ChanServ changed the topic of #etnaviv to: #etnaviv - the home of the reverse-engineered Vivante GPU driver - Logs https://oftc.irclog.whitequark.org/etnaviv
tlwoerner has quit [Remote host closed the connection]
tlwoerner has joined #etnaviv
mvlad has joined #etnaviv
lynxeye has joined #etnaviv
pcercuei has joined #etnaviv
karolherbst has quit [Quit: Konversation terminated!]
<tomeu> I am seeing many more failures when I run piglit or CL CTS tests in parallel, than when running them sequentially
<tomeu> anybody has any ideas on what state might be being trampled on?
<lynxeye> tomeu: I don't see anything totally obvious, especially since you are stalling the FE before writing any new compute state.
<tomeu> ok, something I want to do when I get back to the desktop is to check the caching configs in the cmdstream
<lynxeye> Did you try if using any of the other cache flush bits after the thread walker kicker changes anything?
<tomeu> those don't seem to change ever in all the testing I have done, but the ones at the very top might, as I am not diffing them with Mesa's because the blob sends them in a different submit
chewitt has joined #etnaviv
lynxeye has quit [Quit: Leaving.]
mvlad has quit [Remote host closed the connection]