| 25 Jul 2025 |
 apyh | i have a 7800x3d and it still took 3.5 hours per torch build | 17:30:26 |
| 26 Jul 2025 |
 Tristan Ross | Is that a PR that my 128 cores could be useful with? | 00:34:02 |
| apyh | haha i mean, if you have the ram to match ;) | 01:07:29 |
| apyh | it builds fine on my end - just a verification from someone else would be nice :) | 01:07:40 |
|  oak 🏳️🌈♥️ changed their profile picture. | 08:29:04 |
 Gaétan Lepage | Any objection to merging the nccl bump?
https://github.com/NixOS/nixpkgs/pull/427804 | 09:26:39 |
| 3 Aug 2025 |
 connor (he/him) | As a heads up, I’ll be removing CUDA versions prior to 11.8 from Nixpkgs and moving them to https://github.com/nix-community/cuda-legacy so they’ll be available as an overlay using the cool new functionality I got merged over the last few months to make it easier to extend CUDA stuff out of tree | 19:12:49 |
| 4 Aug 2025 |
 aidalgol | I'm trying to write a derivation for https://github.com/alicevision/popsift/ and the configure phase fails because it cant find "cudadevrt".
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
CUDA_CUDADEVRT_LIBRARY
linked by target "popsift" in directory /build/source/src
linked by target "popsift-match" in directory /build/source/src/application
linked by target "popsift-demo" in directory /build/source/src/application
CUDA_cudadevrt_LIBRARY (ADVANCED)
linked by target "popsift" in directory /build/source/src
I haven't been able to find much about this beyond that it is the CUDA Device Runtime library. A grep for "devrt" on nixpkgs turns up nothing. Anyone here know about this CUDA library?
| 11:00:58 |
 Robbie Buxton | It might be the dev output of cuda_cudart, I.e cuda_cudart.dev | 11:02:21 |
| Robbie Buxton | I remember running into something similar in the past and I think I fixed it that way but might be misremembering | 11:03:16 |
| aidalgol | Yay, now I'm on to compilation errors! | 11:05:39 |
| aidalgol | Looks like this will only build on ooold CUDA versions. | 11:06:05 |
| aidalgol | nvcc fatal : Unsupported gpu architecture 'compute_35'
| 11:06:18 |
| aidalgol | I want to get meshroom in nixpkgs, and it appears that all the CUDA parts are still targeting CUDA pre-11. | 11:07:08 |
| connor (he/him) | Are they manually specifying compute capabilities in the source build system? If so, generally those can be patched out. | 14:41:22 |
| connor (he/him) | Try using the latest commit and make sure to set https://github.com/alicevision/popsift/blob/8623b697ac632f40d74f8773d646d6a804e89893/CMakeLists.txt#L14 | 14:46:54 |
| connor (he/him) | Inside the flags attribute of the CUDA package set (https://github.com/NixOS/nixpkgs/blob/7306c79fa2fd9483e7abd32e8637fc7f20344cbd/pkgs/top-level/cuda-packages.nix#L104) we have a helper attribute (https://github.com/NixOS/nixpkgs/blob/7306c79fa2fd9483e7abd32e8637fc7f20344cbd/pkgs/development/cuda-modules/_cuda/lib/strings.nix#L196) which you can use to set that CMake variable | 14:49:07 |
| connor (he/him) | Example usage: https://github.com/NixOS/nixpkgs/blob/36590ed03f5cd09054978e927af3eb41e89fe838/pkgs/by-name/ll/llama-cpp/package.nix#L140 | 14:49:46 |
| connor (he/him) | It’s not documented because there’s basically never time for anything lmao 😭 | 14:50:15 |
| connor (he/him) | At the end of August when I’m at Nix Camp / Nix Con I hope I get the chance to write a bunch more documentation, including references for different build systems, and maybe even a tutorial for packing CUDA applications (although they’re all so frustratingly, wildly different that there’s barely a generic way to approach packaging them, and can involve bash scripts templating Python scripts which then run and configure CMake lmao) | 14:52:27 |
| aidalgol | Progress!! | 21:06:33 |
| aidalgol | https://bpa.st/MDQA | 21:06:35 |
| 5 Aug 2025 |
| connor (he/him) | CUDA 13 release notes are up, some interesting stuff in there, will hopefully have time this month to take a longer look
Still don’t remember whether our setup works with device LTO or separate compilation, not sure how big an issue that’ll be | 05:16:36 |
|  le-chat joined the room. | 11:09:02 |
| 6 Aug 2025 |
| le-chat | Hi. Does somebody know how to make a meson project to see a PyTorch dependency? I'm trying to compile NNStreamer (these are GStreamer plugins for different NN frameworks) with PyTorch backend, but meson reports it has not found pytorch.
An example flake is https://gist.github.com/le-chat/e5c7f583a4fa51742dcbdf5821350679
I've tried extending CMAKE_PREFIX_PATH but without success. | 13:04:23 |
| connor (he/him) | IIRC PyTorch (like OpenCV) has a cxxdev output you need to include if you want to use it in a C++ project; give that a try | 15:13:52 |
| connor (he/him) | Oh you already tried that, never mind
Have you tried with just the cxxdev output? And torch should only go in buildInputs, not nativeBuildInputs | 15:15:11 |
| connor (he/him) | But generally avoid the dev output and make sure to use the cxxdev output for C++ projects | 15:16:18 |
| 7 Aug 2025 |
 grw00 | hey all- is there a similar channel/effort for rocm in nixos? | 09:57:51 |
| le-chat | Now magma fails to build, I did flake update. Switched to nixos-25.05, reran with torch.cxxdev only in buildInputs. Will see. | 10:09:53 |
