NixOS CUDA | 290 Members | |
| CUDA packages maintenance and support in nixpkgs | https://github.com/orgs/NixOS/projects/27/ | https://nixos.org/manual/nixpkgs/unstable/#cuda | 57 Servers |
| Sender | Message | Time |
|---|---|---|
| 3 Nov 2024 | ||
| do you remember what issue/bug it was? | 00:04:40 | |
In reply to @ss:someonex.netreally nice, I was just gonna propose we run wget --recursive on https://developer.download.nvidia.com/compute/cuda/repos/runfile/x86_64/ | 00:16:04 | |
| much prefer the feature extraction stuff, that's wicked | 00:16:23 | |
| one problem with the runfile json is it excludes certain components, eg: nccl | 00:17:14 | |
| 4 Nov 2024 | ||
| Is there a reason why https://github.com/NixOS/nixpkgs/pull/291471 is not merged? | 13:46:49 | |
In reply to @snektron:matrix.orgStuck on unvendoring qt libraries it seems | 15:58:17 | |
What are your thoughts on nixGL and nix-gl-host? Specifically as it relates to ensuring people use the same driver version on both NixOS and non-NixOS host machines | 22:43:59 | |
| It's a disagreeable opinion but the whole "inspect the system at eval time" business in nixGL makes me want to scream "why". Even the idea of requiring a running nix-daemon in order to load the drivers at the program's startup... I think it's very backwards. I think nix-gl-host decomposes tasks better, but
| 23:07:53 | |
| * It's a disagreeable opinion but the whole "inspect the system at eval time" business in nixGL makes me want to scream "why". Even the idea of requiring a running nix-daemon in order to load the drivers at the program's startup... I think it's very backwards. I think nix-gl-host decomposes tasks better, but
| 23:08:04 | |
| FWIW I see no reason not to implement both approaches (host's libs vs. nixpkgs' libs that match the running kernel) in one tool | 23:09:19 | |
That could be just two flags: --drivers-from={HOST|<nix expr or a flake ref>} and --export=/run/opengl-driver/lib as the alternative to --run <CMD> | 23:12:23 | |
| (and then of course there's this whole libcapsule business that's waiting to be tried) | 23:13:40 | |
| Note there was recently somebody in nix-gl-host issues advertising their Rust rewrite | 23:29:51 | |
| 5 Nov 2024 | ||
| Has anyone successfully built blender on unstable with CUDA support enabled recently? | 02:31:45 | |
| 🤔 Currently, openusd fails in patch phase. | 08:39:43 | |
In reply to @aidalgol:matrix.orglooks like it's broken since end of October: https://hydra.nix-community.org/build/1720981 | 15:13:24 | |
In reply to @ss:someonex.netDid you take a look at https://github.com/NVIDIA/nvidia-container-toolkit already? I think they are battling with similar issues, plus sandboxing on top. It seems like they are using ld.so.cache as a loading mechanism instead of the LD_* env var and that might be more robust? | 15:16:02 | |
In reply to @zimbatm:numtide.comYes that's what we use for docker/podman on nixos | 15:16:44 | |
| I'm not sure if it's worth reusing because nixGL and nix-gl-host are solving a more general problem; the ctk just assumes an FHS environment, we have to patch it and we have to patch its outputs to make them usable on nixos | 15:19:32 | |
| But yes we should keep in mind the general idea of exporting ld.so.cache. We actually used it at least for some time for the singularity containers | 15:21:25 | |
In reply to @zimbatm:numtide.com
Well here's the first offender | 15:22:59 | |
| Ah isn't it nice to have a per-attribute build history | 15:23:36 | |
| So odd that wasn't a thing with hercules | 15:24:15 | |
| 23:25:06 | ||
| 6 Nov 2024 | ||
| Oh my god another TensorRT release | 17:06:11 | |
| I really gotta finish up stuff I’m working on | 17:06:49 | |
| 7 Nov 2024 | ||
| wanting to figure out the cost of evaluating a single attribute led me to make https://github.com/ConnorBaker/nix-eval-jobs-python | 06:06:20 | |
wanting to figure out the impact malloc replacements have on nix eval led me to make https://github.com/ConnorBaker/tune-nix-evalstill excited about this one -- plan to add support to optimize zram parameters at some point to see if it's worthwhile to just always disable GC so long as there's RAM (or compressed RAM) available | 06:07:36 | |
| 8 Nov 2024 | ||
| Pointed out to me today by a coworker:
(yes, it copies the entirety of nixpkgs from one path in the store to another path in the store) | 08:20:41 | |
| * Pointed out to me today by a coworker:
(yes, it copies the entirety of nixpkgs from one path in the store to another path in the store) | 08:20:48 | |