#asahi-gpu · 2022-03-22 00:23
dougallalyssa: Yeah, from where the `tt` bits are in the encoding it'd have to be a completely different instruction, so it's probably not a thing. (I did go looking for a ternary logic op like `vpternlog`, but couldn't find that either)#asahi-gpu · 2022-02-22 01:34
dougall(but yeah, i don't think that's saying what phire was saying)#asahi-gpu · 2022-02-22 01:33
dougall(is this the video? https://developer.apple.com/videos/play/wwdc2020/10632/ - confusingly, they keep making videos with similar titles each year - but you can search the transcript for "depth" and click the links to jump to that part of the video which is handy)#asahi-gpu · 2022-02-22 00:39
dougall(i should really rewatch all the wwdc videos, and the llvm dev meeting video - lots of good stuff that might make more sense now we have more context)#asahi-gpu · 2022-02-11 03:22
dougallyeah, possibly... they seem quite different, but then it's not clear to me why the ring buffers for different types of work are separated to begin with#asahi-gpu · 2022-02-11 03:12
dougallheh, that is weird#asahi-gpu · 2021-05-25 03:33
dougallnice!#asahi-gpu · 2021-05-25 03:16
dougallhmm, yeah, i'm probably doing other weird things that are confusing me#asahi-gpu · 2021-05-25 03:14
dougall(i think i was doing offscreen rasterisation tests, and you don't get a writeout instruction there? it just stores to tile memory?)#asahi-gpu · 2021-05-25 03:12
dougallyeah - having given it some thought, i definitely have no idea :P#asahi-gpu · 2021-05-25 03:02
dougalloh, i'm getting mixed up i guess - i thought it was at the end of the vertex shaders?#asahi-gpu · 2021-05-25 02:58
dougall(not in any depth - i think roughly setting up the cf state for interpolation - but i think you named it and probably understand better)#asahi-gpu · 2021-05-25 02:54
dougall:D#asahi-gpu · 2021-05-25 02:54
dougallhahaha that's awesome! does that imply most/all control flow is working, or do the loop tests avoid mixing other structures?#asahi-gpu · 2021-05-25 02:04
dougall:o well done!#asahi-gpu · 2021-05-24 02:46
dougallooh, conditional execution time - exciting!#asahi-gpu · 2021-05-04 08:17
dougall(i'm guessing yes: https://gitlab.freedesktop.org/mesa/mesa/-/blob/cad54e27/src/asahi/lib/cmdbuf.xml#L50-60 )#asahi-gpu · 2021-04-30 23:38
dougallsure :)#asahi-gpu · 2021-04-30 23:36
dougallhahahaha#asahi-gpu · 2021-04-28 00:26
dougallgood catch - yeah, that makes a bit more sense :)#asahi-gpu · 2021-04-23 02:53
dougallawesome! looking forward to trying that out#asahi-gpu · 2021-04-17 04:56
dougallyeah, i'm sure it'll make more sense as we figure more out, but definitely plenty of puzzles left :)#asahi-gpu · 2021-04-17 04:54
dougall(that's from the announcement video iirc)#asahi-gpu · 2021-04-17 04:53
dougallhttps://www.anandtech.com/show/16252/mac-mini-apple-m1-tested/3 <- the slide#asahi-gpu · 2021-04-17 04:53
dougallyeah, i'm quite confused... "128 execution units", so 16 per core? when a simd-group is 32? is that like AMDs thing where you just put each half through the pipeline over two cycles?#asahi-gpu · 2021-04-17 04:50
dougallyeah... that number is right, but i'm a bit confused where the 24x comes from? like is the register file per-core? (i assume so, and that'd an 8x, but then where's the 3x?)#asahi-gpu · 2021-04-17 04:17
dougallnice!#asahi-gpu · 2021-04-17 04:05
dougallyeah, i think that all makes sense#asahi-gpu · 2021-04-17 04:05
dougallhaha#asahi-gpu · 2021-04-17 03:50
dougall(is one register quadword like r0-r3 or like r0l-r1h?)#asahi-gpu · 2021-04-17 03:47
dougallah, what's it look like?#asahi-gpu · 2021-04-17 03:18
dougallah, yeah, that'd make sense :)#asahi-gpu · 2021-04-17 03:16
dougall(asside from precision)#asahi-gpu · 2021-04-17 03:16
dougall"Implemented as x * rsqrt(x) with special cases handled correctly" - so i guess how does sqrt(x) differ from x * rsqrt(x)?#asahi-gpu · 2021-04-17 02:07
dougalli think a granularity of four is possibly correct... register granularity is observable by using out-of-range register ids. occupancy would be 'floor(total / round_up(register_count))' right?#asahi-gpu · 2021-04-17 02:04
dougallhmm... yeah, i've been wanting to get at those counters, but i'm not quite sure how to approach it#asahi-gpu · 2021-04-16 03:58
dougallyeah, that sounds much better :) g'night!#asahi-gpu · 2021-04-16 00:18
dougallbloom: great solve! that all properly makes sense now :)#asahi-gpu · 2021-04-15 02:52
dougall(or maybe i was trying to say that immediates don't get sign extended? not really the best way to represent that... hmm)#asahi-gpu · 2021-04-15 02:21
dougallyeah, not sure what's up with that encoding... i do think there's _something_ in the high couple of bits in most/all instructions that i haven't figured out, which might make it make a tiny bit more sense#asahi-gpu · 2021-04-15 02:17
dougallfwiw i saw apple's compiler emit 64-bit subtracts and 64-bit add+shift, but (as far as i can recall) not 64-bit adds#asahi-gpu · 2021-04-15 02:14
dougallhmm, yeah, i think you're right about < 64...#asahi-gpu · 2021-04-15 00:27
dougallyeah, C is particularly painful for that... i'd probably use __int128, and i'd probably end up regretting it :p#asahi-gpu · 2021-04-15 00:13
dougallah, good point, yeah... that makes sense#asahi-gpu · 2021-04-15 00:06
dougallyeah, they could definitely be worse - i'm just thinking of moving the Am/Bm/Cm field relative to their 32-bit equivalents (why did they do that?), and the fact that 32-bit ops can have 16-bit sources and destinations too#asahi-gpu · 2021-04-14 23:58
dougallhaha least regular ALU#asahi-gpu · 2021-04-14 23:52
dougallhaha yep, fadd16/fmul16/fmadd16 and 64-bit iadd/imadd come to mind, although i don't remember if that's the worst of it#asahi-gpu · 2021-04-14 23:41
dougallsounds good! yeah, it's mostly really nice and regular :)#asahi-gpu · 2021-04-12 03:15
dougallhaha awesome!#asahi-gpu · 2021-04-12 02:41
dougall:o