips can differ significantly from the reported size. For more information see: https://github.com/actions/upload-artifact#zipped-artifact-downloads 

Error: Error: The process '/usr/bin/sudo' failed with exit code 134

any idea what to conclude from that? The workflow for this run is here: https://github.com/Haydnspass/SplinePSF/blob/0ab69d9722f0abce3de11947a5bccf4946341ba9/.github/workflows/build_upload_test.yaml

Add CUDA versions

Hi,

What is the procedure to add new CUDA versions? For example. if I wanted to add the links to version 11.1.0, would I simply edit the two files under src/links ?

Thanks

Slow setup times on windows

Setting up cuda-toolkit on a normal windows github actions runner can take >>15 min. I tried to speed this up by using the network installer and specify only the sub-packages i actually need but that also doesn't make any difference. The same setup on linux takes ~2:30 min also via the network method. Am i doing something wrong? What is the recommended way to install only specific packages on windows?

[Windows] fatal error C1083: Cannot open include file: 'cuda_runtime.h': No such file or directory

Hi @Jimver , when trying to compile a basic CUDA programs in Windows Github Actions, nvcc complains about being unable to find the cuda_runtime.h:
https://github.com/zalo/matmul_cuda/runs/7678959035?check_suite_focus=true#step:10:9

Perhaps there are some paths that are not getting applied in time?

Also, it can't normally find cl.exe, but that can be resolved by adding:

-ccbin "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Tools\MSVC\14.32.31326\bin\HostX64\x64"

which is just the resolved path of:

"%VCToolsInstallDir%\bin\Hostx64\x64"

Problems with python torch dependency

When I execute main.exe which I have built with github actions and cx_Freeze i am getting the following error:

Could not locate cudnn_ops_infer64_8.dll. Please make sure it is in your library path!

When i look at build/lib/torch/lib, i see that some cuda and cuddn dll's are missing. When I build it locally with cx_Freeze, I see that the required dll's are there. I think it's because it doesn't recognize the cuda installation correctly and therefore doesn't copy the dll's into the build

here is my action.yml:

 windows-backend-build:
    runs-on: windows-2019
    steps:
      - name: Install CUDA
        uses: Jimver/[email protected]
        id: cuda-toolkit
        with:
          cuda: "12.4.0"

      - name: Set environment variables
        run: |
          setx CUDA_PATH "%CUDA_PATH%"
          setx CUDA_PATH_V12_4 "%CUDA_PATH%"

        env:
          CUDA_PATH: ${{ env.CUDA_PATH }}

      - name: Display CUDA Version
        run: |
          echo "Installed cuda version is: ${{steps.cuda-toolkit.outputs.cuda}}"
          echo "Cuda install location: ${{steps.cuda-toolkit.outputs.CUDA_PATH}}"

      - name: Checkout repository
        uses: actions/checkout@v2
        with:
          submodules: true
          fetch-depth: 0

      - name: Set up Python
        uses: actions/setup-python@v2
        with:
          python-version: "3.11.4"

      - name: Install dependencies
        run: |
          cd backend
          python -m pip install --upgrade pip
          pip install -r requirements.txt
          pip install transformers==4.40.1
        shell: bash

      - name: Build application
        run: |
          cd backend
          python setup.py build
        shell: cmd

      - name: Archive build output
        uses: actions/upload-artifact@v2
        with:
          name: build-output-windows-backend
          path: backend/build/

these are some of my requirements of requirements.txt:

torch==2.2.1 --index-url https://download.pytorch.org/whl/cu121
torchvision --index-url https://download.pytorch.org/whl/cu121
torchaudio==2.2.1
faster-whisper==1.0.2

Add Cuda 12.5

With update of VS compiler in github actions it now requires 12.5 to compile. Would be great to get support for 12.5

I tried to fork and update, but my ts skills is not adequate :)

My attempt: xnorpx@05dd9b7

  C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.1\include\crt/host_config.h(153): fatal error C1189: #error:  -- unsupported Microsoft Visual Studio version! Only the versions between 2017 and 2022 (inclusive) are supported! The nvcc flag '-allow-unsupported-compiler' can be used to override this version check; however, using an unsupported host compiler may cause compilation failure or incorrect run time execution. Use at your own risk.

error installing cuda: "E: Unable to locate package cuda-11-4"

error installing cuda: "E: Unable to locate package cuda-11-4"

thanks for maintaining this repo!

My builds recently started failing with the E: Unable to locate package cuda-11-4 error. Wondering if you are seeing the same with v0.2.8?

