#
✕ nick:alyssa
#asahi-gpu · 2023-01-08 02:00
alyssaMy buffer texture code is causing faults but dmesg isn't much help figuring out why
#asahi-gpu · 2023-01-08 01:59
alyssalina: Nod
#asahi-gpu · 2023-01-08 01:51
alyssaI'm also suspicious about the unit names
#asahi-gpu · 2023-01-08 01:50
alyssaas in, I see it set sometimes when it's a read instead
#asahi-gpu · 2023-01-08 01:50
alyssalina: BTW, I think the "write" field in FaultInfo is wrong
#asahi-gpu · 2023-01-08 00:51
alyssawoo!
#asahi-gpu · 2023-01-08 00:41
alyssaye?
#asahi-gpu · 2023-01-08 00:20
alyssa"noticeable" mood
#asahi-gpu · 2023-01-08 00:20
alyssa*sparkles*
#asahi-gpu · 2023-01-07 22:26
alyssaYou know the drill
#asahi-gpu · 2023-01-07 22:26
alyssaPushed the latest asahi/main
#asahi-gpu · 2023-01-07 22:26
alyssa----
#asahi-gpu · 2023-01-07 21:23
alyssahm
#asahi-gpu · 2023-01-07 21:23
alyssaalthough... flipping sampler descriptor bits is definitely causing tests to regress related to LOD so
#asahi-gpu · 2023-01-07 21:22
alyssathe perf impact is probably not *terrible* but it would still suck
#asahi-gpu · 2023-01-07 21:22
alyssaDelight
#asahi-gpu · 2023-01-07 21:19
alyssanow I know why :|
#asahi-gpu · 2023-01-07 21:19
alyssaTellowKrinkle: Apple's GL driver implements every single texture op as textureBias with the bias from a uniform register
#asahi-gpu · 2023-01-07 21:08
alyssamashing bits isn't finding anything yet..
#asahi-gpu · 2023-01-07 20:57
alyssaah yes I forgot that MoltenVK gets to cheat.
#asahi-gpu · 2023-01-07 20:57
alyssa"If the VK_KHR_portability_subset extension is enabled, and VkPhysicalDevicePortabilitySubsetFeaturesKHR::samplerMipLodBias is VK_FALSE, mipLodBias must be zero"
#asahi-gpu · 2023-01-07 20:56
alyssatime to mash random bits
#asahi-gpu · 2023-01-07 20:56
alyssadoesn't look like lod bias is implemented in Apple's driver but there's a chance it's still supported in (deprecated?) hw
#asahi-gpu · 2023-01-07 20:30
alyssalod_bias assert going off. Great.
#asahi-gpu · 2023-01-07 20:29
alyssaworks fine in a release build..
#asahi-gpu · 2023-01-07 20:16
alyssano deqp-gles2 regressions
#asahi-gpu · 2023-01-07 20:09
alyssaGNOME doesn't want to start on my latest rebase, gg
#asahi-gpu · 2023-01-06 01:49
alyssaand 1d
#asahi-gpu · 2023-01-06 01:49
alyssapile of texwrap fails for compression, yep that was expected given the state of this code
#asahi-gpu · 2023-01-06 01:48
alyssa199 transform feedback
#asahi-gpu · 2023-01-06 01:47
alyssaof the remaining 1331..
#asahi-gpu · 2023-01-06 01:47
alyssa1153 of those are clipflat
#asahi-gpu · 2023-01-06 01:46
alyssadown to 2484, yeet
#asahi-gpu · 2023-01-06 01:46
alyssaat 2716 fails+crash
#asahi-gpu · 2023-01-06 01:46
alyssaDec 22 was peak broken
#asahi-gpu · 2023-01-06 01:46
alyssafewer fails than Dec 22, there's that :p
#asahi-gpu · 2023-01-06 01:45
alyssa(this is agx/next, going to have a lot less fails/crashes on the release obviously)
#asahi-gpu · 2023-01-06 01:45
alyssaincluding known broken stuff
#asahi-gpu · 2023-01-06 01:44
alyssai guess that's the problem with enabling more exts
#asahi-gpu · 2023-01-06 01:44
alyssai feel like these piglit results get worse every time i run them
#asahi-gpu · 2023-01-06 01:36
alyssabut certainly not the relevant piglits (-:
#asahi-gpu · 2023-01-06 01:36
alyssaok. have a branch passing the relevant deqp's
#asahi-gpu · 2023-01-06 01:19
alyssai hate border colours so much
#asahi-gpu · 2023-01-06 01:02
alyssaoh, yeah. it is. What a delight.
#asahi-gpu · 2023-01-06 00:58
alyssaunless it's 11-bit unorm internally ... nn
#asahi-gpu · 2023-01-06 00:57
alyssanot sure what R11 maps to yet
#asahi-gpu · 2023-01-06 00:56
alyssasRGB " == sRGB "
#asahi-gpu · 2023-01-06 00:56
alyssaETC2 RGB8A1 == RGBA8
#asahi-gpu · 2023-01-06 00:56
alyssaETC2 RGBA8 == RGBA8
#asahi-gpu · 2023-01-06 00:56
alyssaETC2 RGB8 == RGB8

plus ›