| 23 Apr 2026 |
 Gaétan Lepage | * connor (burnt/out) (UTC-8), do you know if libnvvm is a CUDA>=13.0 thing? Or should it also be packaged for 12.9? | 16:18:49 |
| Gaétan Lepage | * connor (burnt/out) (UTC-8), do you know if libnvvm is a CUDA>=13.0 thing? Or should it also be packaged for 12.9?
It should be... But only the 13.x redist manifests mention it... | 16:21:09 |
| Gaétan Lepage | * connor (burnt/out) (UTC-8), do you know if libnvvm is a CUDA>=13.0 thing? Or should it also be packaged for 12.9?
It should be... But only the 13.x redist manifests mention it...
EDIT: It's under cudaPackages.cuda_nvcc/nvvm for 12.x 🫠 | 16:27:19 |
 connor (burnt/out) (UTC-8) | Oh my god | 16:28:44 |
| connor (burnt/out) (UTC-8) | Oh wait yeah it was bundled with NVCC then was split out into its own library phew | 16:29:05 |
| connor (burnt/out) (UTC-8) | IIRC I symlink it into NVCC to avoid breaking consumers? | 16:29:22 |
| Gaétan Lepage | As you say. phew. | 16:31:54 |
| Gaétan Lepage | Raw fun: https://github.com/NixOS/nixpkgs/pull/512828 | 16:32:17 |
 SomeoneSerge (matrix works sometimes) | I'd rather change the definition of libnvvm to extract itself from nvcc in cuda<13 and declare that the current interface. | 16:55:27 |
| connor (burnt/out) (UTC-8) | The latent concern I have is around packages assuming relative path access to work :( | 18:04:25 |
| SomeoneSerge (matrix works sometimes) | Depending on fu... motivation, this is not necessarily a lost battle;) | 18:44:14 |
| Gaétan Lepage | connor (burnt/out) (UTC-8), I'd like to get this one in if you have a minute: https://github.com/NixOS/nixpkgs/pull/512796 | 20:49:46 |
| 24 Apr 2026 |
|  amarshall joined the room. | 05:39:52 |
 mdietrich | Hi all, this might be a stupid question but why do I have to build some packages still locally when on nixos-unstable with nixpkgs.config.cudaSupport = true; even though I've configured the binary cache in my substituters? Is this just because cuda packages have complicated versions and change very often that not all packages are built and cached on the cachix or am I doing something wrong?
This actually hinders me from updating right now because python3.13-torch-2.11.0 for python3.13-whisper-20250625 and python3.12-torch-2.11.0 for aider-chat-playwright-browser-help-bedrock-0.86.1 are eating my RAM and swap alive.
| 10:25:30 |
| mdietrich | Oh, and I'm sorry if this is the wrong place, I just remembered that there is a forum, duh. | 10:26:21 |
 Robbie Buxton | Just to sanity check this is when you are using the cache.nixos-cuda.org? | 10:28:10 |
| mdietrich | Yeah, this is the full relevant code:
nix = {
registry.nixpkgs.flake = inputs.nixpkgs;
channel.enable = false; # Get rid of both unneeded state files and the nasty "Nix search path entry ... does not exist"
settings = rec {
use-xdg-base-directories = true;
experimental-features = [
"nix-command"
"flakes"
];
substituters = [
"https://nix-community.cachix.org"
"https://ai.cachix.org"
"https://numtide.cachix.org"