Run Jimver/[email protected]
  
/usr/bin/lsb_release -sr
22.04
/usr/bin/wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_[1](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:1).0-1_all.deb -O cuda_keyring.deb
--2022-[12](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:12)-11 03:07:27--  https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.0-1_all.deb
Resolving developer.download.nvidia.com (developer.download.nvidia.com)... 152.195.19.142
Connecting to developer.download.nvidia.com (developer.download.nvidia.com)|152.195.19.142|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 4332 (4.2K) [application/x-deb]
Saving to: ‘cuda_keyring.deb’
     0K ....                                                  100%  267M=0s
2022-[12](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:13)-11 03:07:27 (267 MB/s) - ‘cuda_keyring.deb’ saved [4332/4332]
/usr/bin/sudo dpkg -i cuda_keyring.deb
Selecting previously unselected package cuda-keyring.
(Reading database ... 261272 files and directories currently installed.)
Preparing to unpack cuda_keyring.deb ...
Unpacking cuda-keyring (1.0-1) ...
Setting up cuda-keyring (1.0-1) ...
A deprecated public CUDA GPG key appear to be installed.
To remove the key, run this command:
sudo apt-key del 7fa2af80
/usr/bin/wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-ubuntu2204.pin
--2022-12-11 03:07:28--  https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-ubuntu2204.pin
Resolving developer.download.nvidia.com (developer.download.nvidia.com)... 152.195.19.142
Connecting to developer.download.nvidia.com (developer.download.nvidia.com)|152.195.19.142|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 190 [application/octet-stream]
Saving to: ‘cuda-ubuntu2204.pin’
     0K                                                       100%  363K=0.001s
2022-12-11 03:07:28 (363 KB/s) - ‘cuda-ubuntu2204.pin’ saved [190/190]
/usr/bin/sudo mv cuda-ubuntu2204.pin /etc/apt/preferences.d/cuda-repository-pin-600
/usr/bin/sudo add-apt-repository deb http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/ /
Get:1 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64  InRelease [1581 B]
Hit:2 http://azure.archive.ubuntu.com/ubuntu jammy InRelease
Get:3 http://azure.archive.ubuntu.com/ubuntu jammy-updates InRelease [114 kB]
Get:4 http://azure.archive.ubuntu.com/ubuntu jammy-backports InRelease [99.8 kB]
Get:5 http://azure.archive.ubuntu.com/ubuntu jammy-security InRelease [110 kB]
Get:6 https://packages.microsoft.com/ubuntu/22.04/prod jammy InRelease [10.5 kB]
Get:7 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64  Packages [192 kB]
Get:8 http://azure.archive.ubuntu.com/ubuntu jammy-updates/main amd64 Packages [758 kB]
Get:9 http://azure.archive.ubuntu.com/ubuntu jammy-updates/main Translation-en [169 kB]
Get:10 http://azure.archive.ubuntu.com/ubuntu jammy-updates/main amd64 c-n-f Metadata [11.4 kB]
Get:11 http://azure.archive.ubuntu.com/ubuntu jammy-updates/restricted amd64 Packages [488 kB]
Get:12 http://azure.archive.ubuntu.com/ubuntu jammy-updates/restricted Translation-en [74.9 kB]
Get:[13](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:14) http://azure.archive.ubuntu.com/ubuntu jammy-updates/universe amd64 Packages [761 kB]
Get:[14](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:15) http://azure.archive.ubuntu.com/ubuntu jammy-updates/universe Translation-en [128 kB]
Get:[15](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:16) http://azure.archive.ubuntu.com/ubuntu jammy-updates/universe amd64 c-n-f Metadata [14.0 kB]
Get:[16](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:17) http://azure.archive.ubuntu.com/ubuntu jammy-updates/multiverse amd64 Packages [7300 B]
Get:[17](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:18) http://azure.archive.ubuntu.com/ubuntu jammy-backports/main amd64 Packages [3324 B]
Get:[18](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:19) http://azure.archive.ubuntu.com/ubuntu jammy-backports/main Translation-en [1580 B]
Get:[19](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:20) http://azure.archive.ubuntu.com/ubuntu jammy-backports/main amd64 c-n-f Metadata [272 B]
Get:[20](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:21) http://azure.archive.ubuntu.com/ubuntu jammy-backports/universe amd64 Packages [6740 B]
Get:[21](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:22) http://azure.archive.ubuntu.com/ubuntu jammy-backports/universe amd64 c-n-f Metadata [348 B]
