00:23 fdobridge: <a​irlied> Uapi also enables a lot of tests
00:23 fdobridge: <g​fxstrand> It does but I'm able to run on Turing with 18 threads and not lock up the machine.
00:24 fdobridge: <g​fxstrand> That's a pretty good stress test
00:24 fdobridge: <g​fxstrand> I do get spurious faults but they don't bring it down.
00:25 fdobridge: <g​fxstrand> I should try upping it back to 36 just to see what happens. 🤔
00:29 fdobridge: <g​fxstrand> Oh... You probably want to disable a few, though. I disable the max objects tests as well as graphicsfuzz
05:05 fdobridge: <g​eorgeouzou> I am at the drm-misc-next branch with all the latest nouveau changes.
05:05 fdobridge: <g​eorgeouzou> I have a ryzen 7900 and a rtx 2070 and no wifi at all.
05:06 fdobridge: <g​eorgeouzou> I will disable the tests you mention and report back
09:33 fdobridge: <g​eorgeouzou> Hmm when I run it with 12 threads instead of the full 24 threads it seems to work ok
10:16 fdobridge: <k​arolherbst🐧🦀> @mhenning I was scanning through old nouveau MRs and I saw your aux/pipebuffer one https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18052 wondering if we want to get that in
11:51 fdobridge: <i​shitatsuyuki> How does nvidia update page tables? DMA or BAR writes?
11:51 fdobridge: <i​shitatsuyuki> Wondering because I'm stumbling upon some amdgpu memes where page table updates are currently done one-by-one taking one submission per BO
11:53 fdobridge: <a​irlied> Bar writes now. Maybe dma later esp for sparse maps
15:26 fdobridge: <m​henning> Yeah, I think that one's ready if you want to review it. I suppose I should rebase and test since it's been a minute
15:27 fdobridge: <m​henning> I never really did proper benchmarks though - although it did let me run larger maps on Civ 6 before OOM and I used to have it installed globally
18:52 anholt: gfxstrand: do I need a specific meson for nak, or something? getting "unresolved imports `nak_bindings`"
19:42 gfxstrand: I didn't think so
19:43 gfxstrand: You might need recent meson but nothing special as far as I remember
20:10 anholt: maybe something dependency related, after another meson configure and build things seem to work. :shrug:
22:25 fdobridge: <l​ingm> anholt: did you update your rust toolchain recently? IIRC errors like that show up if a crate has been compiled with a different version of rustc. cargo would automatically rebuild that dependency, but meson doesn't do that and the bug report celebrated its first birthday.
22:31 anholt: lingm: nope, and it was my first build of the branch anyway.
22:46 fdobridge: <l​ingm> ah, wrong guess then. (dear irc bridge, kindly use my display name instead of my account name kthxbye)