| 11 Mar 2026 |
 Gaétan Lepage | (same reasonning for https://github.com/NixOS/nixpkgs/pull/498678#issuecomment-4035473707). | 23:39:46 |
| Gaétan Lepage | * (same reasonning for https://github.com/NixOS/nixpkgs/pull/498678). | 23:39:52 |
 connor (burnt/out) (UTC-8) | Sounds good! I’ll leave a comment on them | 23:51:04 |
| Gaétan Lepage | I'm testing the CUDA bump more thoroughly though.
~1.3k rebuilds left (out of 1.8k) | 23:53:19 |
| Gaétan Lepage | * connor (burnt/out) (UTC-8)
About https://github.com/NixOS/nixpkgs/pull/498681, I plan to build torch and vllm. If this works fine, I will merge it.
With all the CUDA PRs in the queue, I won't have the capacity to exhaustively test all of them.
No objection on your side? | 23:54:04 |
| 12 Mar 2026 |
|  Theuni changed their display name from Christian Theune to Theuni. | 07:18:55 |
 Bryan Honof | It looks like torch's supportedTorchCudaCapabilities was out-of-sync with upstream. https://github.com/NixOS/nixpkgs/pull/499216
How would I use nixpkgs-review to test these changes?
| 10:53:19 |
| Gaétan Lepage | Thanks for the PR!
Well, you don't want to rebuild all torch consumers for this. What you can do is the following:
nixpkgs-review --extra-nixpkgs-config "{ allowUnfree = true; cudaSupport = true; }" -p python3Packages.torch -p python3Packages.vllm -p python3Packages.torchvision
| 12:39:42 |
| Gaétan Lepage | I'll try to have a look at it before next week | 12:39:58 |
| Gaétan Lepage | connor (burnt/out) (UTC-8) actually, neither the current nor the new gpu-burn version work:
❮ ./result/bin/gpu_burn
Run length not specified in the command line. Using compare file: /nix/store/9c2avfi2bxc2aydfl2sdgkp8iamhj8as-gpu-burn-0-unstable-2024-04-09/share/compare.ptx
Burning for 10 seconds.
GPU 0: NVIDIA GeForce RTX 3060 (UUID: GPU-7d08a1e6-4634-499f-d58a-91bf77137f69)
Initialized device 0 with 11911 MB of memory (11788 MB available, using 10609 MB of it), using FLOATS
Results are 268435456 bytes each, thus performing 39 iterations
Couldn't init a GPU test: Error in load module (gpu_burn-drv.cpp:239): a PTX JIT compilation failed
0.0% proc'd: -1 (0 Gflop/s) errors: 0 (DIED!) temps: 36 C
(tested on 2 GPUs) | 19:56:18 |
| Gaétan Lepage | Nevermind, all good.
You need to carefully set cudaCapabilities for it to run fine on a given GPU: https://github.com/NixOS/nixpkgs/pull/499323#issuecomment-4049769046 | 20:25:50 |
 apyh | In reply to @glepage:matrix.org
Nevermind, all good.
You need to carefully set cudaCapabilities for it to run fine on a given GPU: https://github.com/NixOS/nixpkgs/pull/499323#issuecomment-4049769046 should it have an isBroken if cudaCapabilities has more than one item, then? | 20:42:25 |
| Gaétan Lepage | Not really. It selects the highest (techically, the last) capability from your config.cudaCapabilities.
So there's no fundamental reason why a list with additional, lower caps than the one of your GPU could not work. | 22:14:15 |
| 13 Mar 2026 |
| Bryan Honof | Thanks! That seems to have built succesfully. :) | 09:26:19 |
| 16 Mar 2026 |
|  Kevin Mittman (jetlagged/UTC-7) changed their display name from Kevin Mittman (jetlagged/UTC+8) to Kevin Mittman (jetlagged/UTC-7). | 00:57:26 |
| 17 Mar 2026 |
| connor (burnt/out) (UTC-8) | Does https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html#cuda-compiler (improved C++ standards conformance bullet) mean we could drop our patches to NVCC for 13.2? | 22:38:09 |
|  kirillrdy joined the room. | 22:48:24 |
| kirillrdy | connor (burnt/out) (UTC-8): git blame pkgs/development/cuda-modules/packages/tests/onnx-tensorrt/short.nix shows your name, is there onnx-tensorrt support in nixpkgs ? i think not | 23:03:22 |
| connor (burnt/out) (UTC-8) | No, I packaged a bunch of stuff in https://github.com/connorbaker/cuda-packages but haven't had the chance to upstream it because they have massive dependency chains and I also need to support every release since CUDA 11.4 🫠 | 23:47:30 |
| kirillrdy | ah awesome ! thanks for the link | 23:48:25 |
| kirillrdy | i guess i could have searched myself | 23:48:42 |
| 18 Mar 2026 |
| Kevin Mittman (jetlagged/UTC-7) | kirillrdy: do you happen to know how is it related to libnvonnxparser ? | 03:55:22 |
| connor (burnt/out) (UTC-8) | iirc it provides that lib when built; see https://github.com/onnx/onnx-tensorrt/blob/10.15-GA/NvOnnxParser.h. I think previously when I packaged it (with the intention of upstreaming it to Nixpkgs, never got around to it) I preferred to use libnvonnxparser from onnx-tensorrt instead of the one provided by the tensorrt binary archives | 06:00:28 |
| connor (burnt/out) (UTC-8) | ugh there's still a typo on the TensorRT page where ONNX is spelled "ONYX" https://developer.nvidia.com/tensorrt#:~:text=for%20production%20applications.-,ONYX,-%3A | 06:01:30 |
| connor (burnt/out) (UTC-8) | Gaétan Lepage: various patches for CUDA 13 stuff https://github.com/ConnorBaker/blender-temp/blob/7132a52cd929b764e83dc97fb0596b8caae58c95/overlay.nix | 06:43:24 |
| Gaétan Lepage | Thanks for sharing :) | 11:32:30 |
|  ereslibre removed their profile picture. | 15:52:02 |
| ereslibre | hello! was wondering, can we get github.com/NixOS/nixpkgs/pull/489218 merged? Thank you :) | 15:52:49 |
| Gaétan Lepage | connor (burnt/out) (UTC-8), can you stamp https://github.com/NixOS/nixpkgs/pull/501079 whenever you have a minute please? | 17:55:27 |
| ereslibre | thanks connor (burnt/out) (UTC-8) ! | 17:56:59 |
