#asahi-gpu · 2024-02-16 03:11
alyssai gotchu#asahi-gpu · 2024-02-16 03:09
alyssavalgrind, maybe?#asahi-gpu · 2024-02-16 03:09
alyssasuspects shenanigans#asahi-gpu · 2024-02-16 03:09
alyssaHm.#asahi-gpu · 2024-02-16 03:06
alyssathe weird part to me is just that iris is fine#asahi-gpu · 2024-02-16 03:05
alyssathx#asahi-gpu · 2024-02-16 03:04
alyssachergert: right before the assert can yo printf the pt->last_level and the lastLevel?#asahi-gpu · 2024-02-16 03:04
alyssaunless it's transfer helper related maybe#asahi-gpu · 2024-02-16 03:03
alyssaI have no clue how that common code would fail on asaahi but pass on iris#asahi-gpu · 2024-02-16 01:41
alyssachergert: no dice, the replay still doesn't repro#asahi-gpu · 2024-02-16 01:40
alyssathx#asahi-gpu · 2024-02-16 00:42
alyssacross-driver apitraces are a little bit spicy but should be ok here#asahi-gpu · 2024-02-16 00:42
alyssacan you try getting the apitrace from the intel (iris) ?#asahi-gpu · 2024-02-16 00:42
alyssaoh, because it crashed I assume#asahi-gpu · 2024-02-16 00:41
alyssaapitrace replay ~/Downloads/epiphany.trace#asahi-gpu · 2024-02-16 00:41
alyssachergert: trace doesn't replay properly#asahi-gpu · 2024-02-16 00:40
alyssathx#asahi-gpu · 2024-02-16 00:15
alyssafor gles#asahi-gpu · 2024-02-16 00:15
alyssaapitrace trace -a egl#asahi-gpu · 2024-02-16 00:07
alyssathanks#asahi-gpu · 2024-02-16 00:07
alyssaplease send an apitrace#asahi-gpu · 2024-02-16 00:02
alyssa(intel, amd?)#asahi-gpu · 2024-02-16 00:02
alyssachergert: does it happen with non-asahi mesa drivers?#asahi-gpu · 2024-02-16 00:01
alyssathat sounds vaguely familiar but I don't remember why#asahi-gpu · 2024-02-15 23:04
alyssatry stuff? or throw me a branch#asahi-gpu · 2024-02-15 23:03
alyssalooks plausible, idk#asahi-gpu · 2024-02-15 23:03
alyssawhich is plausible#asahi-gpu · 2024-02-15 23:03
alyssa 584: 0501200d08c03200 device_load 0, i32, xy, r0_r1, u0_u1, 130, signed#asahi-gpu · 2024-02-15 23:03
alyssafrom the preamble, u18 is equal to#asahi-gpu · 2024-02-15 23:02
alyssa 0: 12014a421a4a403a icmpsel ult, r0, r5, u18, r5, u18#asahi-gpu · 2024-02-15 23:02
alyssafirst vertex shader instruction is the clamp:#asahi-gpu · 2024-02-15 22:42
alyssaworth looking into#asahi-gpu · 2024-02-15 22:41
alyssado you really upload 455+ scissors?#asahi-gpu · 2024-02-15 22:41
alyssa Scissor: 454#asahi-gpu · 2024-02-15 22:41
alyssaDepth bias/scissor#asahi-gpu · 2024-02-15 22:40
alyssacherrypick "HACK: Wire dougallj's disassembler into wrap+compiler" from my next branch and setup applegpu#asahi-gpu · 2024-02-15 22:39
alyssaoh uh can you build with disasembler#asahi-gpu · 2024-02-15 22:39
alyssawell I'm first going to look at the VS because we're talking about it#asahi-gpu · 2024-02-15 22:16
alyssa(or set the clamp to infinity or whatever)#asahi-gpu · 2024-02-15 22:16
alyssabut agxv doesn't advertise robustness yet so you can just hack that out#asahi-gpu · 2024-02-15 22:15
alyssai509vcb: ultimately vkCmdBindVertexBuffers2, but it should be in the vk_graphics_state#asahi-gpu · 2024-02-15 15:33
alyssaok, will do in a bit#asahi-gpu · 2024-02-15 15:32
alyssai509vcb: lmk if you want me to push that#asahi-gpu · 2024-02-15 15:32
alyssa:)#asahi-gpu · 2024-02-15 14:08
alyssawoo#asahi-gpu · 2024-02-15 12:24
alyssai509vcb: VIEWPORT now also has a VIEWPORT_CONTROL word, see agx_state.c (and blame if you want to know who to blame)#asahi-gpu · 2024-02-15 02:19
alyssai509vcb: new_ppp_update takes a # of control words, needs to be high enough#asahi-gpu · 2024-02-15 01:31
alyssagiven the XML is changing externally, etc#asahi-gpu · 2024-02-15 01:31
alyssaI don't think preserving history is reasonable at this stage, tbh.#asahi-gpu · 2024-02-15 01:30
alyssaI am yeah