<urjaman>
narmstrong: i dont think the - would be seen as a part of the symbol name :P
<narmstrong>
urjaman: lol
<narmstrong>
I have one issue with gitlab MR, we can’t « Cc » a person to notify for review, without this it stays inferior to emails
* urjaman
finds he should have read the mesa MR "guide" whilst doing the MR, not a long time ago and just assume you remember it
<urjaman>
I didnt do even half the things it recommends :P
fysa has joined #panfrost
<urjaman>
In an ideal world i'd like mailing lists, but my mental health doesnt like email for some reason
fysa has quit [Ping timeout: 240 seconds]
maccraft123 has joined #panfrost
klaxa has quit [Ping timeout: 276 seconds]
stikonas has joined #panfrost
klaxa has joined #panfrost
stikonas has quit [Remote host closed the connection]
<urjaman>
okay yeah I guess it'd make sense to label that MR as panfrost, but i'm such a nobody that gitlab will not let me do it
fysa has joined #panfrost
fysa has quit [Ping timeout: 240 seconds]
<daniels>
narmstrong: huh? just mention their username and they get CCed
<narmstrong>
daniels: hmm yeah, sounds dumb but haven’t thought as this
<urjaman>
daniels: thanks for the label :)
kaspter has quit [Ping timeout: 265 seconds]
megi has joined #panfrost
megi has quit [Ping timeout: 246 seconds]
fysa has joined #panfrost
fysa has quit [Ping timeout: 240 seconds]
raster has joined #panfrost
<alyssa>
narmstrong: @them, it's like twitter.
<alyssa>
urjaman: I just had the pleasure of trying out your oq patch -- thank you so much!!
<alyssa>
Works like a charm. See comments but tl;dr occlusion queries work now.
* alyssa
debugging spilling issue ugh
<alyssa>
Bisection is annoying
<alyssa>
Double-spilling patch is broken, I see.
<alyssa>
Well, we can fix it with LCRA.
<alyssa>
("You say that about everything." "Well, it's true.")
<alyssa>
In other news, I swear my computer is getting slower and I don't know why.
<alyssa>
Disk I/O in particular
<alyssa>
/dev/mapper/root 29G 22G 5.3G 81% /
<alyssa>
^ So not terrible there..
stikonas has joined #panfrost
fysa has joined #panfrost
fysa has quit [Ping timeout: 268 seconds]
kaspter has joined #panfrost
maccraft123 has quit [Quit: WeeChat 2.6]
maccraft123 has joined #panfrost
maccraft123 is now known as maccraft1
maccraft1 has quit [Client Quit]
fysa has joined #panfrost
fysa has quit [Ping timeout: 268 seconds]
<urjaman>
alyssa: thanks for the review, also yay that it works :D
fysa has joined #panfrost
<urjaman>
i'll be running the dEQP tests just for fun and practice when i get it all checked out and built, but since this is a C201 i decided to update the MR before playing with that
<urjaman>
(so yeah style things i think addressed, and i put that patch of yours into the series while at it)
fysa has quit [Ping timeout: 240 seconds]
<alyssa>
urjaman: Yup, it's ready for merge. Not sure why CI just collapsed.
<urjaman>
yeah i noticed the meson-windows thing just failed for some reason that i think is totally unrelated
<alyssa>
meson-windows was landed today/yesterday I think so
raster has quit [Quit: Gettin' stinky!]
<alyssa>
Now, to the great swizzle refactor of 2019 :p
<alyssa>
168 instances of "swizzle". Okay. This is fine.
<urjaman>
swizzle swizzle
<alyssa>
Essentially, right now MIR has swizzles packed tightly in a hw-specific format
<alyssa>
and it's a total mess
<alyssa>
I would like to have unpacked generic swizzles (like NIR) to do away with the million different cases
<alyssa>
And then lower the swizzle to the hardware-specific formats (plural!) later on
<alyssa>
Necessary to implement !32-bit vectors in a way that isn't bad.
fysa has joined #panfrost
kaspter has quit [Ping timeout: 245 seconds]
maccraft has joined #panfrost
maccraft is now known as maccraft123
enunes has quit [Read error: Connection reset by peer]
fysa has quit [Ping timeout: 245 seconds]
<alyssa>
aaand we're back, more swizzling
<urjaman>
swizzling is starting to sound like some kind of a funky dance move by now :P
<alyssa>
urjaman: tell me about it
fysa has joined #panfrost
maccraft has joined #panfrost
fysa has quit [Ping timeout: 250 seconds]
maccraft123 has quit [Ping timeout: 250 seconds]
BenG83 has joined #panfrost
fysa has joined #panfrost
fysa has quit [Ping timeout: 245 seconds]
davidlt has quit [Ping timeout: 240 seconds]
enunes has joined #panfrost
<alyssa>
Okay I guess I need to give the same treatment to load/st and tex
<alyssa>
Wonder if --- d'oh, okay I do need to normalize uhm okay
* alyssa
should put a dummy source in src0 for loads to get symmetry with stors
<alyssa>
So much benefit of hindsight
maccraft has quit [Quit: WeeChat 2.6]
maccraft123 has joined #panfrost
maccraft123 has quit [Client Quit]
fysa has joined #panfrost
maccraft123 has joined #panfrost
fysa has quit [Ping timeout: 265 seconds]
megi has joined #panfrost
fysa has joined #panfrost
fysa has quit [Ping timeout: 240 seconds]
<urjaman>
yay i can confirm that the occlusion tests pass 100% on the C201 too :) (== i managed to compile that stuff :P)