CUDA in 22.05

[SomeoneSerge]

Issue description

This is a aggregate issue to track (the selected subset of) that state of CUDA in master before the 22-05 release. It's concern is the world evaluated with { config.cudaSupport = true; config.allowUnfree = true; }

• The release will see the following changes that affect how we package and use cuda:
  • [x] The new scope-based interface for consuming cuda-related packages, introduced together with the redistributable cuda packages in https://github.com/NixOS/nixpkgs/pull/167016/
  • [x] cudnn drops its propagatedBuildInputs which used to include cudatoolkit and cause a lot of obscure errors. Happens in: https://github.com/NixOS/nixpkgs/pull/168748

• We currently have the following failures (we might not fix them before the release, but back-port later)

  • [x] As of (at least) 2022-04-11 jax is failing on master, cf. jax: tests/test_sparse.py

  • [x] test_sparse.py has been fixed upstream, but blocked in nixpkgs because of: https://github.com/NixOS/nixpkgs/issues/169148

  • [ ] catboost

  • [x] As of (at least) 2022-04-11, tensorflow-datasets is failing, cf. https://github.com/NixOS/nixpkgs/issues/167445

  • [x] xgboost

  CMake Error at CMakeLists.txt:140 (message):
  `BUILD_WITH_CUDA_CUB` should be set to `ON` for CUDA >= 11.4

  ...and now it's https://github.com/NixOS/nixpkgs/issues/167931

  • [x] python39Packages.mxnet CMake configurePhase fails

  Likely will be fixed in: https://github.com/NixOS/nixpkgs/pull/167250

  • [ ] cholmod-extra

  Tests fail: too large residuals

  • [x] python3*Packages.chainer

  Complains about cupy version

  Marked as broken in: https://github.com/NixOS/nixpkgs/pull/167250

  • [x] python310Packages.caffe

  CMake configurePhase fails

  • [x] python39Packages.caffe

  Fixed by: https://github.com/NixOS/nixpkgs/pull/166784/

  • [x] blender: failing dependency (opensubdiv)

  • [x] opensubdiv: fails at cmake configurePhase

  Likely will be fixed in: https://github.com/NixOS/nixpkgs/pull/167250

  • [x] Pytorch

  Cmake error: https://github.com/NixOS/nixpkgs/issues/166403 Build error: https://github.com/NixOS/nixpkgs/issues/161843 Fixed by: https://github.com/NixOS/nixpkgs/pull/166784/

cc @NixOS/cuda-maintainers

[samuela]

I think we may just need to mark some of these as broken. For example, chainer is requesting a truly ancient version of cupy: cupy>=7.7.0,<8.0.0. Latest stable cupy is v10.3.0.

cc'ing some maintainers:

• blender: @cillianderoiste @veprbl
• opensubdiv: @edolstra
• cholmod-extra: @jluttine
• python3Packages.chainer: @hyphon81
• python310Packages.caffe: @jb55

[SomeoneSerge]

The state of things at the time of the release landing:

I believe it would be still justifiable to backport some incremental improvements, fixes, and extensions as we make them. In particular, in the release we still

• [ ] haven't finished migrating downstream expressions to the redistributable cudaPackages, even the fundamental packages like
  • [ ] pytorch,
  • [ ] jax,
  • [ ] or tensorflow;
• [ ] don't have a good control of the closure size, e.g. when building docker images with deep learning packages: https://github.com/NixOS/nixpkgs/issues/173462
• ..?