| 25 Oct 2025 |
 asa | also for anyone with possibly more knowledge on this, tensorrt downloads are public endpoints (https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/10.9.0/tars/TensorRT-10.9.0.34.Linux.x86_64-gnu.cuda-12.8.tar.gz) that you can get without authenticating
would it be a license violation to directly include that in nixpkgs with the relevant package license instead of via separately joining their developer program? | 22:24:21 |
 Daniel Fahey | Can't remember where I saw this link, I though it was in here, anyone know of any other discussion / context re: Nix? It reports Flox are signed up? https://developer.nvidia.com/blog/developers-can-now-get-cuda-directly-from-their-favorite-third-party-platforms/ | 22:59:54 |
| Daniel Fahey | DuckDuckGo was better at searching the Discouse, heh, https://discourse.nixos.org/t/nix-flox-nvidia-opening-up-cuda-redistribution-on-nix/69189 | 23:02:41 |
 connor (burnt/out) (UTC-8) | asa: only the 10-series are available without a login.
They also don't publish the URLs anywhere public except on their TensorRT OSS repo (only for current releases). You can see the URLs for previous releases assuming you log in to NVIDIA's website.
See https://github.com/NixOS/nixpkgs/pull/437723; I've got a helper bash script which can be used to generate a manifest file similar to what NVIDIA provides for their other binary archives. | 23:03:12 |
| connor (burnt/out) (UTC-8) | As for having only x86_64-linux releases -- that's mostly been because releases for Jetson were few and far in between and I don't have access to hopper or blackwell to verify ARM releases function correctly | 23:04:10 |
| connor (burnt/out) (UTC-8) | As I'm working on https://github.com/nixos-cuda/cuda-legacy/pull/1 I'll be making manifests for older versions of TensorRT available publicly | 23:06:27 |
| asa | oh i see
do the arm builds only run on jetsons? | 23:06:38 |
| Daniel Fahey | * https://developer.nvidia.com/blog/developers-can-now-get-cuda-directly-from-their-favorite-third-party-platforms/ | 23:06:55 |
| Daniel Fahey | * https://discourse.nixos.org/t/nix-flox-nvidia-opening-up-cuda-redistribution-on-nix/69189 | 23:07:01 |
| asa | also this is very useful thanks, i didn't know something like cuda-legacy existed | 23:08:50 |
| connor (burnt/out) (UTC-8) | NVIDIA has historically published two "arm" builds: linux-sbsa and linux-aarch64. These both map to Nixpkgs' aarch64-linux platform, but linux-sbsa is for server-grade ARM and linux-aarch64 for Jetsons.
Going forward, it seems new Jetson devices will also use linux-sbsa (e.g., the Jetson Thor). It remains to be seen whether older Jetson devices will be supported on CUDA 13. | 23:09:20 |
| connor (burnt/out) (UTC-8) | It's still very much a work in progress and best-effort, so beware 👻 | 23:10:04 |
| Daniel Fahey | After diffing the inputDrvs
diff <(nix derivation show --impure -I nixpkgs=https://github.com/daniel-fahey/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/daniel-fahey/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
37c37
< zv7b6x3i1k6vgmlx4ll74a676znpp19l-python3.13-outlines-1.2.3.drv
---
> b0yp229czgvgsxay9vw8mq0xraa0w94l-python3.13-outlines-1.2.3.drv
| 23:17:03 |
| Daniel Fahey | * After diffing the inputDrvs
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
37c37
< zv7b6x3i1k6vgmlx4ll74a676znpp19l-python3.13-outlines-1.2.3.drv
---
> b0yp229czgvgsxay9vw8mq0xraa0w94l-python3.13-outlines-1.2.3.drv
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.outlines' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.outlines' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
15c15
< fcsb1iizwb50wiag876xzj9kpd8vpa9s-python3.13-jax-0.8.0.drv
---
> 2vxlyjm486swfp89iwsacyx52rslq4xj-python3.13-jax-0.8.0.drv
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.jax' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.jax' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
0a1
> 521737jxrb1akanvm7df9zmx3236rwc6-af5c211d49f3b99447db2252d2cc2b8e0fb54d1c.patch?full_index=1.drv
That last patch is the latest fix to JAX. Maybe a red herring, but I'm running a fresh build of vLLM to see if the new JAX version (and hence Outlines) breaks vLLM.
| 23:21:15 |
| Daniel Fahey | * After diffing the inputDrvs
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
37c37
< zv7b6x3i1k6vgmlx4ll74a676znpp19l-python3.13-outlines-1.2.3.drv
---
> b0yp229czgvgsxay9vw8mq0xraa0w94l-python3.13-outlines-1.2.3.drv
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.outlines' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.outlines' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
15c15
< fcsb1iizwb50wiag876xzj9kpd8vpa9s-python3.13-jax-0.8.0.drv
---
> 2vxlyjm486swfp89iwsacyx52rslq4xj-python3.13-jax-0.8.0.drv
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.jax' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.jax' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
0a1
> 521737jxrb1akanvm7df9zmx3236rwc6-af5c211d49f3b99447db2252d2cc2b8e0fb54d1c.patch?full_index=1.drv
That last patch is the latest fix to JAX. Maybe a red herring, but I'm running a fresh build of vLLM to see if the new JAX version (and hence Outlines) breaks the build.
