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!
jrmuizel has joined #lima
yuq825 has joined #lima
jrmuizel has quit [Remote host closed the connection]
jrmuizel has joined #lima
jrmuizel has quit [Remote host closed the connection]
jrmuizel has joined #lima
Da_Coynul has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
megi has quit [Ping timeout: 272 seconds]
jrmuizel has quit [Remote host closed the connection]
camus has joined #lima
kaspter has quit [Ping timeout: 246 seconds]
camus is now known as kaspter
<anarsoul> glmark2 -b ideas doesn't work, it requires proper liveness analysis, it never exits spill loop :(
<anarsoul> X11 works but cursor is missing :\
<anarsoul> running piglit on cf branch now
<anarsoul> hopefully it won't wedge gpu :)
<anarsoul> looks like gpu is still alive
<anarsoul> 15 new crashes, probably these changed from fail to crash
<anarsoul> will look in a minute
<anarsoul> OK, all the crashes are previous fails, but there's some regressions for complex shaders :\
<anarsoul> oh, that's due to changes in regalloc to fix loops
<anarsoul> not it tries to spill and fails
<anarsoul> OK, we need proper liveness analysis.
<anarsoul> yuq825: enunes: rellla: I'd appreciate some feedback on https://gitlab.freedesktop.org/mesa/mesa/merge_requests/1699
<yuq825> anarsoul: sure, will look into it this weekend
<anarsoul> I hope I'll have some time to implement liveness analysis
<anarsoul> I think I understand the basics but I've never done it before :)
<yuq825> don't worry, just try, you'll be fine
<yuq825> I also struggle on the compiler before especially the gpir
<yuq825> cwabbott saved me
<yuq825> a good experience
<anarsoul> ouch, some of crashes are existing bugs
<anarsoul> it's totally valid to read uninitialized register
<anarsoul> at least we get it from nir
<anarsoul> fixed some crashes
<anarsoul> rerunning piglit
<anarsoul> I added a dummy op to handle cases when we have a read from uninitialized register
<anarsoul> it's basically just a placeholder of ppir_dest for register
<anarsoul> I guess it'll be better to initialize var_nodes[] for registers from the very beginning
<anarsoul> anyway we add reg deps in a separate pass
<anarsoul> 24 regressions, 25 fixes
kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #lima
megi has joined #lima
monstr has joined #lima
yuq8251 has joined #lima
yuq825 has quit [Remote host closed the connection]
yuq8251 has quit [Remote host closed the connection]
Da_Coynul has joined #lima
Da_Coynul has quit [Client Quit]
yuq825 has joined #lima
Da_Coynul has joined #lima
kaspter has quit [Quit: kaspter]
Da_Coynul has quit [Client Quit]
kaspter has joined #lima
Da_Coynul has joined #lima
kaspter has quit [Read error: Connection reset by peer]
Da_Coynul has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
kaspter has joined #lima
yuq825 has quit [Remote host closed the connection]
adjtm has quit [Ping timeout: 268 seconds]
jrmuizel has joined #lima
jrmuizel has quit [Read error: Connection reset by peer]
jrmuizel has joined #lima
afaerber has joined #lima
jrmuizel has quit [Remote host closed the connection]
adjtm has joined #lima
kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #lima
yuq825 has joined #lima
dddddd has joined #lima
yuq825 has quit [Remote host closed the connection]
jrmuizel has joined #lima
monstr has quit [Remote host closed the connection]
megi has quit [Ping timeout: 246 seconds]
<anarsoul> I wonder why cursor isn't drawn in X11 with my cf branch :\
<anarsoul> enunes: re: vector select
<anarsoul> since we don't have vector select with vector condition, should we introduce another nir op for one that we have?
<anarsoul> we can add NIR pass that converts all vector selects with condition that has swizzle xxxx to this new op, and then lower other vector selects to scalars
<anarsoul> sounds pretty easy to me
<enunes> anarsoul: that sounds like a good idea to me, but also more of an optimization, I guess it doesn't block the vector select MR as those ops would be lowered anyway?
<enunes> have to drop in a minute
megi has joined #lima
<anarsoul> enunes: it doesn't block https://gitlab.freedesktop.org/mesa/mesa/merge_requests/1639/ for sure
adjtm has quit [Ping timeout: 248 seconds]
kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #lima
adjtm has joined #lima
adjtm_ has joined #lima
adjtm has quit [Ping timeout: 268 seconds]
jrmuizel has quit [Remote host closed the connection]
Da_Coynul has joined #lima
jrmuizel has joined #lima
jrmuizel has quit [Remote host closed the connection]