Get:[22](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:23) http://azure.archive.ubuntu.com/ubuntu jammy-security/main amd64 Packages [525 kB]
Get:[23](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:24) http://azure.archive.ubuntu.com/ubuntu jammy-security/main Translation-en [113 kB]
Get:[24](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:25) http://azure.archive.ubuntu.com/ubuntu jammy-security/main amd64 c-n-f Metadata [7352 B]
Get:25 http://azure.archive.ubuntu.com/ubuntu jammy-security/restricted amd64 Packages [452 kB]
Get:26 http://azure.archive.ubuntu.com/ubuntu jammy-security/restricted Translation-en [69.2 kB]
Get:27 http://azure.archive.ubuntu.com/ubuntu jammy-security/universe amd64 Packages [622 kB]
Get:[28](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:29) http://azure.archive.ubuntu.com/ubuntu jammy-security/universe Translation-en [82.7 kB]
Get:[29](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:30) http://azure.archive.ubuntu.com/ubuntu jammy-security/universe amd64 c-n-f Metadata [11.0 kB]
Get:[30](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:31) https://packages.microsoft.com/ubuntu/22.04/prod jammy/main arm64 Packages [16.7 kB]
Get:[31](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:32) https://packages.microsoft.com/ubuntu/22.04/prod jammy/main armhf Packages [10.6 kB]
Get:[32](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:33) https://packages.microsoft.com/ubuntu/22.04/prod jammy/main amd64 Packages [60.1 kB]
Hit:[33](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:34) https://ppa.launchpadcontent.net/ubuntu-toolchain-r/test/ubuntu jammy InRelease
Fetched 4922 kB in 1s (42[34](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:35) kB/s)
Reading package lists...
W: Target Packages (Packages) is configured multiple times in /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list:1 and /etc/apt/sources.list.d/cuda-ubuntu2204-x86_64.list:1
W: Target Translations (en) is configured multiple times in /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list:1 and /etc/apt/sources.list.d/cuda-ubuntu2204-x86_64.list:1
W: Target Packages (Packages) is configured multiple times in /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list:1 and /etc/apt/sources.list.d/cuda-ubuntu2204-x86_64.list:1
W: Target Translations (en) is configured multiple times in /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list:1 and /etc/apt/sources.list.d/cuda-ubuntu2204-x86_64.list:1
Repository: 'deb http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/ /'
Description:
Archive for codename: / components: 
More info: http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/
Adding repository.
Adding deb entry to /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list
Adding disabled deb-src entry to /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list
/usr/bin/sudo apt-get update
Hit:1 http://azure.archive.ubuntu.com/ubuntu jammy InRelease
Hit:2 http://azure.archive.ubuntu.com/ubuntu jammy-updates InRelease
Hit:3 http://azure.archive.ubuntu.com/ubuntu jammy-backports InRelease
Hit:4 http://azure.archive.ubuntu.com/ubuntu jammy-security InRelease
Hit:5 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64  InRelease
Hit:6 https://packages.microsoft.com/ubuntu/22.04/prod jammy InRelease
Hit:7 https://ppa.launchpadcontent.net/ubuntu-toolchain-r/test/ubuntu jammy InRelease
Reading package lists...
W: Target Packages (Packages) is configured multiple times in /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_[64](https://github.com/ellisbrown/private-repo/actions/runs/3667122264/jobs/6199374776#step:4:65)_-jammy.list:1 and /etc/apt/sources.list.d/cuda-ubuntu2204-x86_64.list:1
W: Target Translations (en) is configured multiple times in /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list:1 and /etc/apt/sources.list.d/cuda-ubuntu2204-x86_64.list:1
W: Target Packages (Packages) is configured multiple times in /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list:1 and /etc/apt/sources.list.d/cuda-ubuntu2204-x86_64.list:1
W: Target Translations (en) is configured multiple times in /etc/apt/sources.list.d/archive_uri-http_developer_download_nvidia_com_compute_cuda_repos_ubuntu2204_x86_64_-jammy.list:1 and /etc/apt/sources.list.d/cuda-ubuntu2204-x86_64.list:1
/usr/bin/sudo apt-get -y install cuda-11-4
Reading package lists...
Building dependency tree...
Reading state information...
E: Unable to locate package cuda-11-4
Error: Error: The process '/usr/bin/sudo' failed with exit code **[100]