| 23:22:25 |
| Daniel Fahey | * After diffing the inputDrvs
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
37c37
< zv7b6x3i1k6vgmlx4ll74a676znpp19l-python3.13-outlines-1.2.3.drv
---
> b0yp229czgvgsxay9vw8mq0xraa0w94l-python3.13-outlines-1.2.3.drv
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.outlines' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.outlines' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
15c15
< fcsb1iizwb50wiag876xzj9kpd8vpa9s-python3.13-jax-0.8.0.drv
---
> 2vxlyjm486swfp89iwsacyx52rslq4xj-python3.13-jax-0.8.0.drv
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.jax' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.jax' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
0a1
> 521737jxrb1akanvm7df9zmx3236rwc6-af5c211d49f3b99447db2252d2cc2b8e0fb54d1c.patch?full_index=1.drv
That last patch is the latest fix to JAX. Might be a red herring, but I'm running a fresh build of vLLM to see if the new JAX version (and hence Outlines) breaks the build.
| 23:23:44 |
| Daniel Fahey | * After diffing the inputDrvs
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.vllm' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
37c37
< zv7b6x3i1k6vgmlx4ll74a676znpp19l-python3.13-outlines-1.2.3.drv
---
> b0yp229czgvgsxay9vw8mq0xraa0w94l-python3.13-outlines-1.2.3.drv
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.outlines' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.outlines' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
15c15
< fcsb1iizwb50wiag876xzj9kpd8vpa9s-python3.13-jax-0.8.0.drv
---
> 2vxlyjm486swfp89iwsacyx52rslq4xj-python3.13-jax-0.8.0.drv
$ diff <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/3d2db9755e7815937fb7b8f089fad9b44bc416d8.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.jax' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2) <(nix derivation show --impure -I nixpkgs=https://github.com/NixOS/nixpkgs/archive/6a08e6bb4e46ff7fcbb53d409b253f6bad8a28ce.tar.gz --expr 'with import <nixpkgs> { }; python313Packages.jax' | jq -r '.[] | .inputDrvs | keys[]' | sort -t'-' -k2)
0a1
> 521737jxrb1akanvm7df9zmx3236rwc6-af5c211d49f3b99447db2252d2cc2b8e0fb54d1c.patch?full_index=1.drv
That last patch is the latest fix to JAX. Might be a red herring, but I'm running a fresh build of vLLM to see if the new JAX fix (and hence Outlines derivation) breaks the build.
| 23:26:34 |
| 26 Oct 2025 |
| Daniel Fahey | The above is just why the hash changed. The nvcc error : '"$CICC_PATH/cicc"' died due to signal 9 (Kill signal) means the process was forcibly killed, probably by the out-of-memory killer, Hydra should really restart these, possibly with fewer CPU cores dynamically. | 00:28:50 |
| Daniel Fahey | * The above is just why the hash changed. The nvcc error : '"$CICC_PATH/cicc"' died due to signal 9 (Kill signal) means the process was forcibly killed, probably by the out-of-memory (OOM) killer, Hydra should really restart these, possibly with fewer CPU cores dynamically. | 00:29:10 |
| Daniel Fahey | built, lol https://nixbuild.net/builds/4745574?t=EtkBCm8KBWJ1aWxkCgpidWlsZDpyZWFkGAMiCQoHCAcSAxD5BSINCgsIBBIHOgUKAxiBCDImCiQKAggbEgYIBRICCAUaFgoECgIIBQoICgYg-OSj1AYKBBoCCAAyFgoUCgIIGxIOCAISAxiACBIFEObSoQISJAgAEiBakpZI6g3gNTHtwvNqaB3b4-Rv07_-fD8p_DmOt4YJvhpAjaEDvlgwE_0ZXUm23Ey-3fjcQrW_Ss6bRdz-G3HAoFRMVKlpSYx-cvq0vAJ_HuUdirLhAo9KMjHP4aJ50uxSBCIiCiAwUUMo9gXtYC6TYvCVyceuEvf91HzzfSXG_wyt6NV4oA== | 01:00:33 |
 Gaétan Lepage | Introducing ExecuTorch 1.0: Powering the next generation of edge AI | 23:56:03 |
| 27 Oct 2025 |
| connor (burnt/out) (UTC-8) | Meh
Did you see Helion (https://www.youtube.com/watch?v=MBOPzfl1JBo and https://github.com/pytorch/helion)? I swear they’re just going to keep re-inventing compilers with various degrees of auto-tuning built in | 04:41:04 |
| connor (burnt/out) (UTC-8) | SomeoneSerge (back on matrix)I intend to merge the CUDA 13/packaging refactor PR tomorrow; I think it’s important to get it in so we can iterate on it | 05:14:52 |
| Gaétan Lepage | nccl bump should be ready to go: https://github.com/NixOS/nixpkgs/pull/438735 | 14:25:40 |
| connor (burnt/out) (UTC-8) | Set the CUDA 13 PR to merge, here’s hoping that goes in smoothly | 15:35:29 |
| connor (burnt/out) (UTC-8) | Reminder of things which still need to be addressed: https://matrix.to/#/!eWOErHSaiddIbsUNsJ:nixos.org/$7F36aNKTb-Fmt-RJjjM4Tp1c_hyVQaWhXpXYQj6jKFA?via=nixos.org&via=matrix.org&via=nixos.dev | 15:36:09 |
| connor (burnt/out) (UTC-8) | Also remove the stupid aliases in tree like torchWithCuda. They’re footguns and give people false hope. | 18:16:46 |
| Gaétan Lepage | connor (he/him) (UTC-7): why are so many cudaPackages.* packages broken? | 21:52:21 |
| 28 Oct 2025 |
| connor (burnt/out) (UTC-8) | What’s broken? | 01:25:31 |
| connor (burnt/out) (UTC-8) | Also if you do eval or build with —trace-verbose it should show you | 01:25:41 |
