Commit Graph

434 Commits

Author SHA1 Message Date
Elliot Berman
236476c9d8 cudaPackages.tensorrt-samples: Add support for 10.16.1
(cherry picked from commit 5bb67e892a)
2026-06-24 22:44:05 +00:00
Elliot Berman
5d1bf94b76 cudaPackages.tensorrt: Add TensorRT 10.16.1 manifest
(cherry picked from commit 9d3d755e8e)
2026-06-24 22:44:05 +00:00
Elliot Berman
af1765be25 cuda-modules: Allow SBSA Orin to use CUDA 13+
NVIDIA announced an SBSA software stack for the Jetson Orin (capability
8.7) using CUDA 13.2 [1]. Update the redist system selection logic and
assertions to support Orin SBSA.

[1]: https://forums.developer.nvidia.com/t/jetpack-7-2-jetson-linux-r39-2-on-jetson-agx-orin-developer-kit-getting-started-and-feedback-thread/372156

(cherry picked from commit e4d4b7fa70)
2026-06-24 03:04:31 +00:00
Gaetan Lepage
b48bc86453 cudaPackages_13_3: init at 13.3.0
Changelog:
https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html
(cherry picked from commit fe55e1a378)
2026-06-02 08:48:49 +00:00
Else, Someone
b0015daf6f cudaPackages.tensorrt: mark insecure 2026-05-22 23:44:39 +03:00
Gaetan Lepage
135ac68cdc cudaPackages.cudnn: 9.13.0 -> 9.22.0
u
2026-05-17 20:47:35 +00:00
Gaetan Lepage
a2baf7bb69 cudaPackages.libcusolvermp: fix build 2026-05-13 14:30:32 +00:00
Gaetan Lepage
4009183fe6 cudaPackages.libcusolvermp: 0.7.0 -> 0.8.0
Changelog:
https://docs.nvidia.com/cuda/cusolvermp/release_notes/index.html#cusolvermp-v0-8-0
2026-05-13 08:20:44 +00:00
Gaetan Lepage
daa7b35604 cudaPackages.libcublasmp: 0.6.0 -> 0.8.1
Changelog:
https://docs.nvidia.com/cuda/cublasmp/release_notes/index.html
2026-05-05 08:25:00 +00:00
Gaetan Lepage
ad39f6cb9f cudaPackages.nvbandwidth: init at 0.9 2026-04-30 15:58:13 +00:00
Gaetan Lepage
4ed0f75b07 cudaPackages_13_0: 13.0.2 -> 13.0.3
Changelog:
https://docs.nvidia.com/cuda/archive/13.0.3/cuda-toolkit-release-notes/index.html
2026-04-27 18:53:11 +00:00
Gaetan Lepage
01d468622e cudaPackages.cuda_crt: do not patch math headers on >=13.2 2026-04-27 08:57:36 +00:00
Gaetan Lepage
0db3511132 cudaPackages.cuda_crt: patch math_functions.h signatures 2026-04-23 23:05:33 +00:00
Gaetan Lepage
4bab2aa6af cudaPackages.libcuobjclient: init at 1.0.0.26 2026-04-21 15:25:07 +00:00
R. Ryantm
a9128c6c5b cudaPackages.nccl-tests: 2.18.2 -> 2.18.3 2026-04-17 10:37:57 +00:00
Samuel Ainsworth
8e55aea6ee cuda_nvcc: fix host_defines.h __noinline__ conflict under clang
host_defines.h redefines __noinline__ as __attribute__((noinline)).
When clang compiles CUDA code against libstdc++ >=12, which uses
__attribute__((__noinline__)), the macro causes the invalid expansion
__attribute__((__attribute__((noinline)))). Clang natively understands
__noinline__ as an attribute, so the macro is unnecessary — skip it
when clang is the compiler.
2026-04-12 22:06:57 +00:00
Gaetan Lepage
be983daa60 cudaPackages.libnvshmem: make the build customizable 2026-04-01 07:37:35 +00:00
Gaétan Lepage
0ccff482d4 cudaPackages.nccl-tests: 2.17.10 -> 2.18.2 (#499938) 2026-03-29 16:47:13 +00:00
Gaétan Lepage
238897676d cudaPackages.libnvshmem: 3.4.5-0 -> 3.6.5-0 (#503855) 2026-03-28 21:10:36 +00:00
Tim Dortmann
75b7fb4416 cudaPackages.nsight_compute: include necessary files 2026-03-27 14:20:34 +01:00
Tim Dortmann
6a78d900f4 cudaPackages.nsight_systems: support new directory structure 2026-03-26 22:52:22 +01:00
Gaetan Lepage
02e32e37db cudaPackages.libnvshmem: add GaetanLepage to maintainers 2026-03-26 21:21:44 +00:00
Gaetan Lepage
8639b5c1d0 cudaPackages.libnvshmem: 3.4.5-0 -> 3.6.5-0
Diff: https://github.com/NVIDIA/nvshmem/compare/v3.4.5-0...v3.6.5-0

