ChanServ changed the topic of #lima to: Development channel for open source lima driver for ARM Mali4** GPUs - Kernel driver has landed in mainline, userspace driver is part of mesa - Logs at https://oftc.irclog.whitequark.org/lima/
hexdump01 has joined #lima
hexdump0815 has quit [Ping timeout: 480 seconds]
Daanct12 has joined #lima
Daanct12 has quit [Quit: WeeChat 4.6.3]
mripard has quit [Quit: WeeChat 4.6.3]
xdarklight_ has joined #lima
xdarklight has quit [Read error: Connection reset by peer]
dsimic is now known as Guest18783
dsimic has joined #lima
Guest18783 has quit [Ping timeout: 480 seconds]
xdarklight_ is now known as xdarklight
tlwoerner_ has quit [Remote host closed the connection]
tlwoerner has joined #lima
hexdump01 has quit []
hexdump0815 has joined #lima
<sarbes>
It seems that there are quite a few platform specific opcodes. I don't think that introducing a few for Utgard would be so different.
<sarbes>
But having implemented the folding in NIR, I don't think there are advantages for this use case, since the constant load is simple enough.
<sarbes>
I had some teaching my moments during this exercise. I think that some of the currently unused mul/add instructions would benefit from having their own opcodes.
<sarbes>
By "implemented the folding in NIR", what I meant was I just introduced new opcodes and wired them up in Lima.