| 11 Feb 2025 |
 motiejus | I know there are Frigate users here, and they are less likely to point to "dependency issues" on my Frigate NixOS installation. :)
Frigate is consuming a half-to-full core per camera on "detection" process in an OpenMP call, even though Google Coral is configured, visible and seemingly well. I have taken a CPU profile for the process, and the only thing it says is a C function gomp_barrier_wait_end, which is not helpful.
Any insights would be appreciated, here or in the discussion. :)
| 14:13:17 |
| motiejus | * I know there are Frigate users here, and they are less likely to point to "dependency issues" on my Frigate NixOS installation. :)
Frigate is consuming a half-to-full core per camera on "detection" process in an OpenMP call, even though Google Coral is configured, visible and seemingly well. I have taken a CPU profile for the process, and the only thing it says is a C function gomp_barrier_wait_end, which is not helpful for someone not deep into Frigate internals.
Any insights would be appreciated, here or in the discussion. :)
| 14:13:41 |
 @hexa:lossy.network | gomp | 14:13:52 |
| @hexa:lossy.network | sorry, that is a weird sound | 14:13:57 |
| motiejus | now that you wrote it, kinda like garp from Irving's novel. | 14:14:29 |
| @hexa:lossy.network |
Upgrading to 0.15 is quite some effort if done before 2025.04
| 14:20:18 |
| @hexa:lossy.network | lmao | 14:20:19 |
| @hexa:lossy.network | yes, for me. | 14:20:26 |
| @hexa:lossy.network | if not for you. | 14:20:31 |
| @hexa:lossy.network | thanks for debugging this with the upstream | 14:21:28 |
| @hexa:lossy.network | frigate 0.15 depends on https://github.com/NixOS/nixpkgs/pull/380543/commits/b3def7466fa7f0389c0df2246dce02887f94509c | 14:21:44 |
| motiejus |
Josh's suggestion is correct that something is not installed or configured the same way as it is in the docker container
:/
| 14:44:20 |
| motiejus | so I guess my next step would be to run this in a Docker container and see if I get the same | 14:54:43 |
| motiejus | * so I guess my next step would be to run this in a Docker container and see if I get the same behavior | 14:54:48 |
| motiejus | So it looks like it's indeed an environment-related issue: https://github.com/blakeblackshear/frigate/discussions/16450#discussioncomment-12152067
The first thing that comes to mind is looking at all the applications loaded into that process, and see if any of them have different versions. While sounds like work, but feels like shooting in the dark a bit.
If anyone has better tips, I am all ears. | 20:45:36 |
| motiejus | * So it looks like it's indeed an environment-related issue: https://github.com/blakeblackshear/frigate/discussions/16450#discussioncomment-12152067
The first thing that comes to mind is looking at all the applications loaded into that process, and see if any of them have different versions. While sounds like work, but feels like wandering in the dark a bit.
If anyone has better tips, I am all ears. | 20:45:51 |
| motiejus | Filed https://github.com/NixOS/nixpkgs/issues/381280 for visibility | 21:13:08 |
| motiejus | * Filed https://github.com/NixOS/nixpkgs/issues/381280 | 21:13:18 |
| @hexa:lossy.network | yeah, well | 21:19:52 |
| @hexa:lossy.network | I'm here | 21:19:53 |
| @hexa:lossy.network | I'm not quite sure where to start debugging this further | 21:20:08 |
| @hexa:lossy.network | multiprocessing is a python standard library module | 21:20:55 |
| @hexa:lossy.network | our python is not compiled with pgo | 21:21:04 |
 K900 | I would start by attaching a profiler | 21:21:28 |
| @hexa:lossy.network | but the problem could also be gomp | 21:21:33 |
| K900 | Like py-spy | 21:21:43 |
| K900 | And then maybe perf if it's not in Python code | 21:21:50 |
| @hexa:lossy.network | there was a perf already | 21:21:59 |
| @hexa:lossy.network | https://github.com/blakeblackshear/frigate/discussions/16450#discussioncomment-12143059 | 21:22:13 |
| K900 | I'm pretty sure all the interesting bits are too small to see on this graph | 21:23:35 |