Changelog: https://github.com/NVIDIA/nvshmem/releases/tag/v3.6.5-0
2026-03-26 21:21:44 +00:00
Gaetan Lepage
49a82e7410 cudaPackages_12_9.cuda_cccl: patch to fix missing _CCCL_PP_SPLICE_WITH_IMPL20 in preprocessor.h
PR: https://github.com/NVIDIA/cccl/pull/4972

Issue: https://github.com/NVIDIA/cccl/issues/4967
2026-03-23 17:57:32 +00:00
Gaétan Lepage
35e9cb5e37 cudaPackages.gdrcopy: 2.5.1 -> 2.5.2 (#500610) 2026-03-23 14:23:56 +00:00
Gaétan Lepage
b6a834e180 cuda_nvcc: also patch math_functions.hpp for glibc 2.42 (#498678) 2026-03-22 09:08:27 +00:00
Gaetan Lepage
9084f40a63 cudaPackages.cuda_cccl: patch to fix torch build on cuda>=13.0
Co-authored-by: Connor Baker <ConnorBaker01@gmail.com>
2026-03-18 17:14:10 +00:00
R. Ryantm
5bac571559 cudaPackages.gdrcopy: 2.5.1 -> 2.5.2 2026-03-17 02:32:25 +00:00
R. Ryantm
f5b25da559 cudaPackages.nccl-tests: 2.17.10 -> 2.18.2 2026-03-14 22:45:03 +00:00
Samuel Ainsworth
62cf59f29d cuda_nvcc: also patch math_functions.hpp for glibc 2.42
The existing glibc 2.42 compatibility patch fixes function signatures
in math_functions.h but misses math_functions.hpp, which has the same
rsqrt/sinpi/cospi functions wrapped in __func__() macros. Without
throw() annotations on these, C++ compilation fails when glibc 2.42
declares the same functions with throw().
2026-03-10 20:24:50 -04:00
Gaetan Lepage
cfa1f3da48 cudaPackages_13_2: init at 13.2.0 2026-03-10 11:51:03 +00:00
R. Ryantm
1ae2bed8e6 cudaPackages.nccl-tests: 2.17.9 -> 2.17.10 2026-03-06 22:55:32 +00:00
Connor Baker
d6997f7145 cudaPackages_13_1: init at 13.1.1 (#486717) 2026-02-21 16:56:26 +00:00
Gaetan Lepage
71c7f2d505 cudaPackages.cuda_nvcc: do not patch crt/math_functions.h on cuda 13 2026-02-09 17:19:33 +00:00
Gaetan Lepage
dedd0d1c8f cudaPackages.cuda_nvcc: optionals -> optionalString 2026-02-08 22:55:04 +00:00
Gaetan Lepage
eb6a294bd8 cudaPackages.cuda_nvcc: patch math_functions.h signatures
Fix compatibility with glibc 2.42
2026-02-07 22:57:23 +00:00
Gaetan Lepage
2017d6d515 cudaPackages_13_1: init at 13.1.1 2026-02-06 12:55:06 +00:00
R. Ryantm
f607495e0e cudaPackages.nccl-tests: 2.17.8 -> 2.17.9 2026-02-04 16:31:31 +00:00
jopejoe1
f21b68b3cb cudatoolkit: set pname instead of name 2026-02-03 22:19:33 +01:00
K900
520e2f27a5 treewide: remove unnecessary dependencies on mesa
None of those should be real. Things should not depend on mesa
directly for any reason ever, except llvmpipeHook.
2026-01-31 18:42:58 +03:00
Gaétan Lepage
e64e0da362 cudaPackages: fix redistName for NVPL packages (#485211) 2026-01-30 13:54:49 +00:00
Connor Baker
b04ad574ca cudaPackages.nvpl_sparse: remove non-existent static output
Signed-off-by: Connor Baker <ConnorBaker01@gmail.com>
2026-01-29 18:10:24 -08:00
Connor Baker
64e602559a cudaPackages: fix redistName for NVPL packages
Signed-off-by: Connor Baker <ConnorBaker01@gmail.com>
2026-01-29 17:50:25 -08:00
nixpkgs-ci[bot]
24bf3a7395 Merge master into staging-nixos 2026-01-27 00:21:30 +00:00
Connor Baker
3922461091 cudaPackages.cudnn: add missing nvrtc runtime dependency (#471394) 2026-01-26 23:25:40 +00:00
nixpkgs-ci[bot]
4f189086d0 Merge master into staging-nixos 2026-01-26 18:08:53 +00:00
quantenzitrone
7d8132a92c treewide: remove references to the xorg namespace in pkgs (automated)
this creates some eval errors that will be fixed in the next commit

