02:37mareko: why does Mesa run libdisplay-info tests at ninja test, and why are they all failing?
03:21kode54: anything known about why geekbench 6 particle physics test fails validation?
03:21kode54: is it trying to use float16?
04:13jnoorman: anholt: great, I also ran into two compiler crashes when running the dx12 fossils 😅
08:32karolherbst: robclark: already merged
08:33karolherbst: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36158
08:33karolherbst: robclark: fyi, phoronix also did some benchmarks against ROCm the other day: https://www.phoronix.com/review/rocm-rusticl-strix-halo and the results looked quite okayish overall
08:57karolherbst: I'll also looking into the kernel launch overhead thing, because there are multiple smaller places where I could ditch 0.5-1% of CPU overhead each
12:41robclark: karolherbst: ahh, nice
13:22rgallaispou: sima: Hi, I cannot find https://lore.kernel.org/all/20250328114148.260322-1-rgallaispou@gmail.com anywhere in the latest drm-misc. As it been applied ?
13:22rgallaispou: If it's just an oversight, would you mind if I apply it myself with a Link to the upstream commit (as you stated) ? I'll also add mripard's r-b that I forgot to append.
13:51sima: rgallaispou, oops yeah somehow this one didn't make it and was stuck in my tree, sorry about that :-/
13:52sima: rgallaispou, and yeah if you have drm-misc commit rights meanwhile, please feel free to push yourself
14:07rgallaispou: sima: No worries ! I'll modify and push it then
14:08sima: thx a lot!
14:30misyl: Is there a way as an application, I can check if Mesa was compiled with video codecs disabled?
14:30misyl: Aside from just the extensions being missing?
15:13misyl: I guess I can check for video_encode_queue + no h265 I guess
15:18rodrigovivi: tzimmermann: your ack also valid on this other patch: https://lore.kernel.org/dri-devel/aJ-xgHEZEw4am8-i@intel.com/ ?
15:20rodrigovivi: tzimmermann: mlankhorst: your ack also need in this https://lore.kernel.org/dri-devel/aKc4eSmYqhSWgwMI@intel.com/T/#u please
15:21tzimmermann: rodrigovivi, acked
15:32rgallaispou: rodrigovivi: there was a conflict in drivers/gpu/drm/i915/display/intel.tc.c that I had to solve when pushing drm-tip. Can you confirm that I got it right ?
15:38MrCooper: gfxstrand: out of curiosity, in the first graph in the description of https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36827, why exactly did the VS for frame n+1 need to wait for the FS of frame n? What resource do they both use?
15:39rodrigovivi: tzimmermann: thank you!
15:39rodrigovivi: rgallaispou: checking right now
15:43rodrigovivi: rgallaispou: looks good to me...
15:43rodrigovivi: thank you
15:45rgallaispou: rodrigovivi: Thanks a lot ! :)
19:45kisak: I've got a problem with mesa 25.2.1. We got combined arch vulkan icd files that have "library_arch": "64", in them, which makes it non coinstallable with the 32 bit build of mesa-vulkan-drivers.
19:46kisak: I think I need to revert https://gitlab.freedesktop.org/mesa/mesa/-/commit/f7aa6ba9064f64885a2ae24af8d51c7671d20b6d to get the build usable for now.
20:17psykose: "combined arch"? why would a line in a json file make something non coinstallable, if the 64 one says 64 and the 32 one says 32?
20:17psykose: or do you have one icd.json for both somehow
20:28kisak: yeah https://salsa.debian.org/xorg-team/lib/mesa/-/commit/82eed41b2456c20eac16f86456a8c583753a94f2 and that commit are fighting
20:32kisak: I'll disable the conflict on the packaging side