| 8 Dec 2024 |
 kaya π€ | * Pretty sure the biggest difference between the flake and my upstreaming attempt is that in the flake i have allowUnfree and cudaSupport as true, but those options should carry over (i assume) as i also have those enabled on my nixos config and im building with --impure | 16:50:40 |
| 9 Dec 2024 |
 SomeoneSerge (back on matrix) |
For some reason its complaining about CUDA_HOME being missing even though im specifying it which im kind of confused
It might just want some more components than just nvcc and cudart. Also off the top of my head not sure which outputs are propagated into the symlinkJoin
| 00:11:43 |
| SomeoneSerge (back on matrix) | Could you publish the full logs? | 00:11:56 |
 hexa (UTC+1) | is there a relationship between cuda and the open nvidia kmod? | 03:40:17 |
| hexa (UTC+1) | because my cuda things stopped working some time after migrating to 24.11 | 03:40:27 |
| hexa (UTC+1) | though nvidia-smi is working | 03:40:47 |
| hexa (UTC+1) | but ollama and wyoming-faster-whisper can't init cuda | 03:41:08 |
| hexa (UTC+1) | will try to drop hardening next, as usual π | 03:42:19 |
| hexa (UTC+1) | ok, DevicePolicy related π | 03:46:59 |
| hexa (UTC+1) | ok, apparently not | 15:44:08 |
| hexa (UTC+1) | nvidia_uvm doesn't get loaded at boot anymore | 15:44:25 |
| hexa (UTC+1) | * it seems like nvidia_uvm doesn't get loaded at boot anymore | 15:44:31 |
| SomeoneSerge (back on matrix) | https://github.com/NixOS/nixpkgs/issues/334180 | 15:52:57 |
| hexa (UTC+1) | fbdcdde Kiskae 2024-05-22 13:46 +0200 308β # Don't add `nvidia-uvm` to `kernelModules`, because we want
fbdcdde Kiskae 2024-05-22 13:46 +0200 309β # `nvidia-uvm` be loaded only after `udev` rules for `nvidia` kernel
fbdcdde Kiskae 2024-05-22 13:46 +0200 310β # module are applied.
fbdcdde Kiskae 2024-05-22 13:46 +0200 311β #
fbdcdde Kiskae 2024-05-22 13:46 +0200 312β # Instead, we use `softdep` to lazily load `nvidia-uvm` kernel module
fbdcdde Kiskae 2024-05-22 13:46 +0200 313β # after `nvidia` kernel module is loaded and `udev` rules are applied.
fbdcdde Kiskae 2024-05-22 13:46 +0200 314β extraModprobeConfig = ''
fbdcdde Kiskae 2024-05-22 13:46 +0200 315β softdep nvidia post: nvidia-uvm
fbdcdde Kiskae 2024-05-22 13:46 +0200 316β '';
| 16:03:03 |
| SomeoneSerge (back on matrix) | Yeah, somehow softdep breaks with the open driver? | 16:03:59 |
| hexa (UTC+1) | nope, reformat | 16:03:59 |
| hexa (UTC+1) | ok, has been there since 2023.11 | 16:04:16 |
| hexa (UTC+1) | and yeah, I'm on the open driver | 16:04:19 |
| SomeoneSerge (back on matrix) | https://github.com/NixOS/nixpkgs/issues/334180#issuecomment-2284518816 | 16:04:30 |
| SomeoneSerge (back on matrix) | atry | 16:04:31 |
| SomeoneSerge (back on matrix) | * Atry | 16:04:33 |
| hexa (UTC+1) | we should probably deduplicate all issues to this one | 16:05:04 |
| 11 Dec 2024 |
|  @magic_rb:matrix.redalder.org joined the room. | 00:50:41 |
| @magic_rb:matrix.redalder.org | cross post from #dev:nixos.org
anyone touch the nvidia driver code? packaging i mean. The long standing bug of "use xrandr twice and you get a segfault in X11" doesn't happen to me anymore apparently
| 00:51:08 |
| hexa (UTC+1) |
copying path '/nix/store/62vk99s9kdcjj4x64wcw22a7rwbfnm36-python3.12-onnxruntime-1.20.1' from 'ssh://hexa@build2.darmstadt.ccc.de'
| 20:11:21 |
| hexa (UTC+1) | π₯³ | 20:11:23 |
| hexa (UTC+1) | now if only the cuda build was working | 20:11:35 |
| hexa (UTC+1) | https://github.com/microsoft/onnxruntime/issues/22855#issue-2662882047 | 20:58:10 |
| hexa (UTC+1) | ah, very coool π« | 20:58:14 |
| 12 Dec 2024 |
 connor (burnt/out) (UTC-8) | Ah right that thing
I was told CMake is supposed to understand itβs a header-only library and not try to actually link against a shared object file, so not sure why itβs doing exactly that and causing the build to fail | 01:48:43 |
