deflated8837_ has quit [Ping timeout: 480 seconds]
possiblemeatball has quit [Quit: Quit]
possiblemeatball has joined #asahi-gpu
possiblemeatball has quit []
possiblemeatball has joined #asahi-gpu
deflated8837 has joined #asahi-gpu
possiblemeatball has quit [Quit: Quit]
possiblemeatball has joined #asahi-gpu
possiblemeatball has quit [Quit: Quit]
possiblemeatball has joined #asahi-gpu
possiblemeatball has quit [Quit: Quit]
possiblemeatball has joined #asahi-gpu
possiblemeatball has quit []
possiblemeatball has joined #asahi-gpu
possiblemeatball has quit []
possiblemeatball has joined #asahi-gpu
phire_ has joined #asahi-gpu
phire is now known as Guest5202
phire_ is now known as phire
hightower3 has joined #asahi-gpu
Guest5202 has quit [Ping timeout: 480 seconds]
phire has quit [Quit: No Ping reply in 180 seconds.]
hightower2 has quit [Ping timeout: 480 seconds]
phire has joined #asahi-gpu
phire has quit [Read error: Connection reset by peer]
phire has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
crabbedhaloablut has quit []
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
cylm_ has joined #asahi-gpu
phire has quit [Read error: Connection reset by peer]
phire has joined #asahi-gpu
cylm has joined #asahi-gpu
cylm_ has quit [Ping timeout: 480 seconds]
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
PyroPeter has joined #asahi-gpu
pyropeter3 has quit [Ping timeout: 480 seconds]
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
r3muxd has joined #asahi-gpu
<r3muxd>
does anyone know if the working set limitations on the GPU are implemented in hardware or software? on asahi it says i can use all 16gb of unified memory as vram, but macos only lets me use about 10gb
<lina>
r3muxd: It's a software thing, on Asahi you can use as much as you want until things start crashing
<r3muxd>
opencl in asahi isnt functional yet right? just making sure because the only thing that shows up for opencl asahi is your streams
zkrx_ has joined #asahi-gpu
zkrx is now known as Guest5227
zkrx_ is now known as zkrx
Guest5227 has quit [Ping timeout: 480 seconds]
Cyrinux9 has quit [Ping timeout: 480 seconds]
Cyrinux9 has joined #asahi-gpu
r3muxd has quit [Read error: Connection reset by peer]
cylm has quit [Ping timeout: 480 seconds]
bisko has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
nsklaus has joined #asahi-gpu
i509vcb has quit [Quit: Connection closed for inactivity]
possiblemeatball has quit [Quit: Quit]
possiblemeatball has joined #asahi-gpu
possiblemeatball has quit [Quit: Quit]
nela2 has joined #asahi-gpu
possiblemeatball has joined #asahi-gpu
nela has quit [Ping timeout: 480 seconds]
nela2 is now known as nela
possiblemeatball has quit []
possiblemeatball has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
ChaosPrincess has quit [Quit: WeeChat 3.8]
ChaosPrincess has joined #asahi-gpu
possiblemeatball has quit [Quit: Quit]
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
Cyrinux94 has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
Cyrinux9 has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
akemin_dayo has quit [Ping timeout: 480 seconds]
akemin_dayo has joined #asahi-gpu
<_jannau__>
lina: please test KHR-GLES31.core.draw_indirect.advanced-twoPass-transformFeedback-arrays and KHR-GLES31.core.draw_indirect.advanced-twoPass-transformFeedback-elements for the barrier bits
<_jannau__>
before you start the full cts run
cylm has joined #asahi-gpu
mikelee has quit [Remote host closed the connection]
mikelee has joined #asahi-gpu
chadmed has quit [Quit: Konversation terminated!]
chadmed has joined #asahi-gpu
possiblemeatball has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
<jannau>
lina: for i in $(seq 1000); do if ./glcts -n dEQP-EGL.functional.render.multi_context.gles2.rgb888_window | grep -q 'Failed: *1/1'; then echo "$i repeat(s) until fail"; break; fi ; done
<jannau>
lina: the pbuffer variant of that test fails as well
<jannau>
but I'm using x11_egl as target
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
<_jannau__>
lina: no, it usually fails after 5 to 50 runs for me
<_jannau__>
running on the 10-Core GPU variant although I'd hope that that makes no difference
<lina>
250 repeat(s) until fail
<lina>
Now running 13.5 with the latest kernel, same mesa, no hypervisor
<lina>
204 repeat(s) until fail
<lina>
;;
<lina>
Guess it's just more heisen for me...
<_jannau__>
sometimes it results in render timeouts here which might or might be a separate issue
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
possiblemeatball has quit [Quit: Quit]
possiblemeatball has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
i509vcb has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
maximbaz has quit [Quit: bye]
maximbaz has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
darkapex has quit [Remote host closed the connection]
darkapex has joined #asahi-gpu
mikelee has joined #asahi-gpu
possiblemeatball has quit [Quit: Quit]
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
bisko has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
mikelee has joined #asahi-gpu
yamii has joined #asahi-gpu
yamii has quit []
roban has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
roban has quit []
mikelee has joined #asahi-gpu
yamii has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
mikelee has joined #asahi-gpu
mikelee has quit [Ping timeout: 480 seconds]
<alyssa>
Today's AGXV update:
<alyssa>
- more banging on descriptor sets. ended up ~~plagiarising~~ adapting NVK's implementation. hopefully that means we picked up a bunch of new features for free.
<alyssa>
- vk_meta implementations of vkCmdFillBuffer, vkCmdUpdateBuffer, and vkCmdCopyBuffer2.
<alyssa>
that's passing all the relevant deqp-vk tests
<alyssa>
and gets rid of all the internal hardcoded dispatches, which is a big deal for portability
<alyssa>
(Lina just found out that G14X needs a tweaked CDM header. With the new approach, there's only one place to make that change in the VK driver instead of 3 or more.)
<alyssa>
In addition to being totally generic (i.e. it'll Just Work for panvk2), I implemented one bit of cleverness that the AGX-specific one lacked -- dynamically choosing a block size based on the alignments
<alyssa>
in the happy path, it'll copy 16 bytes at a time
<alyssa>
(this still isn't fully optimal, since the thread dispatch overhead will dominate. but it's better than copying 1 byte at a time!)
<alyssa>
(actually let me go fix that, should be easy)
<alyssa>
--actually I don't care it's 7pm on a friday
mikelee has joined #asahi-gpu
<alyssa>
next week I'll look at the image copies, now that I've built up a little bit of self-confidence with hacking on vk_meta