alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - - Logs - <daniels> avoiding X is a huge feature
<HdkR> Looks like Bifrost has either a CLZ or CTZ instruction. Found it while testing findMSB/findLSB
<HdkR> Need to read the shader to see wtf it is actually doing
<alyssa> HdkR: Probably same as Midg
<alyssa> HdkR: Namely, a clz instruction
<alyssa> And then they do 31-clz() or whatever
<HdkR> yea I see a sub 31-op
cwabbott has quit [Ping timeout: 258 seconds]
<HdkR> Yea, just a clz
<alyssa> This is really the biggest mystery of the lot:
<alyssa> ADD, (1 - src_alpha), src_alpha
<HdkR> Feels weird to have a clz instead of a lsb/msb op since GLSL doesn't define a way to clz
<alyssa> Requires a negate on the destination
<HdkR> bweh?
<alyssa> IKR
<alyssa> So either the clip bit doesn't do what I think
<alyssa> Or the negate doesn't
<alyssa> Maybe there's some special case, like...
<alyssa> If you complement the dominant via clip, it negates the non-dominant as well. For some reason.
<alyssa> Smells like errata more than intended functionality but I dunno
<alyssa> Regardless, though, with that bizarre assumption for one half the question, I'm passing add tests
<alyssa> It's so tempting to force blend shaders but that'll bite me, I know
vstehle has quit [Ping timeout: 252 seconds]
<alyssa> Integers all ove
hlmjr has joined #panfrost
cwabbott has joined #panfrost
<alyssa> Blend shader stuff on the ML
<alyssa> Would love a CI run, hmm
<HdkR> ooo
<HdkR> Gonna have to take a look at that after dinner :D
<alyssa> :D
cwabbott has quit [Ping timeout: 264 seconds]
cwabbott has joined #panfrost
indy has quit [Remote host closed the connection]
vstehle has joined #panfrost
<tomeu> alyssa: ok, I'm going to schedule another exact run and see if the failures change
<alyssa> tomeu: Er, I just pushed, awkward timing
<HdkR> alyssa: I guess a limitation of the blend shader work is lack of support for blend_equation_advanced as well?
<HdkR> Theoretically easy to extend to support that in the future once anyone cares
<HdkR> Also, oh snap, called out
<alyssa> HdkR: There's already a lowering pass for that so I didn't bother yet
<HdkR> ah
<alyssa> Biggr issue I forgot about is color_mask, that's for v2 I guess
<alyssa> tomeu: Oh, wait, you're pushing to armhf and I to aarch64, different targets? :)
hlmjr has quit [Quit: Konversation terminated!]
herbmillerjr has joined #panfrost
<tomeu> alyssa: oh, what did you push?
<alyssa> tomeu: Bunch of random fixes I had queued up... I want to know what the net score turns out to be
<tomeu> ah, ok
<HdkR> How is store_deref supposed to work? Looking in to it a little bit, does deref_var allow you to calculate a memory offset?
<alyssa> HdkR: Call lower_io
<alyssa> Your backend should never see a deref
<alyssa> If you're writing a nir pass that's different :p
<alyssa> Unless you mean from reviewing my pass
<HdkR> alyssa: I'm seeing it with SSBO accesses. Curious about it :P
<alyssa> Oh..
<alyssa> Dunno
<alyssa> Haven't got there
<HdkR> :D
<HdkR> I won't get until this weekend to work on it more
<HdkR> That roughly makes up for the week of work I missed as well :P
Moiman has joined #panfrost
NeuroScr has quit [Quit: NeuroScr]
cwabbott has quit [Quit: cwabbott]
yann has joined #panfrost
cwabbott has joined #panfrost
stikonas has joined #panfrost
stikonas_ has joined #panfrost
stikonas has quit [Ping timeout: 245 seconds]
<urjaman> whoops, i didnt reboot the C201 at all during the week after i installed the rc7 kernel ... and now it's already too old
<urjaman> (that is, i "installed" it and was like meh it'll get tested whenever i get to rebooting ... nope, didnt get tested before now :P)
narmstrong has quit [Ping timeout: 264 seconds]
narmstrong has joined #panfrost
stikonas_ has quit [Remote host closed the connection]
afaerber has quit [Quit: Leaving]
afaerber has joined #panfrost
MoeIcenowy has quit [Quit: ZNC 1.6.5+deb1+deb9u1 -]
MoeIcenowy has joined #panfrost
herbmillerjr has quit [Quit: Konversation terminated!]
metaverse has joined #panfrost
<alyssa> So today wasn't going to be a Panfrost work day, buuuuuuut
<alyssa> I woke up just now with a cosmic vision on how to implement that feature I tried to do the other day efficiently :p
<alyssa> Namely, csel with vectorized conditions... Midg can only do scalar conditions... but note the identity holds:
<alyssa> C ? A : B = (C ? A : 0) | (C ? 0 : B) = (C & A) | (~C & B)
<alyssa> iand r0, C, A
<alyssa> iandnot r0, B, C
<alyssa> er
<alyssa> iand r0, C, A
<alyssa> iandnot r1, B, C
<alyssa> ior out, r0, r1
<alyssa> 3 ops to do a vec4 compare
<alyssa> (Compare to 8 we presently emit, or 4 in the best case but that's probably not possible to hit)
cwabbott has quit [Read error: Connection reset by peer]
<sphalerite> alyssa: I've been supporting you via liberapay, but apparently your account isn't configured to receive new donations or something? Are you in principle still accepting donations?
<urjaman> didnt liberapay just end the stuff ... or actually, i didnt check what happened after their problem way back
* urjaman checks
<urjaman> I remember sending her about a year worth of $1/month at some point when liberapay said they have payment processor problem of some sort
cwabbott has joined #panfrost
<alyssa> sphalerite: Something happened with liberapay last July, I haven't gotten around to figuring out how to reenable it since. I probably should have officially shut the account off but there hasn't been any real activity so it hasn't been hi-prio in either way
<sphalerite> ah ok
<alyssa> And donations were supposed to be tied to my blog more than anything \shrug/
<sphalerite> urjaman: yeah they have a new model which pays the money out directly in advance so they don't have to hold funds
<alyssa> sphalerite: ..So if I'm not configured, what's happening...?
<sphalerite> " Cannot be renewed because the account of the recipient isn't ready to receive new payments."
<alyssa> Ah.
<alyssa> FWIW the integer arithmetic alone is 3 ops and I don't think it's possible to do lower than that, but they're all scheduler-friendly ops
stikonas has joined #panfrost
<urjaman> sphalerite: yeah i read that now (and to me them not holding funds kinda breaks the service...)
herbmillerjr has joined #panfrost
metaverse has quit [Quit: metaverse]
belgin has joined #panfrost
belgin has quit [Quit: Leaving]
yann has quit [Ping timeout: 248 seconds]
pH5 has joined #panfrost
afaerber has quit [Quit: Leaving]
NeuroScr has joined #panfrost
_whitelogger has quit [Ping timeout: 248 seconds]