"https://cache.garnix.io"
"https://cache.nixos-cuda.org"
];
trusted-substituters = substituters;
trusted-public-keys = [
"nix-community.cachix.org-1:mB9FSh9qf2dCimDSUo8Zy7bkq5CX+/rkCWyvRCYg3Fs="
"ai.cachix.org-1:N9dzRK+alWwoKXQlnn0H6aUx0lU/mspIoz8hMvGvbbc="
"numtide.cachix.org-1:2ps1kLBUWjxIneOy1Ik6cQjb41X0iXVXeHigGmycPPE="
"cache.garnix.io:CTFPyKSLcx5RMJKfLo5EEPUObbA78b0YQ2DTCJXqr9g="
"cache.nixos-cuda.org:74DUi4Ye579gUqzH4ziL9IyiJBlDpMRn9MBN8oNan9M="
];
};
optimise = {
automatic = true;
dates = [ "weekly" ];
};
};
nixpkgs.config.allowUnfreePredicate = (
pkg:
builtins.elem (lib.getName pkg) [
"nvidia-x11"
"cuda_cccl"
"libcublas"
"libcurand"
"libcusparse"
"libnvjitlink"
"libcufft"
"cudnn"
"cuda_nvrtc"
"cuda_cudart"
"cuda_nvcc"
"libnpp"
"cuda_cupti"
"cuda_nvml_dev"
"cuda_nvtx"
"libcufile"
"libcusolver"
"libcusparse_lt"
"cuda_profiler_api"
"cuda_cuobjdump"
"cuda_nvdisasm"
"cuda-merged"
"cuda_gdb"
"cuda_nvprune"
"cuda_cuxxfilt"
"cuda_sanitizer_api"
"nvidia-settings"
"obsidian"
"mathematica"
"spotify"
"discord"
"zoom"
"steamcmd"
"steam"
"steam-unwrapped"
]
);
| 10:33:53 |
| Gaétan Lepage | Which revision of nixpkgs are you tracking?
Torch should be cached: https://hydra.nixos-cuda.org/job/nixos-cuda/cuda-packages-unstable/python313Packages.torch.x86_64-linux | 11:48:59 |
| mdietrich | nixpkgs unstable with ref github:NixOS/nixpkgs/b12141ef619e0a9c1c84dc8c684040326f27cdcc?narHash=sha256-ZSK0NL4a1BwVbbTBoSnWgbJy9HeZFXLYQizjb2DPF24%3D according to nix flake info | 14:31:09 |
|  ereslibre set a profile picture. | 19:59:01 |
 ccicnce113424 | Is there any remaining blocker?
https://github.com/NixOS/nixpkgs/pull/498612 | 20:35:23 |
| 25 Apr 2026 |
| Gaétan Lepage | 🗞️Jax update: 0.10.0
Changelog: https://docs.jax.dev/en/latest/changelog.html
PR: https://github.com/NixOS/nixpkgs/pull/511567
Tracker: https://nixpkgs-tracker.ocfox.me/?pr=511567 | 08:22:30 |
| 26 Apr 2026 |
 hexa (UTC+1) | https://github.com/NixOS/nixpkgs/issues/310772#event-24879453622 | 12:46:21 |
| hexa (UTC+1) | is this nixos-cuda spam? | 12:46:24 |
| hexa (UTC+1) | https://github.com/nixos-cuda/infra/blob/master/hosts/hydra/hydra-github-app.nix | 12:46:47 |
| SomeoneSerge (matrix works sometimes) | Experimental/not-yet-official work, I think the backlink is more of an unintended github artifact...
Cc @yorik.sar:matrix.org
| 13:10:01 |
 yorik.sar | hexa (UTC+1) we’re keeping master branch in our fork up to date with the upstream branch. For each commit that references an issue GitHub creates an event in that issue. I wouldn’t call this spam, it’s just how GitHub works. | 14:50:26 |
| ccicnce113424 | Seeking a merge for #498612. It’s a 500+ line improvement with 6 approvals. I've kept it rebased, and since the breaking-change window is closing in 15h, I'd hate to see it stall. Thanks!
https://github.com/NixOS/nixpkgs/pull/498612 | 15:59:50 |
| SomeoneSerge (matrix works sometimes) | Got hands back on my laptop last night, hoping to have one last look-over on it today | 16:32:21 |
|  Arne joined the room. | 22:49:09 |
