Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions doc/languages-frameworks/cuda.section.md
Original file line number Diff line number Diff line change
Expand Up @@ -12,11 +12,11 @@ Nixpkgs provides a number of CUDA package sets, each based on a different CUDA r
- `cudaPackages_x`: A major-versioned alias to the major-minor-versioned CUDA package set with the latest widely supported major CUDA release.
- `cudaPackages`: An unversioned alias to the major-versioned alias for the latest widely supported CUDA release. The package set referenced by this alias is also referred to as the "default" CUDA package set.

It is recommended to use the unversioned `cudaPackages` attribute. While versioned package sets are available (e.g., `cudaPackages_12_2`), they are periodically removed.
It is recommended to use the unversioned `cudaPackages` attribute. While versioned package sets are available (e.g., `cudaPackages_12_8`), they are periodically removed.

Here are two examples to illustrate the naming conventions:

- If `cudaPackages_12_8` is the latest release in the 12.x series, but core libraries like OpenCV or ONNX Runtime fail to build with it, `cudaPackages_12` may alias `cudaPackages_12_6` instead of `cudaPackages_12_8`.
- If `cudaPackages_12_9` is the latest release in the 12.x series, but core libraries like OpenCV or ONNX Runtime fail to build with it, `cudaPackages_12` may alias `cudaPackages_12_8` instead of `cudaPackages_12_9`.
- If `cudaPackages_13_1` is the latest release, but core libraries like PyTorch or Torch Vision fail to build with it, `cudaPackages` may alias `cudaPackages_12` instead of `cudaPackages_13`.

All CUDA package sets include common CUDA packages like `libcublas`, `cudnn`, `tensorrt`, and `nccl`.
Expand Down Expand Up @@ -146,7 +146,7 @@ These settings ensure that the CUDA setup hooks function as intended.
When using `callPackage`, you can choose to pass in a different variant, e.g. when a package requires a specific version of CUDA:

```nix
{ mypkg = callPackage { cudaPackages = cudaPackages_12_2; }; }
{ mypkg = callPackage { cudaPackages = cudaPackages_12_6; }; }
```

::: {.caution}
Expand Down
2 changes: 2 additions & 0 deletions doc/release-notes/rl-2511.section.md
Original file line number Diff line number Diff line change
Expand Up @@ -69,6 +69,8 @@

- `mongodb-6_0` was removed as it is end of life as of 2025-07-31.

- CUDA versions below 12.6 have been removed, as they are unmaintained upstream and depend on end‐of‐life compilers.

- `vmware-horizon-client` was renamed to `omnissa-horizon-client`, following [VMware's sale of their end-user business to Omnissa](https://www.omnissa.com/insights/introducing-omnissa-the-former-vmware-end-user-computing-business/). The binary has been renamed from `vmware-view` to `horizon-client`.

- `neovimUtils.makeNeovimConfig` now uses `customLuaRC` parameter instead of accepting `luaRcContent`. The old usage is deprecated but still works with a warning.
Expand Down
14 changes: 6 additions & 8 deletions pkgs/by-name/ca/catboost/package.nix
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,6 @@
gitUpdater,
cudaSupport ? config.cudaSupport,
cudaPackages ? { },
llvmPackagesCuda ? llvmPackages,
pythonSupport ? false,
}:
let
Expand Down Expand Up @@ -45,10 +44,6 @@ stdenv.mkDerivation (finalAttrs: {
shopt -s globstar
for cmakelists in **/CMakeLists.*; do
sed -i "s/OpenSSL::OpenSSL/OpenSSL::SSL/g" $cmakelists
${lib.optionalString (cudaPackages.cudaOlder "11.8") ''
sed -i 's/-gencode=arch=compute_89,code=sm_89//g' $cmakelists
sed -i 's/-gencode=arch=compute_90,code=sm_90//g' $cmakelists
''}
done
'';

Expand Down Expand Up @@ -91,7 +86,7 @@ stdenv.mkDerivation (finalAttrs: {
# catboost requires clang 14+ for build, but does clang 12 for cuda build.
# after bumping the default version of llvm, check for compatibility with the cuda backend and pin it.
# see https://catboost.ai/en/docs/installation/build-environment-setup-for-cmake#compilers,-linkers-and-related-tools
CUDAHOSTCXX = lib.optionalString cudaSupport "${llvmPackagesCuda.stdenv.cc}/bin/cc";
CUDAHOSTCXX = lib.optionalString cudaSupport "${stdenv.cc}/bin/cc";
NIX_CFLAGS_LINK = lib.optionalString stdenv.hostPlatform.isLinux "-fuse-ld=lld";
NIX_LDFLAGS = "-lc -lm";
NIX_CFLAGS_COMPILE = toString (
Expand Down Expand Up @@ -139,7 +134,10 @@ stdenv.mkDerivation (finalAttrs: {
natsukium
];
mainProgram = "catboost";
# /nix/store/hzxiynjmmj35fpy3jla7vcqwmzj9i449-Libsystem-1238.60.2/include/sys/_types/_mbstate_t.h:31:9: error: unknown type name '__darwin_mbstate_t'
broken = stdenv.hostPlatform.isDarwin && stdenv.hostPlatform.isx86_64;
broken =
# See: <https://github.com/catboost/catboost/issues/2755>
cudaSupport
# /nix/store/hzxiynjmmj35fpy3jla7vcqwmzj9i449-Libsystem-1238.60.2/include/sys/_types/_mbstate_t.h:31:9: error: unknown type name '__darwin_mbstate_t'
|| (stdenv.hostPlatform.isDarwin && stdenv.hostPlatform.isx86_64);
};
})
110 changes: 0 additions & 110 deletions pkgs/by-name/dc/dcgm/fix-includes.patch

This file was deleted.

Loading
Loading