!eWOErHSaiddIbsUNsJ:nixos.org

NixOS CUDA

212 Members
CUDA packages maintenance and support in nixpkgs | https://github.com/orgs/NixOS/projects/27/ | https://nixos.org/manual/nixpkgs/unstable/#cuda43 Servers

Load older messages


SenderMessageTime
23 Jul 2024
@kaya:catnip.eekaya changed their profile picture.05:30:28
@ezzobirbezziou:matrix.orgEzzobir Bezziou joined the room.08:22:21
@sellout:matrix.orgsellout joined the room.13:32:24
@ss:someonex.netSomeoneSerge (utc+3)
In reply to @ss:someonex.net
Sounds like "why aren't we doing this yet?"
Speaking of which, we need to stop copy-pasting this stuff because why aren't these default: https://github.com/NixOS/nixpkgs/pull/328713/files#diff-2d862bf9684df6fdface7aabf2af2b1767eb17ba2f9ab8b2e7db03d22c3c0660R196-R216
13:49:06
@ss:someonex.netSomeoneSerge (utc+3)Should be like a single option13:49:23
@ss:someonex.netSomeoneSerge (utc+3)A single option to blacklist everything, and then you add a few exceptions13:49:45
@hexa:lossy.networkhexa (UTC+1)ideally13:58:01
@trofi:matrix.org@trofi:matrix.org left the room.19:05:15
24 Jul 2024
@phirsch:matrix.orgPascalDoes anyone else have trouble building ollama-cuda from nixos-unstable lately? 05:14:53
@hexa:lossy.networkhexa (UTC+1) Pascal: anything concrete that you are seeing? 09:40:09
@kaya:catnip.eekaya changed their profile picture.21:52:00
25 Jul 2024
@phirsch:matrix.orgPascalCompilation of nixpkgs#ollama-cuda fails with: '/nix/store/4fxpzb7rqf19r5hypxpzbf72jaa6mphw-cuda_cudart-12.2.140-dev/include/cuda_runtime.h:82:10: fatal error: crt/host_config.h: No such file or directory'05:54:18
@phirsch:matrix.orgPascal@hexa: 👆05:55:35
@phirsch:matrix.orgPascalThis is on non-NixOS (Ubuntu 22.04).06:00:40
@phirsch:matrix.orgPascal *

@hexa (UTC+1) 👆

06:02:11
@julius:mtx.liftm.deˈt͡sɛːzaɐ̯I can't reproduce that on a nixos, github:NixOS/nixpkgs/289eafaaf02177e5814d8738cf57e259f9eae46e#ollama-cuda builds fine. (should master from an ~hour ago.)06:26:48
@julius:mtx.liftm.deˈt͡sɛːzaɐ̯* I can't reproduce that on a nixos, github:NixOS/nixpkgs/289eafaaf02177e5814d8738cf57e259f9eae46e#ollama-cuda builds fine. (should master from half an ~hour ago.)06:27:09
@julius:mtx.liftm.deˈt͡sɛːzaɐ̯ * Pascal: I can't reproduce that on a nixos, github:NixOS/nixpkgs/289eafaaf02177e5814d8738cf57e259f9eae46e#ollama-cuda builds fine. (should master from half an ~hour ago.) 06:27:22
@phirsch:matrix.orgPascal
In reply to @julius:mtx.liftm.de
Pascal: I can't reproduce that on a nixos, github:NixOS/nixpkgs/289eafaaf02177e5814d8738cf57e259f9eae46e#ollama-cuda builds fine. (should master from half an ~hour ago.)
Thanks! I'll give that one a go later.
09:17:32
@phirsch:matrix.orgPascal @julius:mtx.liftm.de: Just checked: Works on NixOS, but still fails on Ubuntu (using nix run --impure). Same error: crt/host_config.h not found. Dang... it worked until 3 weeks or so ago. It fails during 'Compiling the CUDA compiler identification source file' oddly using /usr/local/cuda/bin/nvcc despite finding /nix/store/...-cuda_nvcc... 🤷‍♂️ 15:07:36
@ss:someonex.netSomeoneSerge (utc+3)
In reply to @phirsch:matrix.org
@julius:mtx.liftm.de: Just checked: Works on NixOS, but still fails on Ubuntu (using nix run --impure). Same error: crt/host_config.h not found. Dang... it worked until 3 weeks or so ago. It fails during 'Compiling the CUDA compiler identification source file' oddly using /usr/local/cuda/bin/nvcc despite finding /nix/store/...-cuda_nvcc... 🤷‍♂️
Try entering a shell with cudaPackages.cuda_nvcc I guess
15:13:03
@phirsch:matrix.orgPascalSince I don't have any CUDA env vars set nor nvcc on the PATH, I suspect that the nvcc location is explicitly specified in the derivation's build script...15:13:14
@ss:someonex.netSomeoneSerge (utc+3)I didn't check but I'd guess ollama doesn't propagate nvcc (not necessarily bad) and then it ends up picking up the bad nvcc from ubuntu (/usr/...)15:14:17
@phirsch:matrix.orgPascalNow this is weird: Simply running nix run github:NixOS/nixpkgs/289eafaaf02177e5814d8738cf57e259f9eae46e#ollama serve (without -cuda!) works, showing 'library=cuda' and accurate GPU info... Can't claim to understand why, but it looks like I don't actually need the '-cuda' variant. (Suspect that it's somehow managing to dload() the system CUDA SOs?)15:31:57
@phirsch:matrix.orgPascal @ss:someonex.net: Thanks for the tips (only saw your replies after refreshing just now). Might try that as well (although I might already be good according to the above). 15:41:55
26 Jul 2024
@jasonschnitzer:matrix.orgJason Schnitzer joined the room.01:35:00
@nathan72419:matrix.orgnathan72419 joined the room.16:11:45
@connorbaker:matrix.orgconnor (he/him) (UTC-7)As a heads up, I’m moving this week so my availability will probably be more limited than it was already for a few weeks :/16:27:09
27 Jul 2024
@phirsch:matrix.orgPascal

@SomeoneSerge (UTC+3) @ˈt͡sɛːzaɐ̯ No dice... While ollama (without '-cuda') somehow manages to get GPU serial and VRAM allocation into, it doesn't use the GPU when actually running a model (outputs 'Not compiled with GPU offload support'). And unfortunately, using 'nix run --impure' as above from within a nix shell with 'nvcc' from nixpkgs still fails because it's using nvcc from /usr/local/...

07:04:48
28 Jul 2024
@matthewcroughan:defenestrate.itmatthewcroughan changed their display name from matthewcroughan to matthewcroughan - going to nix.camp.16:12:17

Show newer messages


Back to Room ListRoom Version: 9