ChanServ changed the topic of #lima to: Development channel for open source lima driver for ARM Mali4** GPUs - Kernel has landed in mainline, userspace driver is part of mesa - Logs at https://people.freedesktop.org/~cbrill/dri-log/index.php?channel=lima and https://freenode.irclog.whitequark.org/lima - Contact ARM for binary driver support!
yuq825 has joined #lima
yann has quit [Ping timeout: 240 seconds]
yann has joined #lima
yuq825 has quit [Remote host closed the connection]
jernej has quit [Ping timeout: 252 seconds]
jernej has joined #lima
cp- has quit [Quit: Disappeared in a puff of smoke]
Barada has joined #lima
dddddd has quit [Ping timeout: 250 seconds]
buzzmarshall has quit [Remote host closed the connection]
cp- has joined #lima
yuq825 has joined #lima
chewitt has joined #lima
bhoj has quit [Remote host closed the connection]
bhoj has joined #lima
warpme_ has joined #lima
yuq825 has quit [Ping timeout: 256 seconds]
yuq825 has joined #lima
cwabbott_ has joined #lima
cwabbott has quit [Ping timeout: 265 seconds]
cwabbott_ is now known as cwabbott
jonkerj has quit [Remote host closed the connection]
jonkerj has joined #lima
<rellla> yuq825, anarsoul: https://gitlab.freedesktop.org/lima/mesa/-/issues/129#note_452970 - ugly, but works
enunes has joined #lima
monstr has joined #lima
Barada1 has joined #lima
Barada has quit [Ping timeout: 256 seconds]
Barada1 is now known as Barada
dddddd has joined #lima
<rellla> yuq825: with shadow bo you just mean a cpu memory area that also keeps the index buffer?
<yuq825> yes, but it's embedded inside the lima_resource, so mesa just see normal pipe_resource and update it using transfer map
<yuq825> so we can keep this cpu mem sync with gpu mem
<rellla> so i have to add some "void *index_buffer" to lima_resource and change lima_transfer_map to update that one?
<yuq825> yes
<yuq825> better name with void *shadow
<rellla> ok, let's see if i can do that :)
yuq825 has quit [Quit: Leaving.]
Barada has quit [Quit: Barada]
monstr has quit [Remote host closed the connection]
buzzmarshall has joined #lima
<anarsoul> rellla: have you benchmarked your MR?
yann has quit [Ping timeout: 265 seconds]
yann has joined #lima
<rellla> anarsoul: not yet. atm i don't really know, how to do that...
<anarsoul> try q3a timedemo?
<anarsoul> btw sorting indices every draw is probably expensive
<anarsoul> you have to implement some caching (see minmax cache that we're using)
<rellla> probably, i have to think about many things still :)
<rellla> atm, i was just glad to get it coded and nothing gets broken
<anarsoul> yeah, that's cool
<anarsoul> I guess the point is to get it performing better than current code
<anarsoul> so you need some kind of benchmark to establish baseline
<anarsoul> and measure improvement
<rellla> anarsoul: yes, without benchmarking it's pointless. is there any possibility to do some measurement without gui?
<anarsoul> good question
<anarsoul> I don't know the answer :)
e has quit [*.net *.split]
e has joined #lima
buzzmarshall has quit [Remote host closed the connection]