02:05
mripard has joined #lima
03:38
kaspter has quit [Quit: kaspter]
03:42
kaspter has joined #lima
03:46
kaspter has quit [Ping timeout: 264 seconds]
03:48
kaspter has joined #lima
06:22
chewitt has quit [Quit: Adios!]
07:28
kaspter has quit [Ping timeout: 246 seconds]
07:28
<
linkmauve >
I live in CET btw, soon CEST. :)
07:38
<
linkmauve >
It works!
08:07
kaspter has joined #lima
08:36
kaspter has quit [Remote host closed the connection]
08:40
kaspter has joined #lima
12:57
kaspter has quit [Ping timeout: 260 seconds]
13:26
_whitelogger has joined #lima
13:59
zombah has quit [Ping timeout: 260 seconds]
14:19
zombah has joined #lima
15:35
egbert has quit [Ping timeout: 256 seconds]
15:40
egbert has joined #lima
16:06
<
anarsoul|2 >
linkmauve: PDT here
16:45
enunes has joined #lima
16:47
<
enunes >
anarsoul|2: ping
16:49
<
anarsoul|2 >
enunes: hey
16:49
<
anarsoul|2 >
did you look into what needs to be done to enable shareable shaders on lima?
16:49
<
enunes >
hmm probably not, what do you mean by shareable shaders?
16:50
<
anarsoul|2 >
PIPE_CAP_SHAREABLE_SHADERS
16:50
<
enunes >
I never looked into that
16:50
<
anarsoul|2 >
OK, I'll look into it
16:51
<
anarsoul|2 >
also I'll work on disk cache for shaders
16:53
<
enunes >
it was not easy, but the mali results are on the same platform, same kernel, same OS everything, only graphics stack changing
16:54
<
enunes >
'effect2d' can be fixed relatibely easily
16:55
<
enunes >
'loop3' seems to be because we dont take uniform loads out of loops, seems to be a nir thing thats why vc4 is also pretty bad there
16:55
<
enunes >
'shadow' is what has been puzzling me
16:56
<
anarsoul|2 >
enunes: IIRC blob uses Z16 in shadow while lima always uses Z24X8
16:56
<
anarsoul|2 >
but we don't know how to reload Z16 so it's disabled
16:56
<
enunes >
yeah this is what I have been considering
16:56
<
enunes >
I also thought it could be the varying size thing that we never implemented
16:56
<
anarsoul|2 >
yes, also this
16:57
<
enunes >
but overall better than what I expected
16:57
<
anarsoul|2 >
that also affects refract (and anything with complex geometry)
16:59
<
enunes >
is it too hard to figure out Z16, or just needed someone to spend time in that?
17:02
<
anarsoul|2 >
enunes: blob never attempts to reload Z16 and uses Z24X8 in any tests that requires depth reload
17:03
<
anarsoul|2 >
and sampling it as Z16 in reload shader doesn't work
17:09
<
anarsoul|2 >
so it's not a low-hanging fruit for sure
17:11
<
enunes >
hmm ok, I'll give it a try, if that doesnt work I'll try fp16 varying next to see if that helps
17:15
<
anarsoul|2 >
I wonder what's wrong with loop-3
17:15
<
enunes >
loop-3 is loading uniform inside a loop
17:16
<
enunes >
blob compiler puts the uniform load outside of the loop, nir doesnt
17:16
<
anarsoul|2 >
I guess we copy it into the loop ourselves :)
17:17
<
enunes >
I looked into that a bit but I think its not too much our fault
17:17
<
anarsoul|2 >
but I wonder why it misses the cache
17:18
<
enunes >
I did a local hack to that shader to force it to be loaded outside, and it already improved the results
17:18
<
anarsoul|2 >
what's wrong with effect2d?
17:21
<
enunes >
effect2d uses a RGB textures to which we apply a swizzle to make alpha '1.0' and that complicates the shader quite a bit
17:22
<
enunes >
it comes from mesa which defines the swizzle to alpha 1.0
17:23
<
anarsoul|2 >
wouldn't it be out of spec?
17:24
<
anarsoul|2 >
some shaders may depend on alpha being 1.0
17:24
<
enunes >
I agree, which is why I didnt submit it so far
17:24
<
enunes >
I dont know if all mesa drivers do this swizzle though, neither does the blob, I think they just treat alpha as dontcare
17:25
<
anarsoul|2 >
actually blob supports RGB format
17:25
<
anarsoul|2 >
one that is 3 bytes per pixel
17:26
<
anarsoul|2 >
we used to support it but since it's not renderable (and GL requires it to be renderable) we dropped it
17:28
<
enunes >
I didnt go as far as to check what format it uses, just checked the shader on that one, so its possible it was using that format too
17:37
<
anarsoul|2 >
btw, are you using deqp-runner?
17:39
<
anarsoul|2 >
it's very convenient, but looks like our skips and fails lists were outdated
17:40
<
enunes >
I have it set up but I also use the regular old deqp-gles2
17:40
<
anarsoul|2 >
it's just a runner, it still requires deqp-gles2
17:42
<
enunes >
makes sense, I also have a similar script
17:44
<
enunes >
results were outdated since we are not in CI anymore, I want to enable it again with the stack of test boards I got but then I need to finish my secure setup to expose them and power control to the boards
17:46
anarsoul|c has joined #lima
17:53
<
anarsoul|2 >
sounds good
17:53
<
anarsoul|2 >
anyway, I sent an MR to update the lists since I use them with deqp-runner
22:20
egbert has quit [Ping timeout: 264 seconds]
22:26
egbert has joined #lima
23:37
megi has quit [Quit: WeeChat 3.0.1]
23:37
megi has joined #lima
23:48
mmind00 has joined #lima