kaspter has quit [Read error: Connection reset by peer]
kaspter has joined #lima
Barada has joined #lima
kaspter has quit [Ping timeout: 265 seconds]
kaspter has joined #lima
buzzmarshall has quit [Remote host closed the connection]
chewitt has quit [Quit: Zzz..]
chewitt has joined #lima
dddddd has quit [Ping timeout: 258 seconds]
<anarsoul>
cwabbott: could you push your latest version to your gitlab repo?
yuq825 has joined #lima
kaspter has quit [Ping timeout: 265 seconds]
kaspter has joined #lima
camus1 has joined #lima
kaspter has quit [Ping timeout: 260 seconds]
camus1 is now known as kaspter
yann has joined #lima
<cwabbott>
anarsoul: yes, but that's a hard problem because you don't actually know where the users of the loads are going to go until after the post-RA scheduler
<cwabbott>
so there isn't really a good way to know which loads to try to put in the same vec4 register
Barada has quit [Quit: Barada]
Barada has joined #lima
chewitt has quit [Quit: Adios!]
chewitt has joined #lima
dddddd has joined #lima
anarsoul has quit [Remote host closed the connection]
anarsoul has joined #lima
yuq825 has quit [Quit: Leaving.]
Barada has quit [Remote host closed the connection]
yann has quit [Remote host closed the connection]
<anarsoul>
cwabbott: we can try to swap registers in scheduler
buzzmarshall has joined #lima
chewitt has quit [Quit: Zzz..]
chewitt has joined #lima
yann has joined #lima
<anarsoul>
i.e. if we failed to insert ld_reg or st_reg into instruction we can check whether another ld_reg/st_reg node with the same reg has already been scheduled and if not attempt to change register number to unused or swap with some other not-yet-scheduled reg
yann has quit [Ping timeout: 255 seconds]
yann has joined #lima
cwabbott has quit [Quit: cwabbott]
dllud has quit [Remote host closed the connection]