#
✕ nick:alyssa
#asahi-gpu · 2023-06-29 16:48
alyssaI highly recommend https://gitlab.freedesktop.org/anholt/deqp-runner
#asahi-gpu · 2023-06-29 16:47
alyssawhy piglit btw?
#asahi-gpu · 2023-06-29 15:23
alyssa_jannau__: curious
#asahi-gpu · 2023-06-29 15:01
alyssahang how?
#asahi-gpu · 2023-06-29 14:49
alyssawill fix that
#asahi-gpu · 2023-06-29 14:49
alyssaalright.
#asahi-gpu · 2023-06-29 14:49
alyssajannau: dammit
#asahi-gpu · 2023-06-29 14:46
alyssaok, that does what I want
#asahi-gpu · 2023-06-29 14:28
alyssalina: ^
#asahi-gpu · 2023-06-29 14:28
alyssaand see if the shaders diff (other than adding an atomic, obviously) and/or if the structures differ
#asahi-gpu · 2023-06-29 14:28
alyssathe way to go about checking this is to diff the agxdecode on t600x for a compute kernel without atomic + compute kernel with
#asahi-gpu · 2023-06-29 14:28
alyssa3. pass a flag to the kernel telling the kernel to pass a flag to the hardware to force things together?
#asahi-gpu · 2023-06-29 14:28
alyssa2. set a flag in the CDM dispatch structs forcing things to stay together?
#asahi-gpu · 2023-06-29 14:27
alyssa1. add more barriers?
#asahi-gpu · 2023-06-29 14:27
alyssaand so the compiler needs to report whether atomics are used, and if so either
#asahi-gpu · 2023-06-29 14:27
alyssabecause the atomics aren't coherent across clusters
#asahi-gpu · 2023-06-29 14:27
alyssaor need to have barriers on multi-cluster parts that aren't needed on single-cluster parts
#asahi-gpu · 2023-06-29 14:27
alyssaI wonder if compute kernels with atomics need to be pinned to a single cluster for correct results
#asahi-gpu · 2023-06-29 14:26
alyssaOr, geez, I wonder
#asahi-gpu · 2023-06-29 14:26
alyssathinking either the kernel is configuring the CDM badly on t600x, or Mesa is doing something silly, or..
#asahi-gpu · 2023-06-29 14:23
alyssaflush_finish.* are the stupid tests that we skip in CI
#asahi-gpu · 2023-06-29 14:22
alyssaYes.
#asahi-gpu · 2023-06-29 14:22
alyssathis really could be either Mesa or kernel, tbh
#asahi-gpu · 2023-06-29 14:10
alyssaAll of these tests use atomics.
#asahi-gpu · 2023-06-29 14:09
alyssathere are also single invocation/group versions of all those, which I assume are passing
#asahi-gpu · 2023-06-29 14:09
alyssaThey're all compute shader tests, but tests with multiple invocations/groups
#asahi-gpu · 2023-06-29 14:09
alyssajannau: Those tests at the end are interesting
#asahi-gpu · 2023-06-29 13:57
alyssaspicy.
#asahi-gpu · 2023-06-29 00:57
alyssaand this is despite low, ~constant memory usage according to top.
#asahi-gpu · 2023-06-29 00:57
alyssa(Well before it gets to the spicy Big Framebuffer test that causes the run to fail.)
#asahi-gpu · 2023-06-29 00:57
alyssaIf I `./glcts -n dEQP-GLES31.functional.*`, I get "collect_garbage: failed to reserve space" in under 5 minutes
#asahi-gpu · 2023-06-29 00:56
alyssalina: This is interesting
#asahi-gpu · 2023-06-29 00:52
alyssalina: https://rosenzweig.io/hrm.txt
#asahi-gpu · 2023-06-29 00:51
alyssaevidently did not fix the leaks hard enough
#asahi-gpu · 2023-06-29 00:51
alyssaOh come on. Failed in the exact same place.
#asahi-gpu · 2023-06-29 00:33
alyssalina: that leaves you with M1 Pro, if we even want a separate run for that?
#asahi-gpu · 2023-06-29 00:32
alyssajannau: said you can do M1 Max/Ultra/M2
#asahi-gpu · 2023-06-29 00:32
alyssaI'll supply classic M1
#asahi-gpu · 2023-06-29 00:32
alyssalina: jannau: we should coordinate who's doing which hardware
#asahi-gpu · 2023-06-29 00:31
alyssakicking off another run with the leak fixes
#asahi-gpu · 2023-06-29 00:02
alyssagot the valgrind symbols
#asahi-gpu · 2023-06-29 00:02
alyssafedora is my new love
#asahi-gpu · 2023-06-28 23:43
alyssagrr why is valgrind not giving any symbols
#asahi-gpu · 2023-06-28 23:41
alyssatries to valgrind
#asahi-gpu · 2023-06-28 23:30
alyssawrites leaky code.
#asahi-gpu · 2023-06-28 23:16
alyssaand watching memory usage in top
#asahi-gpu · 2023-06-28 23:16
alyssafast growing leak can be seen running dEQP-GLES31.functional.ssbo.* in succession
#asahi-gpu · 2023-06-28 23:13
alyssaso I think I have a leak in Mesa and it just hit the kernel first
#asahi-gpu · 2023-06-28 23:13
alyssawhat I do observe is memory slowly creeping up when running deqp
#asahi-gpu · 2023-06-28 23:12
alyssaHmm, I don't see any evidence of a kernel bug specifically.

plus ›