mrfixit2001 has quit [Remote host closed the connection]
stikonas has quit [Remote host closed the connection]
_whitelogger has joined #panfrost
marcodiego has quit [Quit: Leaving]
_whitelogger has joined #panfrost
<urjaman>
mk
<urjaman>
updating my system (and with that, LLVM) fixed that llvmpipe thing it seems
<rcf>
Come to think of it, that was actually causing issues for me at some point in the past.
<rcf>
I just can't remember what it was that broke even with software rendering, because it was a dumb idea to try anyway....
rcf1 has joined #panfrost
rcf1 has quit [Client Quit]
stikonas has joined #panfrost
warpme_ has joined #panfrost
<alyssa>
The constraints are being tracked~
<daniels>
narmstrong: can you please allow more concurrent jobs on the baylibre runner
<narmstrong>
daniels: only one board is tagged for Mesa right now, how much concurrency is needed ?
<daniels>
narmstrong: given that there are two jobs which run, having them run concurrently would certainly be extremely helpful
<daniels>
the two lima boards are different architectures
<daniels>
what I mean is to make the GitLab CI runner greedy, so it can submit many jobs to LAVA at the same time, and then let LAVA queue them
<narmstrong>
daniels: on yes you’re right there is 3 board for lima, I’ll update to 2 asap
<daniels>
narmstrong: thanks, though again if you can make it 4-8 then we can at least parallelise submission to LAVA, which might allow the dispatcher to download the initramfs, etc
<daniels>
and LAVA can take care of not scheduling jobs when it shouldn't
<narmstrong>
daniels: sure, I’ll also try to add more of these lima boards on a second slave at some point