#asahi-gpu · 2022-12-20 02:55
alyssaso apply just... generates asm that looks like mine... but it doesn't work for me...#asahi-gpu · 2022-12-20 02:55
alyssanope no difference#asahi-gpu · 2022-12-20 02:55
alyssa 32: 3800 wait 0#asahi-gpu · 2022-12-20 02:55
alyssa 2a: 0559880e00c8f200 device_load 0, i32, xyzw, r11_r12_r13_r14, u4_u5, r4, unsigned, lsl 2#asahi-gpu · 2022-12-20 02:55
alyssa 22: 0531840e00c8f200 device_load 0, i32, xyzw, r6_r7_r8_r9, u2_u3, r4, unsigned, lsl 2#asahi-gpu · 2022-12-20 02:55
alyssa 1a: 0501800e00c8f200 device_load 0, i32, xyzw, r0_r1_r2_r3, u0_u1, r4, unsigned, lsl 2#asahi-gpu · 2022-12-20 02:55
alyssa 14: 3e9108e02500 convert f_to_u32, r4, r15, rtz#asahi-gpu · 2022-12-20 02:52
alyssasame index, different base, maybe that's part#asahi-gpu · 2022-12-20 02:50
alyssai get it, it's only okay when apple's compiler does it..#asahi-gpu · 2022-12-20 02:48
alyssaso... what gives?!#asahi-gpu · 2022-12-20 02:45
alyssawhy isn't this work for my compiler#asahi-gpu · 2022-12-20 02:44
alyssaso if I can get the non-TODO one in a VS#asahi-gpu · 2022-12-20 02:44
alyssawtfff#asahi-gpu · 2022-12-20 02:44
alyssaindeed I don't get it in a vertex shader#asahi-gpu · 2022-12-20 02:42
alyssathis is in a fragment shader. I wonder if that makes a difference.#asahi-gpu · 2022-12-20 02:42
alyssaTellowKrinkle: and the plot thickens#asahi-gpu · 2022-12-20 02:35
TellowKrinkle(Compute shader that loads 3 float4s, does an fma, and stores the result)#asahi-gpu · 2022-12-20 02:35
TellowKrinkle 2c: 0551240e01c8f200 device_load 0, i32, quad, r10_r11_r12_r13, u2_u3, r9, unsigned, lsl 2#asahi-gpu · 2022-12-20 02:35
TellowKrinkle 1c: 0501040e00c8f200 device_load 0, i32, quad, r0_r1_r2_r3, u2_u3, r0, unsigned, lsl 2#asahi-gpu · 2022-12-20 02:35
TellowKrinkle 34: 3800 wait 0#asahi-gpu · 2022-12-20 02:35
TellowKrinkle 24: 0529a40e00c8f200 device_load 0, i32, quad, r5_r6_r7_r8, u2_u3, r5, unsigned, lsl 2#asahi-gpu · 2022-12-20 02:35
TellowKrinkleWeird, I'm getting this on my M1 (no .TODO):#asahi-gpu · 2022-12-20 02:20
alyssawhat in the world could this single celebrity bit do#asahi-gpu · 2022-12-20 02:20
alyssaso, so many questions#asahi-gpu · 2022-12-20 02:20
alyssabut not scalars#asahi-gpu · 2022-12-20 02:20
alyssaand vec2#asahi-gpu · 2022-12-20 02:19
alyssadoes it for vec3#asahi-gpu · 2022-12-20 02:18
alyssamuch ado about a single bit.#asahi-gpu · 2022-12-20 02:17
alyssaain't this what we saw this morning?#asahi-gpu · 2022-12-20 02:17
alyssa 36: 3800 wait 0#asahi-gpu · 2022-12-20 02:17
alyssa 2e: 0551800e00c8f200 device_load 0, i32, xyzw, r10_r11_r12_r13, u0_u1, r4, unsigned, lsl 2#asahi-gpu · 2022-12-20 02:17
alyssa 26: 0529a00e00c8f200 device_load 0, i32, xyzw, r5_r6_r7_r8, u0_u1, r5, unsigned, lsl 2#asahi-gpu · 2022-12-20 02:17
alyssa 1e: 0501000a00c8f200 device_load.TODO 0, i32, xyzw, r0_r1_r2_r3, u0_u1, r0, unsigned, lsl 2, 0, 0, 4, 0#asahi-gpu · 2022-12-20 02:17
alyssanow when I make it load vec4 it instead changes to#asahi-gpu · 2022-12-20 02:17
alyssaoh, hold on#asahi-gpu · 2022-12-20 02:15
alyssatheirs#asahi-gpu · 2022-12-20 02:15
alyssawhy does my wait not work but there's does? :P#asahi-gpu · 2022-12-20 02:15
alyssawhich is strange because Apple's compiler is happy to generate sequences like that#asahi-gpu · 2022-12-20 02:10
alyssa2 loads seems to work, but 3 no?#asahi-gpu · 2022-12-20 02:10
alyssaI just want multiple loads in flight :d#asahi-gpu · 2022-12-20 02:07
alyssaI wonder why this code sequence doesn't want to work :V#asahi-gpu · 2022-12-20 02:07
alyssa 20: 3800 wait 0#asahi-gpu · 2022-12-20 02:07
alyssa 18: 0541080e00c0f200 device_load 0, i32, xyzw, r8_r9_r10_r11, u4_u5, r0, unsigned#asahi-gpu · 2022-12-20 02:07
alyssa 10: 0529040e00c0f200 device_load 0, i32, xyzw, r5_r6_r7_r8, u2_u3, r0, unsigned#asahi-gpu · 2022-12-20 02:07
alyssa 8: 0509000e00c0f200 device_load 0, i32, xyzw, r1_r2_r3_r4, u0_u1, r0, unsigned#asahi-gpu · 2022-12-19 17:11
alyssaWoof.#asahi-gpu · 2022-12-19 17:09
alyssawhich means our current txf implementation is broken. gg.#asahi-gpu · 2022-12-19 17:09
alyssalikewise for texture_load, the level is passed in a 16-bit int uniform.#asahi-gpu · 2022-12-19 17:07
alyssaok, that's one mystery down#asahi-gpu · 2022-12-19 17:02
alyssaok, easy enough to handle