| 28 May 2025 |
 srestegosaurio (on tchncs.de) | In reply to @pj:place.org
anyone else have nvidia build failures when upgrading to 25.05 ? I'm getting errors but I'm using both the beta drivers and a custom kernel so idk. | 09:20:41 |
|  crgre joined the room. | 15:17:24 |
 K900 | Probably last Mesa bump for 25.05: https://github.com/NixOS/nixpkgs/pull/411793 | 16:42:57 |
|  catzzo joined the room. | 16:54:36 |
|  @normalcea:matrix.org changed their display name from normalcea to normalcea 🏳️⚧️. | 21:28:10 |
| @normalcea:matrix.org | The performance of NVK isn't at the level of the proprietary driver for very demanding tasks. | 21:32:30 |
| @normalcea:matrix.org | At least when I tested Hitman: WOA on my 1650 card. | 21:33:15 |
| @normalcea:matrix.org | * At least when I tested Hitman: WOA on my 1650 super card. | 21:35:18 |
| K900 | 1650 Super is probably the worst case for NVK | 21:37:18 |
| K900 | It's the lowest end thing they support | 21:37:23 |
 olivia | yeah, at least last time I was looking nvk is faster on some cpu-bound tasks, but generally slower on gpu-bound | 21:49:25 |
| olivia | (and much slower on ~everything pre-turing because no reclocking) | 21:49:41 |
| K900 | I think it's also faster on newer hardware because more of it is reverse engineered by now? | 21:51:49 |
