#asahi-gpu · 2022-10-14 22:49
TramtristThe issue is theyre coming from matrix?#asahi-gpu · 2022-10-14 22:49
alyssaAlso +c for good measure because lol colour#asahi-gpu · 2022-10-14 22:48
alyssaI guess we'll pilot that here and when marcan wakes up I'll talk to him about doing it for the other channels#asahi-gpu · 2022-10-14 22:48
alyssaThat should not affect any legitimate traffic to the development channels, IMHO.#asahi-gpu · 2022-10-14 22:47
alyssaI set the channel to +M -- "client may speak only when registered and identified to NickServ"#asahi-gpu · 2022-10-14 22:46
alyssaOh, I see.#asahi-gpu · 2022-10-14 22:46
alyssamodes#asahi-gpu · 2022-10-14 22:46
alyssaI'm also a bit confused about channel moes.#asahi-gpu · 2022-10-14 22:43
karolherbstyeah.. "used by" I think is it#asahi-gpu · 2022-10-14 22:43
karolherbstused by?#asahi-gpu · 2022-10-14 22:43
karolherbstehh.. from?#asahi-gpu · 2022-10-14 22:43
karolherbstI am sure that's some weird code for idiots#asahi-gpu · 2022-10-14 22:40
alyssaI'm not sure "dilate" means what they think it does.#asahi-gpu · 2022-10-14 22:40
karolherbst:(#asahi-gpu · 2022-10-14 22:36
Tramtristmessaged marcan .. we may need ops in more timezones for this if its going to continue :/#asahi-gpu · 2022-10-14 22:29
nigger[m]dilate trannies#asahi-gpu · 2022-10-14 02:14
phire*easy#asahi-gpu · 2022-10-14 02:14
phireshould be reasonably east to find once I get rid of this headache#asahi-gpu · 2022-10-14 02:14
phireI think there is just a way of passing arguments to the installer pkg to make it install a "full" installer, instead of just the online one that it does by default#asahi-gpu · 2022-10-14 02:11
karolherbstor binary nibs for that matter... the glorious times where you could mod apples apps by just messing with their interface builder files#asahi-gpu · 2022-10-14 02:10
karolherbstthey all have compiled plists by now, don't they? (used macos when there weren't binary plists)#asahi-gpu · 2022-10-14 02:09
phirelol, I'm in such a bad headspace that I'm failing at fighting macos installer pkgs#asahi-gpu · 2022-10-14 02:01
alyssasince I am not in a headspace to r/e uarch details#asahi-gpu · 2022-10-14 02:00
alyssaswitched to preamble shaders#asahi-gpu · 2022-10-14 00:48
alyssaWe do expect there's some ILP handled in HW, even for ALU#asahi-gpu · 2022-10-14 00:47
phirewith that variable instruction width, I'd hope not#asahi-gpu · 2022-10-14 00:47
alyssaUnclear#asahi-gpu · 2022-10-14 00:47
phiredo we know how fast the shader's instruction decoder is? Does it go beyond 1 instruction per cycle?#asahi-gpu · 2022-10-14 00:45
alyssaWithout the specs it's hard to know for sure, of course#asahi-gpu · 2022-10-14 00:44
alyssaSubtle uarch behaviour that would be a single footnote in a spec and the compiler just has to deal with, more likely ;-p#asahi-gpu · 2022-10-14 00:44
alyssaHW bug, maybe#asahi-gpu · 2022-10-14 00:44
phireI guess it's possible there is a hardware bug? the nop is a workaround?#asahi-gpu · 2022-10-14 00:37
alyssaand this workaround is only because of a missing optimization!#asahi-gpu · 2022-10-14 00:37
alyssait's all hw so far#asahi-gpu · 2022-10-14 00:37
alyssanope.#asahi-gpu · 2022-10-14 00:34
phirethere hasn't been any static scheduling anywhere else?#asahi-gpu · 2022-10-14 00:33
alyssathis is.. really iffy.#asahi-gpu · 2022-10-14 00:27
alyssaunclear if that's actually flushing anything (acting as a barrier), or it's just masking the timing-dependent bug#asahi-gpu · 2022-10-14 00:27
alyssaif I insert an intervening nop, the problem isn't visible#asahi-gpu · 2022-10-14 00:25
alyssaso IDK if they have a workaround#asahi-gpu · 2022-10-14 00:20
alyssaor they can just statically guarantee it doesn't happen#asahi-gpu · 2022-10-14 00:20
alyssaso IDK if they have a workaround#asahi-gpu · 2022-10-14 00:20
alyssaI don't know if it's possible to emit that sequence with the blob#asahi-gpu · 2022-10-13 23:44
alyssaon Mali that's a special instruction#asahi-gpu · 2022-10-13 23:44
alyssai.e. a fast path for "texture2D(tex, v_texcoords.xy)"#asahi-gpu · 2022-10-13 23:44
alyssaDebugging *did* lead me to a bit that controls forwarding of varyings to sample coordinates#asahi-gpu · 2022-10-13 23:43
alyssaAGX seems to be inconsistent?#asahi-gpu · 2022-10-13 23:43
alyssaEvery GPU uarch does something different here#asahi-gpu · 2022-10-13 23:43
alyssaTBD#asahi-gpu · 2022-10-13 23:42
cr1901I thought write-after-writes was a performance penalty (but doesn't cause incorrect program operation). Do GPUs forego those checks and expect you to just "don't do that"?