00:50 anholt: coolcool. s390x incorrectly gives +0.0 for -1.0 * 0.0. love that for us.
00:54 airlied: gfxstrand, alyssa : know if anyone has thought about something like LLVM SCEV for NIR?
06:10 mareko: [A
08:42 digetx: alyssa: reviews and testing are welcomed for intel nctx
13:18 alyssa: airlied: could be cool
13:18 alyssa: tarceri: is probably the one to talk to
13:26 alyssa: digetx: ack, thanks, would love to see that landed :)
13:55 llyyr: is it better to pick bgr(a) surfaces of rgb(a) surfaces for vulkan apps? if the application could output either
13:55 llyyr: s/of/or
13:56 MrCooper: depends on other constraints, from the Vulkan API PoV it should make no difference
13:57 llyyr: i meant if it would avoid a swizzle in the drivers to pick a specific format
13:57 llyyr: metal/moltenvk only does direct-to-display rendering with bgr(a) surfaces, so I was wondering if it's fine for a cross platform renderer to pick bgra always to simplify surface selection code or if there's a reason to pick rgba over bgra on radv/anv etc.
13:58 glehmann: for amd it doesn't matter at all
13:58 llyyr: thanks!
15:09 llyyr: anv doesn't seem to support VK_IMAGE_USAGE_STORAGE_BIT for BGR, so i guess it does matter
15:15 dj-death: llyyr: it should on the last 3 generations
15:15 dj-death: llyyr: except the SRGB variant
15:16 llyyr: is UHD Graphics 770 among last 3 generations? wiki says gen12.2 which should be
15:17 dj-death: llyyr: nope
15:17 llyyr: ah alright
15:17 dj-death: llyyr: I'm counting pantherlake ;)
15:17 dj-death: maybe it's a bit early...
15:17 alyssa: is PTL out yet
15:18 glehmann: it was announced this week
15:18 dj-death: there are laptop prices etc...
15:20 alyssa: LNL crowd assemble
15:21 dj-death: anybody seen those errors when trying to replay trace files from traces-db?
15:21 dj-death: eglretrace: apitrace/lib/trace/trace_file_brotli.cpp:124: virtual size_t BrotliFile::rawRead(void*, size_t): Assertion `result == BROTLI_DECODER_RESULT_NEEDS_MORE_OUTPUT || result == BROTLI_DECODER_RESULT_SUCCESS' failed.
15:22 dj-death: oh... missing git-lfs...
15:41 austriancoder: alyssa: If you have time, it would be really nice if you could take another look at !38929
15:42 alyssa: oh right yes