!eWOErHSaiddIbsUNsJ:nixos.org

NixOS CUDA

275 Members
CUDA packages maintenance and support in nixpkgs | https://github.com/orgs/NixOS/projects/27/ | https://nixos.org/manual/nixpkgs/unstable/#cuda56 Servers

Load older messages


SenderMessageTime
1 Nov 2025
@eveeifyeve:matrix.org@eveeifyeve:matrix.org joined the room.04:47:07
@eveeifyeve:matrix.org@eveeifyeve:matrix.org A good question does cuda_cccl include thrust/context.h because with hydra it's receiving a lot of errors saying it's missing for pytorch.
https://hydra.nixos.org/build/311450648/log
04:47:31
@eveeifyeve:matrix.org@eveeifyeve:matrix.orgI am trying to build pytorch right now.04:47:44
@ss:someonex.netSomeoneSerge (back on matrix)
$ nix-build '<nixpkgs>' --arg config '{allowUnfree = true;}' -A cudaPackages.cuda_cccl.dev
$ readlink result-dev
/nix/store/yvvs83nys6i78fq1p5dgliqlhlk2svq0-cuda_cccl-12.8.90-dev
$ ls result-dev/include/thrust/context.h
ls: cannot access 'result-dev/include/thrust/context.h': No such file or directory
05:37:53
@eveeifyeve:matrix.org@eveeifyeve:matrix.orgThis effects building pytorch.05:39:49
@eveeifyeve:matrix.org@eveeifyeve:matrix.org* This effects building pytorch which is will be a major blocker for zero builds hydra 25.11.05:41:36
@ss:someonex.netSomeoneSerge (back on matrix) Perhaps Gaétan Lepage's looked into this? 05:46:57
@ss:someonex.netSomeoneSerge (back on matrix)In any case, see you tomorrow05:47:21
@eveeifyeve:matrix.org@eveeifyeve:matrix.orgI might try to debug this. Because I am happy to.05:53:52
@eveeifyeve:matrix.org@eveeifyeve:matrix.orgIt's complex.h sorry..06:00:31
@eveeifyeve:matrix.org@eveeifyeve:matrix.orgIt's there, but idk something werid about pytorch package. 06:01:30
@eveeifyeve:matrix.org@eveeifyeve:matrix.orghttps://github.com/NixOS/nixpkgs/issues/45745606:02:36
@glepage:matrix.orgGaétan Lepage
In reply to @eveeifyeve:matrix.org
A good question does cuda_cccl include thrust/context.h because with hydra it's receiving a lot of errors saying it's missing for pytorch.
https://hydra.nixos.org/build/311450648/log
I am aware that torchWithRocm is broken. I mentioned it to @lt1379:matrix.org a few days ago.
09:02:40
@glepage:matrix.orgGaétan LepageThe regular pytorch builds fine AFAIK09:03:36
@leona:leona.isleona joined the room.09:15:42
@lt1379:matrix.orgLunTracked it down to USE_FBGEMM_GENAI doing some hacky things with hip flags, will have a PR up soon if that was the only build issue. (Just turning it off, it's some specialized quantized kernels for one ISA only right now so really doesn't seem worth trying to fix the jank)18:55:34
@glepage:matrix.orgGaétan Lepage changed their profile picture.22:52:58
@glepage:matrix.orgGaétan Lepage changed their profile picture.22:54:00
@eveeifyeve:matrix.org@eveeifyeve:matrix.org left the room.23:19:24
2 Nov 2025
@connorbaker:matrix.orgconnor (burnt/out) (UTC-8) changed their display name from connor (burnt/out) (UTC-7) to connor (burnt/out) (UTC-8).08:13:06
@glepage:matrix.orgGaétan Lepage RE cuda_nvcc leaking into nccl:
As a sidenote, I realized that one could remove cuda_nvcc from nativeBuildInputs and (getInclude cuda_nvcc) from buildInputs without breaking nccl's build.
This most probably works because of the makeFlags.
Unfortunately, this does not help with the leakage.
10:58:02
@felix:neode.sefelix joined the room.14:30:39
@connorbaker:matrix.orgconnor (burnt/out) (UTC-8) Gaétan Lepageis https://github.com/NixOS/nixpkgs/pull/457803 ready to merge? I’ll approve and merge if so 15:15:24
@glepage:matrix.orgGaétan Lepage It fixes the leak for nccl, but firefox gets gcc-wrapper from onnxruntime too. 16:42:16
@glepage:matrix.orgGaétan LepageI'm about to push a commit that handles that too. I'm compiling rn.16:42:28
@glepage:matrix.orgGaétan Lepage Rebuilt onnxruntime. It now doesn't depend on cuda_nvcc at runtime.
I'm now rebuilding firefox which should not have cuda_nvcc in its closure anymore.
17:07:36
@glepage:matrix.orgGaétan Lepage

😭 cudaPackages.cuda_cudart depends on cudaPackages.cuda_nvcc at runtime too!!!

Not because of a path leak in the binary this time, just because nvcc is in cudart's propagatedBuildInputs (I think?)

❯ nix why-depends --precise $(nom-build --arg config '{ allowUnfree = true; cudaSupport = true; }' -A firefox-unwrapped) $(nom-build --arg config '{ allowUnfree = true; cudaSupport = true; }' -A cudaPackages.cuda_nvcc)
Finished at 18:16:53 after 1s
Finished at 18:16:53 after 0s
/nix/store/yy1z5y3iql9r4kpslxnjdwcygx52ssl8-firefox-unwrapped-144.0.2
└───lib/firefox/libonnxruntime.so: …st be specified....../nix/store/jk4a7v44fc83ykc15b31r4m21yqc92sp-onnxruntime-1.22.2/lib/.....onn…
    → /nix/store/jk4a7v44fc83ykc15b31r4m21yqc92sp-onnxruntime-1.22.2
    └───lib/libonnxruntime_providers_cuda.so: …nn-9.13.0.50-lib/lib:/nix/store/80x699lyc99dahf85iqdv6z1f0vv6vz2-cuda12.8-cuda_cudart-12.8.90/li…
        → /nix/store/80x699lyc99dahf85iqdv6z1f0vv6vz2-cuda12.8-cuda_cudart-12.8.90
        └───nix-support/propagated-build-inputs: …fhjm-setup-cuda-hook /nix/store/ygd3s9zm1pf77n3q3ac63v58www5scbc-cuda12.8-cuda_nvcc-12.8.93 /nix…
            → /nix/store/ygd3s9zm1pf77n3q3ac63v58www5scbc-cuda12.8-cuda_nvcc-12.8.93
18:19:31
@glepage:matrix.orgGaétan Lepage Actually, rebasing my PR on top of [SomeoneSerge (back on matrix)'s](https://github.com/NixOS/nixpkgs/pull/457424) worked! 20:15:56
@glepage:matrix.orgGaétan Lepage * Actually, rebasing my PR on top of Serge's worked! 20:16:12
3 Nov 2025
@connorbaker:matrix.orgconnor (burnt/out) (UTC-8)Are they good to go or do they need more testing?00:25:39

Show newer messages


Back to Room ListRoom Version: 9