| 10 Aug 2024 |
 Gaétan Lepage | Should we remove the rm -rf cmake/Modules_CUDA_fix/{upstream,FindCUDA.cmake} line in our torch derivation or should we patch the regex in the cmake derivation to properly parse 9.0a ? | 21:00:00 |
 connor (burnt/out) (UTC-8) | That seems familiar | 23:08:17 |
| connor (burnt/out) (UTC-8) | Oh maybe check these out | 23:08:27 |
| connor (burnt/out) (UTC-8) | https://github.com/ConnorBaker/nixpkgs/commits/feat/python3Packages.torch/ | 23:08:31 |
| connor (burnt/out) (UTC-8) | I know I was looking at fixes for the torch derivation in general (I guess two months ago at this point now) and that was one of the things I ran into | 23:09:08 |
| connor (burnt/out) (UTC-8) | My preference is always to patch; IMHO Nixpkgs is a really easy way to reliably build these tools for cutting edge compute capabilities and I’d like that to continue | 23:10:00 |
| 11 Aug 2024 |
| Gaétan Lepage | Ok, thank you very much for sharing your patch ! | 08:48:12 |
| 12 Aug 2024 |
 SomeoneSerge (back on matrix) | Kevin Mittman any idea why the softdep trick might not work with the open driver? $ sudo nvidia-settings
ERROR: libEGL setup error : libEGL.so.1: cannot open shared object file: No such file or directory | 17:16:38 |
| SomeoneSerge (back on matrix) | * Kevin Mittman any idea why the softdep trick might not work with the open driver? $ sudo nvidia-settings
https://github.com/NixOS/nixpkgs/issues/334180#issuecomment-2284535323 | 17:16:47 |
| SomeoneSerge (back on matrix) | * Kevin Mittman any idea why the softdep trick might not work with the open driver?
https://github.com/NixOS/nixpkgs/issues/334180#issuecomment-2284535323 | 17:16:51 |
| SomeoneSerge (back on matrix) | * Kevin Mittman any idea why the softdep trick might not work with the open driver?
https://github.com/NixOS/nixpkgs/issues/334180 | 17:16:54 |
| connor (burnt/out) (UTC-8) | In reply to @glepage:matrix.org
Ok, thank you very much for sharing your patch ! Please feel free to use more than just the patch too; it might be a while before I’m able to revisit that particular work | 18:07:32 |
| connor (burnt/out) (UTC-8) | I’m probably going to be focused fairly heavily on Jetson related stuff for a while | 18:07:46 |
