| 1 Jan 2025 |
|  NixOS Moderation Botchanged room power levels. | 14:27:04 |
 hexa | well you seem to manage the big-parallel feature manually for arm64 and I only deployed on non-big-parallel builder so far | 15:04:28 |
| hexa | * well you seem to manage the big-parallel feature manually for the existing arm64 box and I only deployed on non-big-parallel builder so far | 15:04:40 |
| hexa | it is certainly not for a lack of compute 🙂 | 15:04:52 |
 VladimÃr ÄŒunát | Yeah. The complication is that for a proper big-parallel machine you need to also increase the make-level parallelism, so that's a little harder to switch as needed. | 15:05:51 |
| VladimÃr ÄŒunát | But maybe we'll figure out something in time. | 15:06:08 |
| VladimÃr ÄŒunát | Build ingestion speed is what bothers me more. | 15:06:30 |
| hexa | I can spin up a machine to always handle big-parallel builds | 15:06:30 |
| hexa | yeah, absolutely | 15:06:36 |
| hexa | I' | 15:06:42 |
| hexa | * I'm saying we might not need all the machines we requested after all | 15:06:55 |
| hexa | but that remains to be seen 🙂 | 15:07:03 |
 Gaétan Lepage | In reply to @emilazy:matrix.org
the new parallel eval stuff is significantly faster Well, actually not really.
I have been working on the PR in nixpkgs-review. The code is ready, but I was very disappointed by my initial testing.
Maybe I have missed something, but using the parallel eval is actually 2x slower than the naive nix-env based implementation.
If you have some hindsights to share, feel free to hop in the PR thread :)
It would be great to improve eval times there indeed | 19:43:15 |
| 2 Jan 2025 |
|  error_pro joined the room. | 10:44:26 |
 K900 | Is Hydra down | 19:43:13 |
 emily | works for me (the site) | 19:46:01 |
| hexa | hm, yes | 19:56:15 |
| hexa | 
Download image.png | 19:56:28 |
| hexa | no scrape, no bueno | 19:56:32 |
| hexa | maybe we should ask the person running uptime-kuma against hydra 🤔 | 19:58:42 |
| hexa | 1.23.13 — points to nixos-24.05 | 20:00:09 |
| hexa | so on secon thought, maybe not 😄 | 20:00:15 |
| hexa | * so on second thought, maybe not 😄 | 20:00:20 |
| 3 Jan 2025 |
|  elikoga changed their display name from elikoga (@38c3 📞488{0,1,9}) to elikoga. | 10:27:59 |
|  @perchun:matrix.org joined the room. | 14:33:46 |
| @perchun:matrix.org | how often does public darwin builder change keys? | 14:34:07 |
| hexa | the community builder? rarely | 14:34:35 |
| hexa | they recently migrated to an m4 machine, which resulted in new hostkeys | 14:34:56 |
| @perchun:matrix.org | I got REMOTE HOST IDENTIFICATION HAS CHANGED again (third time for the 6 month I believe) | 14:34:59 |
