2026-08-25
00:04aurora has joined #asahi-gpu
01:08alessaba has quit [Read error: Connection reset by peer]
01:09alessaba has joined #asahi-gpu
01:28chadmed has quit [Read error: Connection reset by peer]
01:29chadmed has joined #asahi-gpu
01:29chadmed has quit [Read error: Connection reset by peer]
01:30chadmed has joined #asahi-gpu
01:54al3xtjames2 has quit [Quit: al3xtjames2]
02:00yurix has quit [Remote host closed the connection]
02:00yurix has joined #asahi-gpu
02:26maettu102 has quit [Remote host closed the connection]
02:26lederstr- has joined #asahi-gpu
02:26maettu102 has joined #asahi-gpu
02:29lederstrumpf has quit [Ping timeout: 480 seconds]
03:34waltmckchaos_princess: increasing the occlusion query limit to 1M increases the bitmap from 4KiB to 128KiB. So not a very significant change
03:36waltmckI don't really understand this stuff very well, but I think that the issue is that on honeykrisp the buffer object table can't grow
03:36waltmcksee https://gitlab.freedesktop.org/mesa/mesa/-/blob/mesa-26.1.5/src/asahi/vulkan/hk_descriptor_table.c#L35
03:36waltmckso you need to allocate the entire table up front regardless of how much is actually used
03:54JamesCalligeros[m]Did you run the CTS?
03:54JamesCalligeros[m]We cannot accept gpu patches that regress the vulkan or opengl cts
04:13JamesCalligeros[m]yeah that one
04:14JamesCalligeros[m]building and running it is a little uhhhh cumbersome
04:40TheLink has joined #asahi-gpu
04:43olivial has quit [Read error: Connection reset by peer]
04:43olivial has joined #asahi-gpu
04:48ddxtanx has quit [Remote host closed the connection]
04:49ddxtanx has joined #asahi-gpu
05:41JamesCalligeros[m]Theres also a list of tests we expect to fail somewhere in .gitlab-ci
06:18Lyude has quit [Quit: Bouncer restarting]
06:19Lyude has joined #asahi-gpu
06:20al3xtjames2 has joined #asahi-gpu
06:25lyudess has joined #asahi-gpu
06:28Lyude has quit [Ping timeout: 480 seconds]
06:39ddxtanx_ has joined #asahi-gpu
06:40ddxtanx has quit [Read error: Connection reset by peer]
06:57ddxtanx_ has quit [Remote host closed the connection]
06:59ddxtanx has joined #asahi-gpu
07:02HerbY_NL has joined #asahi-gpu
07:09chadmed has quit [Quit: Konversation terminated!]
07:12maettu1027 has joined #asahi-gpu
07:13maettu102 has quit [Read error: Connection reset by peer]
07:13maettu1027 is now known as maettu102
07:14lederstrumpf has joined #asahi-gpu
07:16lederstr- has quit [Ping timeout: 480 seconds]
07:24foltz51 has joined #asahi-gpu
07:30foltz5 has quit [Ping timeout: 480 seconds]
07:30foltz51 is now known as foltz5
07:30pb17 has quit [Ping timeout: 480 seconds]
07:39pb17 has joined #asahi-gpu
08:01chadmed has joined #asahi-gpu
08:08olivial has quit [Ping timeout: 480 seconds]
08:15olivial has joined #asahi-gpu
08:17HerbY_NL has quit [Ping timeout: 480 seconds]
08:32jannau_ is now known as jannau
09:32HerbY_NL has joined #asahi-gpu
09:41al3xtjames2 has quit [Read error: Connection reset by peer]
09:41al3xtjames2 has joined #asahi-gpu
10:23picnoir has quit [Ping timeout: 480 seconds]
10:31picnoir has joined #asahi-gpu
10:31HerbY_NL has quit [Quit: Leaving]
11:27chadmedi dont think im setting up the gpu tiling layout properly, no matter what i do iomfb complains about the plane size being too small for whatever sanity check it's calculated itself
11:28chadmedah well interchange will suffice for now i suppose, not like macos ever does this anyway
11:31chadmedthat was my thinking, and i thought "no that sounds stupid when have apple ever done anything like an htpc"
11:32chadmedmy only thinking is that its a remnant of some sort of apple tv specific format that predates both AGX and Interchange
11:32chaos_princessthey did briefly do a mac mini as htpc thing, with an ir remote & stuff
11:32chaos_princessbut yea, looks more atv-related
12:18chadmed has quit [Quit: Konversation terminated!]
12:22chadmed has joined #asahi-gpu
13:21alyssa has joined #asahi-gpu
13:58chadmed has quit [Quit: Konversation terminated!]
13:58chadmed has joined #asahi-gpu
14:21user890104_ has joined #asahi-gpu
14:28user890104 has quit [Ping timeout: 480 seconds]
14:32user890104_ is now known as user890104
14:32user890104 has quit []
14:32user890104 has joined #asahi-gpu
14:39lion328 has quit [Ping timeout: 480 seconds]
14:52lion328 has joined #asahi-gpu
15:00lion328 has quit [Ping timeout: 480 seconds]
15:06lion328 has joined #asahi-gpu
15:20alyssaso hk limits to 32k occlusion queries VkDevice-wide, because doing otherwise would require major shenanigans
15:33alyssaThe relevant limit is "Maximum visibility query offset", which on all M1-M5 chips is 256 KB = 32k queries at 64-bits each
16:56HerbY_NL has joined #asahi-gpu
17:14HerbY_NL has quit [Remote host closed the connection]
17:26HerbY_NL has joined #asahi-gpu
17:27HerbY_NL has quit [Remote host closed the connection]
18:01amarioguy has quit [Ping timeout: 480 seconds]
18:01amarioguy has joined #asahi-gpu
19:53okb has joined #asahi-gpu
19:58waltmckalyssa: very interesting, thank you. What is expected to happen if I just raise the limit anyway? Everything seems to work fine, but I haven't running the vulkan conformance tests
19:59waltmckvisually, no mans sky looks completely fine with the limit raised and looks completely broken with it left as is
20:00waltmckbut yeah it sounds like the driver is doing the "correct" behavior and NMS is just not handling it well
20:01waltmckI don't think I've ever seen it use more than 32K queries at once, it just starts a query pool that is massive
20:02waltmckprobably I'll just keep running my local patch with the limit raised since this shouldn't be upstreamed
21:05JamesCalligeros[m]Also the vk cts should not still be running lol, it should only take half an hour or so
21:46waltmckJamesCalligeros[m]: yeah I had to do something else so I stopped it. will run it later when I have time
21:49rakyah has joined #asahi-gpu
21:51lyudess has quit []
21:51Lyude has joined #asahi-gpu
21:59rakyah has quit [Quit: rakyah]
22:43al3xtjames2 has quit [Quit: al3xtjames2]