done with the following script:

```fish
\#!/usr/bin/env fish

set packagesjson (nix eval --impure --json --expr '
let
  lib = import ./lib;
in
import pkgs/servers/x11/xorg/default.nix (lib.mapAttrs (
  name: _:
  if name == "lib" then
    lib
  else if name == "config" then
    { allowAliases = false; }
  else
    name
) (__functionArgs (import pkgs/servers/x11/xorg/default.nix))) { }
' | jq)

set one (grep '^    [A-Za-z0-9_-]*$' pkgs/servers/x11/xorg/default.nix | string trim | string replace -r '$' Z | sort | string sub -e -1)
set two (grep '^  [A-Za-z0-9_-]* = [A-Za-z0-9_-]*;$' pkgs/servers/x11/xorg/default.nix | cut -d= -f1 | string trim | string replace -r '$' Z | sort | string sub -e -1)

for arg in $one $two
    set oname $arg
    set nname (echo $packagesjson | jq -r .$oname)

    if test $nname = null
        echo (set_color red)warn:(set_color normal) unknown package xorg.$oname >&2
        continue
    end

    echo $oname "->" $nname

    # replace basic xorg.$name references
    for file in (rg -F "xorg.$oname" --files-with-matches pkgs)
        # special cases
        sd -F "$oname = xorg.$oname;" "$nname = $nname;" $file

        # replace
        sd -F "xorg.$oname" "$nname" $file

        # fixup function arguments

        # prevent duplicate function args
        if grep -E " ($oname|$nname),\$" $file >/dev/null
            continue
        end

        if grep 'xorg\..' $file >/dev/null # case1: there is more so we can't just remove the function arg

            if grep ' xorg,$' $file >/dev/null
                sd ' xorg,$' " xorg,
                $nname," $file

            else if grep ' xorg ? .*,$' $file >/dev/null
                sd 'xorg( ? .*),$' "xorg\$1,
                $nname," $file

            else
                sd -F 'xorg,' "$nname,
                xorg," $file
            end

        else # case there is no more xorg..* so we can just replace the function arg
            sd 'xorg(| ? .*),.*$' "$nname," $file
        end
    end
end

nix fmt
```
2026-01-25 22:28:09 +01:00
Ari Lotter
d893175673 cudaPackages.cudnn: add nvrtc to runpath
cudnn uses nvrtc at runtime on some hardware (e.g. H100). without this
runpath, we fail when triggering cudnn runtime compilation from pytorch
(see issue #461334)

closes #461334
2026-01-24 20:51:37 -05:00
Michael Daniels
006ecdbdeb treewide: fix typos (#479869) 2026-01-24 21:36:44 +00:00
Dyego Aurélio
28096cc5e3 treewide: apply nixfmt 1.2.0 2026-01-22 18:37:56 -03:00