NixOS CUDA | 288 Members | |
| CUDA packages maintenance and support in nixpkgs | https://github.com/orgs/NixOS/projects/27/ | https://nixos.org/manual/nixpkgs/unstable/#cuda | 58 Servers |
| Sender | Message | Time |
|---|---|---|
| 12 Mar 2025 | ||
Just in case, this doesn't merge lists | 01:19:38 | |
Yeah. Then I think that it makes sense not to filter them out automatically. If someone explicitly added somepackage = linux; then I think that we should have an eval error if somepackage doesn't support linux. | 01:20:14 | |
| Yes that was the intent | 01:20:37 | |
| Yes. That was the intended effect. | 01:20:38 | |
So that, for example, someone could do cudaPackages.some_package = [ ]; or something. | 01:21:27 | |
* So that, for example, someone could do cudaPackages.some_package = [ ]; in the "explicit" attrset or something. | 01:21:42 | |
| I see | 01:21:46 | |
| If you want, I can add it back in. But I think that this reordering was one of the reasons for the indentation change that you wanted to avoid))) | 01:23:08 | |
Well, ok. Let's swap the order. And for the packageSets variable and for evalPackageSet they actually don't seem to be doing anything useful. The packageSets there really only serves the purpose of automatically identifying the cuda package sets, so let's rename it to cudaPackageSets as, I think, you've already suggested. Let's not introduce the extra variable for evalPackageSet nor for its image, but instead just use packagePlatforms directly in jobs = let in HERE // { ... = linuxl } | 01:24:30 | |
* Well, ok. Let's swap the order. And for the packageSets variable and for evalPackageSet they actually don't seem to be doing anything useful. The packageSets there really only serves the purpose of automatically identifying the cuda package sets, so let's rename it to cudaPackageSets as, I think, you've already suggested. Let's not introduce the extra variable for evalPackageSet nor for its image, but instead just use packagePlatforms directly in jobs = let in HERE // { ... = linuxl } | 01:24:37 | |
* Well, ok. Let's swap the order. And for the packageSets variable and for evalPackageSet they actually don't seem to be doing anything useful. The packageSets there really only serves the purpose of automatically identifying the cuda package sets, so let's rename it to cudaPackageSets as, I think, you've already suggested. Let's not introduce the extra variable for evalPackageSet nor for its image, but instead just use packagePlatforms directly in jobs = let in HERE // { ... = linux; } | 01:24:42 | |
| wdyt? | 01:24:45 | |
| If we want to add any other package set well we just add another function call there 🤷 | 01:25:52 | |
Ah we maybe want to give a name to mapTestOn . packagePlatforms? | 01:28:28 | |
| How about
| 01:36:27 | |
| Also, if you have any ideas on how to remove the "load-bearing comment" without changing the indentation - be my guest))) | 01:38:26 | |
| * How about
| 01:41:00 | |
| * How about
| 01:41:55 | |
ruro: I'm ok with that, I just thought applying mapAttrs packagePlatforms is not such a complex thing that we really need to be so verbose about it | 08:50:34 | |
| 14:27:37 | ||
| 17 Mar 2025 | ||
| Any ideas why tensorrt from 24.11 is failing to build? I downloaded and imported the sources into nix store as instructed, but it doesn't want to unpack it.
| 12:29:56 | |
The wheel in the archive is called tensorrt-10.3-cp312-none-linux_x86_64.whl, so not sure what's going on | 12:31:21 | |
* The wheel in the archive is called tensorrt-10.3.0-cp312-none-linux_x86_64.whl, so not sure what's going on | 15:14:31 | |
| 16:10:28 | ||
| 18 Mar 2025 | ||
| Michal Koutenský: try manually unpacking and verifying the tarball has it at the same path its expecting; also try adding an echo with the current directory in the derivation to make sure the script is printing where its searching from | 14:58:01 | |
| the tarball has it without the last version number (.26). i managed to get it working yesterday by modifying the unpack phase to extract the correct wheel. is there a chance nvidia changed how the tarball contents look and no one caught it yet? if i had a different tarball i assume the hashes wouldn't match with what is at https://github.com/NixOS/nixpkgs/blob/nixos-24.11/pkgs/development/cuda-modules/tensorrt/releases.nix, but since the build process needs manual intervention i assume there's not much automation for it nor do many people use it | 15:04:09 | |
| looking at the python package, the listed maintainer only had one commit for it in 2022 | 15:05:33 | |
| TRT 10.3 is pretty old | 18:17:09 | |
| Michal Koutenský: i recommend you to use CUDA stuff from master/nixpkgs-unstable - many fixes go there and are not being backported to stable (24.11) | 18:35:39 | |
| for example I updated TRT to 10.8 there (and there is already 10.9, yay) | 18:36:04 | |