<anarsoul>
nhp: I guess kwin shaders need some love?
<anarsoul>
from kwin devs
<nhp[m]>
Also the blue effect in KWin is broken with panfrost
<nhp[m]>
blur*
stikonas has quit [Remote host closed the connection]
stikonas_ has quit [Remote host closed the connection]
MoeIcenowy has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
MoeIcenowy has joined #panfrost
cwabbott_ has joined #panfrost
cwabbott has quit [Ping timeout: 264 seconds]
cwabbott_ is now known as cwabbott
vstehle has quit [Ping timeout: 246 seconds]
bbrezillon has quit [Ping timeout: 246 seconds]
<nhp[m]>
Turns out using KWin with GLES does not require rebuilding Qt
<nhp[m]>
Setting `KWIN_COMPOSE=O2ES` will do it
<alyssa_>
nhp[m]: how's perf there?
<nhp[m]>
Great
<nhp[m]>
I haven't measured it but it looks to be 60FOS quite consistently as long as nothing demanding is going on
<nhp[m]>
FPS*
<nhp[m]>
I noticed there's a bit of stuttering with RetriArch running in fullscreen without compositing disabled, not I'd have to test that on another system to see if that's really a panfrost issue
<alyssa_>
BTW, to anyone wondering if current events will impact panfrost dev - https://xkcd.com/2276/
<nhp[m]>
Relatable
alyssa_ is now known as alyssa
icecream95 has joined #panfrost
<icecream95>
mrfixit2001: The blob has a combined EGL / GLES library, but Mesa separates them. If you want to use SDL_VIDEO_*_DRIVER, you'll need to specify the actual library for each and not just libGLESv2.so.
<alyssa>
icecream95: evening :-)
<icecream95>
alyssa: It's still the afternoon here. :)
<alyssa>
Fair enough!
_whitelogger has joined #panfrost
marcodiego has quit [Ping timeout: 250 seconds]
davidlt has joined #panfrost
<icecream95>
For some of the shaders in Tesseract panfrost_dep_graph_contains_batch can take more than 5 seconds for debug builds
_whitelogger has joined #panfrost
_whitelogger has joined #panfrost
vstehle has joined #panfrost
<icecream95>
I've got Tesseract running now, but I can't actually see anything yet.
<icecream95>
Performance starts at about 30fps, but soon drops significantly, and memory usage climbs.
<icecream95>
panfrost_batch_fence_is_signaled is using up most of CPU time...
<HdkR>
spin wait, waiting for GPU to complete?
<icecream95>
HdkR: It looks like the fences are never signaled, so the fence readers array just keeps growing to thousands of elements.
<HdkR>
ah, neat
<icecream95>
If I try running it with PAN_MESA_DEBUG=trace, I get `Access to unknown memory bdb72fdfbdb72fd0 in ../src/panfrost/pandecode/decode.c:1919`
_mixfix41_ has joined #panfrost
guillaume_g has joined #panfrost
nerdboy has joined #panfrost
mias has joined #panfrost
nerdboy has quit [Ping timeout: 246 seconds]
_whitelogger has joined #panfrost
bbrezillon has joined #panfrost
anarsoul|c has quit [Ping timeout: 272 seconds]
ric96 has quit [Ping timeout: 272 seconds]
anarsoul|c has joined #panfrost
ric96 has joined #panfrost
<icecream95>
mrfixit2001: The DRM GLES code you linked to Works For Me (TM). Maybe permissions aren't being set up correctly on your side - try chmod a+rw /dev/dri/*
afaerber has quit [Remote host closed the connection]
afaerber has joined #panfrost
<icecream95>
HdkR: I had accidentaly commented out the fence referencing code, so the fences stuff was just a false alarm.
stikonas has joined #panfrost
_whitelogger has joined #panfrost
icecream95 has quit [Ping timeout: 264 seconds]
jernej has quit [Read error: Connection reset by peer]
raster has joined #panfrost
pak0st has joined #panfrost
_mixfix41_ has quit [Ping timeout: 264 seconds]
jernej has joined #panfrost
pak0st has quit [Quit: Konversation terminated!]
pak0st has joined #panfrost
pak0st has quit [Remote host closed the connection]
_whitelogger has quit [Remote host closed the connection]