| 1 Oct 2025 |
 Winter | * runtime, after i've successfully imported tf -- i then get some JIT compilation error | 20:15:31 |
| Winter | seems like XLA_FLAGS may be my friend here | 20:16:43 |
| Winter | ok yeah setting XLA_FLAGS=--xla_gpu_cuda_data_dir=/nix/store/y16bl5h9nxdbyfs922x4bz9lkk51kx1d-cuda_nvcc-12.8.93 fixed it! | 20:20:09 |
| Winter | i think this is just a simple fix in tensorflow-bin | 20:20:19 |
| Winter | i don’t particularly want to build tf-bin again but i’m going to get a PR up to fix this :-) | 21:57:28 |
| Winter | i assume we have the concept of a check for whatever cuda version is in use — connor (he/him) (UTC-7) do you want a check for <13 before 13 is actually in, as a reminder of sorts to test this out with 13? (still need an MRE tho) | 21:58:24 |
| Winter | for clarity this is some JIT’d XLA stuff | 21:59:17 |
 connor (burnt/out) (UTC-8) | IIRC on 13 I symlink it so it should still be available in NVCC’s bin output | 22:38:35 |
| Winter | awesome | 22:39:52 |
| 2 Oct 2025 |
 Daniel Fahey | Hello there. Anyone got any spare time (and compute, lol) to help me upgrade vLLM in Nixpkgs? https://github.com/NixOS/nixpkgs/pull/447722 | 09:46:34 |
 Gaétan Lepage | Sure! | 12:22:51 |
| Gaétan Lepage | I pushed to the branch and fixed the CUDA build (hopefully) | 12:37:39 |
| Gaétan Lepage | Well, unfortunately:
/nix/store/ycslb9zwv05yjp79ysvbwflwa9s9ffa8-source/include/cutlass/epilogue/collective/detail.hpp(719): error: no instance of function template "cutlass::epilogue::collective::detail::Sm100TmaWarpSpecializedAdapter<EpilogueOp>::operator()
| 12:41:39 |
| Daniel Fahey | Legend! Beware I just force pushed | 12:56:59 |
| Daniel Fahey | (to fix a typo in the commit message) | 12:57:07 |
| Gaétan Lepage | Pushed new fixes. I initially had the wrong version of cutlass | 12:57:12 |
| Daniel Fahey | woops, could you add you co-authorship again please | 12:58:19 |
| Daniel Fahey | lost it when I did a git commit --amend | 12:58:37 |
| Gaétan Lepage | I might have force-pushed right after you | 13:04:07 |
| Daniel Fahey | No worries. | 13:04:15 |
| Daniel Fahey | Looks okay, I'm going to force push again with the corrected commit message crediting you | 13:05:36 |
| Gaétan Lepage | Ok! | 13:12:44 |
| Gaétan Lepage | It built successfully on my system! Running a full nixpkgs-review with cuda support right now. | 13:28:48 |
| Daniel Fahey | Looking forward to seeing it build on https://hydra.nixos-cuda.org hehe | 13:32:04 |
| 3 Oct 2025 |
| Daniel Fahey | Gaétan Lepage: good morning! Think I got to the bottom of the failing Python 3.12 CPU build. Can I please have your opinion on https://github.com/NixOS/nixpkgs/pull/447722#issuecomment-3364822620 think I'll just mark Python 3.12 CPU builds as broken? | 08:49:51 |
| 4 Oct 2025 |
|  @palasso:matrix.org left the room. | 10:43:55 |
 lon | Oh btw, v0.11 is out now https://github.com/vllm-project/vllm/releases/tag/v0.11.0 | 12:33:08 |
| lon | Thank you all for making this package work and keeping it so up to date, you rock ❤️ | 12:33:56 |
| Daniel Fahey | Gaétan Lepage: I think this PR is good to merge, let's ship | 14:09:13 |
| Daniel Fahey | No one likes us, we don't care! We are Nix, super Nix https://repology.org/project/python%3Avllm/versions | 14:14:54 |
