01:31 ngcortes: anholt, so I'm setting our deqp gl runs to 500 tests per group on 8 cpu cores, but randomly some systems will still take like ~30 minutes to run, which normally a run will take like ~5
01:31 ngcortes: *normally=calling the binary directly
01:31 anholt: 30 minutes to run one 500-test caselist?
01:33 ngcortes: that was for an 8000 test gles2 run
01:34 anholt: if things run slower, then in my experience that's generally hitting swap from spawning that many deqps. not that gles2 should be heavy, but that's the usual explanation.
01:35 ngcortes: anholt, I see, I'll restrict the thread count by installed ram, maybe that would work
01:36 anholt: if it's not that, then throw the usual tools at things -- perf, intel-gpu-top.
01:37 ngcortes: yup roger. thanks anholt!
02:09 ngcortes: anholt, does intel_gpu_top work with the xe kernel?
02:09 anholt: I don't know, I haven't done intel in years.
02:09 ngcortes: word
03:12 robclark: ngcortes: nvtop
09:22 phasta: Trying to rebase a feature branch onto misc-next and it fails – without reason. Has anyone ever seen sth like that with git? https://paste.debian.net/1412192/
09:22 phasta: dma-fence.c doesn't contain any conflict...
10:43 ukleinek: phasta: the lowlevel tooling for that is: `git ls-files -u` and you can do: `git diff :{1,2}:drivers/dma-buf/dma-fence.c` to look at the differences of the different stages
12:02 phasta: I think I found the issue
12:02 phasta: We don't have 033559473dd3b55558b535aa37b8848c207b5cbb in drm-misc-next yet. It's in master, though.
12:03 phasta: also in drm-misc-fixes
12:43 tzimmermann: airlied, sima, FYI there's nothing in drm-misc-fixes this week. hence no PR
13:10 daniels: phasta: does 'Resolved 'drivers/dma-buf/dma-fence.c' using previous resolution.' not imply that you or someone else had already dealt with the conflict and it was saved in rerere, so all you have to do is acknowledge it with update-index?
13:19 phasta: daniels: I think I had accidentally touched that commit on my branch already, yes
13:20 phasta: I can however go back the the original branch.
13:20 phasta: Question is now how I should deal with getting it into drm-misc-next
17:15 dcbaker: zzyiwei: thanks! you got the last commit in 25.3.1