| 24 Mar 2025 |
 Gaétan Lepage | (I made sure to --max-jobs 1 --cores 0 so I hope the LAV is not over 24) | 16:16:36 |
 emily | --max-jobs 1 --cores 0 will frequently make load average go > nproc | 16:17:55 |
| emily | since the leaf build children are never the only process running on the system | 16:18:24 |
| emily | * since the leaf build children are never the only processes running on the system | 16:18:33 |
| emily | load average: 24.65, 25.85, 26.11 | 16:19:04 |
| Gaétan Lepage | Is 25.5 (current 15min LAV) significantly worse than 24? | 16:19:40 |
| Gaétan Lepage | I can set --cores n-1 next time if this makes a tangible difference. | 16:19:57 |
 Max | make invocations that don't set `-l` are a crime against humanity | 16:20:08 |
| emily | yes, an overloaded system thrashes endlessly and makes it take 10 seconds to fork a single command etc. | 16:22:26 |
| emily | not that 23 is great either on a shared resource | 16:22:33 |
| Gaétan Lepage | Ok I thought that context switching was negligible for LAV~n_cores. | 16:23:40 |
| Gaétan Lepage | Ok it's finally ending... 1h50min to build this monster on 24cores... | 16:24:12 |
| emily | if there's one free core, the system can satisfy any one core job as normal. if there's zero free cores, the system cannot satisfy any request for more compute | 16:31:31 |
| Gaétan Lepage | Ok! | 16:32:28 |
|  appaquet joined the room. | 16:40:30 |
| 25 Mar 2025 |
|  freedbeats joined the room. | 15:24:52 |
|  aleksana 🏳️⚧️ (force me to bed after 18:00 UTC) changed their display name from aleksana (force me to bed after 18:00 UTC) to aleksana 🏳️⚧️ (force me to bed after 18:00 UTC). | 16:59:51 |
|  Marie changed their display name from Marie to Marie Ramlow. | 17:42:38 |
| Marie changed their display name from Marie Ramlow to Marie. | 18:55:51 |
|  @luxzi:matrix.org changed their profile picture. | 19:47:09 |
| emily | Gaétan Lepage:
emily@darwin01 ~ % uptime
23:36 up 2 days, 16:53, 3 users, load averages: 23.45 21.10 19.50
emily@darwin01 ~ % sysctl hw.ncpu
hw.ncpu: 10
| 23:36:52 |
| emily | how many times do we have to do this? | 23:36:55 |
| Gaétan Lepage | killed... python tests leaking. Maybe I should just stop building for darwin entirely. | 23:38:40 |
|  mkg20001 joined the room. | 23:43:42 |
| mkg20001 | for the namespace builders: is there any difference in billing how much resources are used per build? or is it simply the minutes that count? | 23:47:06 |
| 26 Mar 2025 |
|  @ayatsfer:matrix.org joined the room. | 07:37:29 |
| @ayatsfer:matrix.org | Hey, I'm reposting here for visibility: https://github.com/orgs/nix-community/discussions/1748 | 07:37:54 |
| 27 Mar 2025 |
|  sinan changed their profile picture. | 13:10:48 |
| 28 Mar 2025 |
|  @fxomt:pub.solar joined the room. | 14:18:11 |
| 30 Mar 2025 |
| sinan changed their profile picture. | 15:33:33 |
