00:34
kaspter has quit [Ping timeout: 276 seconds]
00:35
kaspter has joined #lima
00:46
kaspter has quit [Ping timeout: 245 seconds]
01:13
kaspter has joined #lima
01:50
egbert has quit [Ping timeout: 260 seconds]
01:55
egbert has joined #lima
02:05
mripard has quit [Ping timeout: 264 seconds]
02:47
warpme_ has quit [Quit: Connection closed for inactivity]
06:23
maciejjo has quit [Remote host closed the connection]
09:23
_whitelogger has joined #lima
09:33
dev1990 has quit [Quit: Konversation terminated!]
10:46
linkmauve has left #lima ["Disconnected: closed"]
13:04
smaeul has quit [*.net *.split]
13:04
bshah has quit [*.net *.split]
13:04
mmind00 has quit [*.net *.split]
13:04
smaeul has joined #lima
13:05
bshah has joined #lima
13:06
linkmauve has joined #lima
14:35
kaspter has quit [Quit: kaspter]
14:52
kaspter has joined #lima
14:57
kaspter has quit [Ping timeout: 264 seconds]
14:59
kaspter has joined #lima
16:20
<
linkmauve >
I printed the st->shader_has_one_variant decision process for each shader being compiled, they all gave the same results:
16:20
<
linkmauve >
vs: shader_has_one_variant=n, has_shareable_shaders=n, clamp_frag_depth_in_shader=n, clamp_vert_color_in_shader=n, lower_point_size=n, lower_ucp=y
16:20
<
linkmauve >
fs: shader_has_one_variant=n, has_shareable_shaders=n, lower_flatshade=y, lower_alpha_test=y, clamp_frag_color_in_shader=n, clamp_frag_depth_in_shader=n, force_persample_in_shader=n, lower_two_sided_color=y, lower_texcoord_replace=n
18:21
<
anarsoul|2 >
linkmauve: it could be missing support for shareable shaders
18:24
<
anarsoul|2 >
I can look into what needs to be done to enable it later this week
18:25
<
linkmauve >
Thanks. :)
18:25
<
anarsoul|2 >
but I expect it to fail miserably :)
18:26
<
linkmauve >
Let’s try!
18:29
<
linkmauve >
It didn’t fail in any way (GTK is single-threaded, that might be why?) but wasn’t any faster.
18:30
<
linkmauve >
(17.18s this time.)
18:50
<
anarsoul|2 >
btw, is it compilation time?
18:50
<
anarsoul|2 >
can you capture the shader for me please?
18:51
<
linkmauve >
It’s probably still reproducible with blend.glsl.
18:51
<
linkmauve >
Do you want the full shader-db output?
18:51
<
linkmauve >
Err, input.
18:52
<
anarsoul|2 >
I want a shader captured with MESA_SHADER_DUMP_PATH
18:57
<
anarsoul|2 >
please capture shaders with MESA_SHADER_CAPTURE_PATH, I want something that I can feed into compiler without much effort
18:58
<
anarsoul|2 >
so yeah, I want *.shader_test for mesa-shader-db
18:59
<
linkmauve >
Same URL.
18:59
<
linkmauve >
(I was intrigued why you wanted a MESA_SHADER_DUMP_PATH instead of MESA_SHADER_CAPTURE_PATH but didn’t question it. ^^')
19:03
<
anarsoul|2 >
I keep forgetting which one generates shader_test
19:04
<
anarsoul|2 >
which one takes 17s to compile?
19:07
<
anarsoul|2 >
oh, so it's not a single shader that takes 17s? but all of them?
19:09
<
anarsoul|2 >
well, that's 96 shaders (49 vs, 49 fs) after all...
19:09
<
linkmauve >
Yes, including some huge ones with like 21 spills.
19:10
<
linkmauve >
But before 1d5ce66b02016a8c985e3e3cc44df65ccfacca48 it took only ~7s to build everything.
19:11
<
linkmauve >
Still very long (not using the GLES backend drops the total startup time to 1.5s) but already much more bearable.
19:11
<
anarsoul|2 >
because it built it once
19:11
<
anarsoul|2 >
meh, I'll just implement disk cache to paper it over
19:12
<
linkmauve >
Every compilation which does occur would still take twice as much time as before.
19:15
<
anarsoul|2 >
OK, I asked anholt on #dri-devel about this commit
19:16
<
anarsoul|2 >
I doubt he'll respond earlier than Monday though
19:16
<
anarsoul|2 >
it's his daytime job :)
19:17
<
anarsoul|2 >
anyway, I'll work on disk cache for lima, that should help with startup times, we'll figure out how to avoid compiling it second time in parallel
19:19
<
linkmauve >
Thanks!
19:21
<
linkmauve >
Oh, GTK people did a whole bunch of optimisations since then, time to rebuild.