Error: Got no files in tool cahce

I am following the example and using the following in my workflow on an ubuntu-20.04 runner...

- uses: Jimver/[email protected]
      id: cuda-toolkit
      with:
        cuda: '12.3.2'

Here is the output...

Run Jimver/[email protected]
  with:
    cuda: 1[2](https://github.com/robandpdx-org/habitat-lab/actions/runs/8238324873/job/22529020320#step:4:2).[3](https://github.com/robandpdx-org/habitat-lab/actions/runs/8238324873/job/22529020320#step:4:3).2
    sub-packages: []
    non-cuda-sub-packages: []
    method: local
    linux-local-args: ["--toolkit", "--samples"]
    use-github-cache: true
    use-local-cache: true
    log-file-suffix: log.txt
  env:
    CODECOV_TOKEN: xxxx91d3
    FPS_THRESHOLD: 900
  
/usr/bin/tar --posix -cf cache.tzst --exclude cache.tzst -P -C /home/runner/work/habitat-lab/habitat-lab --files-from manifest.txt --use-compress-program zstdmt
Cache Size: ~[4](https://github.com/robandpdx-org/habitat-lab/actions/runs/8238324873/job/22529020320#step:4:4)1[5](https://github.com/robandpdx-org/habitat-lab/actions/runs/8238324873/job/22529020320#step:4:5)7 MB (435857[6](https://github.com/robandpdx-org/habitat-lab/actions/runs/8238324873/job/22529020320#step:4:6)5[7](https://github.com/robandpdx-org/habitat-lab/actions/runs/8238324873/job/22529020320#step:4:7)1 B)
Cache saved successfully
Error: Error: Got no files in tool cahce

Skip '--samples' by default

I propose to skip --samples from linux-local-args by default.
I dare say that developers using your Action typically don't need them, and are only interested in the rest of the toolkit to test that cuda compilation works properly.
This simple change can save some time for each build.

CUDA v 12.3.X on windows

Hi, @Jimver

i'm trying to use CUDA version 12.3.1 in my github workflow, but installing cuda fails with error:
\cuda_installer-windows-10.0.20348-12.3.0\cuda_installer-windows-10.0.20348_12.3.0.exe' failed with exit code 2147944003

worker on windows 10, any ideas? v.12.2 installs fine.

I've tried network and local installers, both are not working.

regards

CUDA v 12.3

Hi,

I.m using cuda toolkit and recently I've updated cuda to version 12.3, but i have message that that version is not supported, is there any chance for support for that version?

regards

Error: File /var/log/cuda-installer.log does not exist when installer an older cuda

thanks for developing this action script!

I spent the past few hours trying to use it for one of my CUDA projects. I found two issues and would like to see if there is a workaround

1. I kept getting an error (Error: Error: File /var/log/cuda-installer.log does not exist) at the end of the cuda installation, see https://github.com/fangq/mcx/actions/runs/4401047764/jobs/7706904559. I even forked this repo and removed the artifact uploading code inside src/installer.ts, yet, it still kept giving me this error. I was trying to install cuda 9.2 for its smaller size. is there a workaround?

2. the default cuda 12.x is over 4GB in size. when I used the default setting, I got "No space left on device" error, see https://github.com/fangq/mcx/actions/runs/4400016129/jobs/7704933495; same for the the CI tests: https://github.com/fangq/cuda-toolkit/actions/runs/4401127506/jobs/7707048051. what is the most compact installation option? I only need nvcc and libcudart/libcudartstatic for my project.

Error: Unexpected HTTP response: 404

All builds started failing with above error a few hours ago. I verified that direct links appear to be still valid, so I'm not sure what caused it.
Here is one of the CI jobs that failed: https://github.com/subspace/subspace/runs/5759646238?check_suite_focus=true

Cannot skip samples installation

Hi,

using Ubuntu 16.04, one cannot skip the installation of the samples:

Run Jimver/[email protected]
  with:
    cuda: 10.2.89
    subPackages: []
    method: local
/usr/bin/sudo /opt/hostedtoolcache/cuda_installer-linux/10.2.89/x64/cuda_installer_10.2.89.run --silent --toolkit --samples
Total size of all the files uploaded is 157428 bytes

Notify CMake on Windows about CUDA

Hey,

Thanks for your awesome Github action! My question might be a bit out of scope of this repo, but I wonder whether you could still help me.
I am trying to build something on windows and ubuntu and need nvcc for that. The build process is like:
conda-build --> ptyhon setup.py --> CMake --> nvcc

Locally, the build runs fine, but I wonder how I can tell CMake the CUDACXX on a windows runner?
If you want to have a look, you can find the workflow below. I tried setting the CUDACXX via powershell but CMake does not find it.
Any ideas?

https://github.com/TuragaLab/SplinePSF/blob/9aa4e4569fce8550a7b22295191447f88ee592e9/.github/workflows/build_upload_test.yaml#L54-L80