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.

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

CUDA-related updates and cleanups #30433

Merged
merged 13 commits into from Oct 19, 2017
Merged

CUDA-related updates and cleanups #30433

merged 13 commits into from Oct 19, 2017

Conversation

abbradar
Copy link
Member

Motivation for this change
  1. Update CUDA toolkit to the latest version, make it default;
  2. Make a unified config flag cudaSupport which controls whether CUDA support is desired in applications;
  3. Update and cleanup various machine learning libraries which make use of CUDA.
Things done
  • Tested using sandboxing (nix.useSandbox on NixOS, or option build-use-sandbox in nix.conf on non-NixOS)
  • Built on platform(s)
    • NixOS
    • macOS
    • other Linux distributions
  • Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
  • Tested compilation of all pkgs that depend on this change using nix-shell -p nox --run "nox-review wip"
  • Tested execution of all binary files (usually in ./result/bin/)
  • Fits CONTRIBUTING.md.

@abbradar abbradar requested a review from FRidh as a code owner October 15, 2017 12:48
@abbradar
Copy link
Member Author

Notice that I don't have access to NVIDIA GPU now (My Google Cloud free credit ran empty :D) so it'd be nice to have someone test those changes (I used all of them but some time ago and rebase may have broken things).

@abbradar abbradar mentioned this pull request Oct 15, 2017
Merged
8 tasks
@abbradar abbradar force-pushed the cudatoolkit branch 4 times, most recently from 7a83f07 to 9120f78 Compare October 16, 2017 09:36
@abbradar
Copy link
Member Author

I've managed to make my laptop's hybrid Nvidia to work with CUDA so these changes are mostly tested now. Let's merge this after #29814 lands so that our default Boost plays nicely with new toolkit.

This was referenced Oct 17, 2017
Merged
Merged
@abbradar
cudatoolkit9: init at 9.0.176
fe7c7a8 
Make CUDA Toolkit 9 the default choice.
@abbradar
cudatoolkit: use recommended gcc by default
7cf6010 
This way one may build packages requiring old CUDA Toolkit with usual stdenv;
only bits which are compiled with NVCC would use older GCC.
@abbradar
cudnn_cudatoolkit9: init at 7.0
ce9059a 
Make it the default.
@abbradar
cudnn: refactor to use shared derivation
6309e5f 
Also return 8.0-6.0 version for TensorFlow.
@abbradar
treewise: respect config.cudaSupport in packages
85d3c00
@abbradar
treewide: use default compiler for cuda
378bfba
@abbradar
caffe: 1.0-rc5 -> 1.0
57f82de 
Fix build with cudatoolkit9.
@abbradar
clblas: 20160505 -> 2.12
50a223b
@abbradar
python.pkgs.libgpuarray: -9998.0 -> 0.6.9
efd2444
@abbradar
python.pkgs.Theano: merge CUDA and non-CUDA versions
771e403
@abbradar
opensubdiv: 3.2.0 -> 3.3.0
b317d7c 
Fix build with new CUDA toolkit.
@abbradar
blender: fix build with new cudatoolkit
22b5828
@abbradar
opencv3: use older cudatoolkit
59f51eb
@abbradar abbradar merged commit 15bfc8c into NixOS:master Oct 19, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@FRidh FRidh Awaiting requested review from FRidh

Assignees
No one assigned
Labels
8.has: package (update)
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
1 participant
@abbradar