00:23 <alyssa> Uh

01:35 <alyssa> hnngh

01:35 <alyssa> 1080p llvmpipe gnome on the m1 feels snappier than 2400x1600 panfrost gnome on my chromebook

01:35 <alyssa> so, so, so not happy about this.

01:36 <alyssa> (hello from m1 gnome :-p)

01:47 <HdkR> Hello!

01:55 <alyssa> Asahi Linux: still less buggy than mainline rk3399 😜

01:57 <HdkR> Looking forward to when I can use it :>

01:59 <alyssa> i'm using it right now :]

02:01 <HdkR> I need that wifi and GPU acceleration then an M1X :D

02:08 <alyssa> there are patches ffor wifi

02:09 <alyssa> not sure why GPU acceleration is taking so long though :-p

02:11 <HdkR> Confusing isn't it

02:11 <HdkR> Maybe one day we'll learn what the hold up is

02:11 <HdkR> Very secretive group :P

02:17 <alyssa> 👻

02:19 <icecream95> cc1plus has finished giving me OOMs, so now I can finally check all the ways increasing the max texture size will break things

02:28 <alyssa> O:)

02:30 <HdkR> What's the current set max texture size in Panfrost? 8kx8k?

02:31 <alyssa> 4kx4k, there's an MR to bump it to 8kx8k

02:38 <HdkR> ah

03:41 <icecream95> ugh, why does Rust have to be so hard?

03:41 <icecream95> I might have to suspend my plans to rewrite Panfrost in Rust and achieve world domination until I can get a PhD in Rust

04:41 <icecream95> dEQP-GLES31.functional.image_load_store.cube.atomic.comp_swap_r32ui_return_value just flaked a couple of times for me, but I can't reproduce it any more, even if I rerun the test 100 times

05:58 <tomeu> alyssa: yeah, I think it will be great to be able to track perf in some way, you can see in the logs how CI uses apitrace to do those measurements

06:29 <macc24> alyssa: why not ur not running cadmium :|

11:44 <robmur01> Hmm, apparently my results yesterday may have had more to do with the debugging stuff I put in the mmap path... Today glmark2 seems plenty happy running with 64K pages :/

11:45 <robmur01> (might as well do something interesting while I wait for a kernel build to debug another issue)

12:47 <alyssa> robmur01: 🤷

13:48 <alyssa> assigns #12312 to marge

13:48 <alyssa> fun number

13:49 <macc24> hm was there a #2137?

17:37 <alyssa> writer's block intensifies