alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - https://gitlab.freedesktop.org/panfrost - Logs https://freenode.irclog.whitequark.org/panfrost - <daniels> avoiding X is a huge feature
vstehle has quit [Ping timeout: 248 seconds]
herbmillerjr has joined #panfrost
NeuroScr has quit [Quit: NeuroScr]
NeuroScr has joined #panfrost
fysa has joined #panfrost
fysa has quit [Ping timeout: 258 seconds]
herbmillerjr has quit [Ping timeout: 272 seconds]
vstehle has joined #panfrost
NeuroScr has quit [Ping timeout: 268 seconds]
jolan has quit [Quit: leaving]
NeuroScr has joined #panfrost
<tomeu> who was offering boards with bifrost in them?
<HdkR> Do you need an N2?
<tomeu> well, I think I only need for someone to run the IGT tests on one of those
<tomeu> I don't have a particular interest in crowding my office with more boards :p
<HdkR> What would the IGT tests entail? It would just be blob status atm
<tomeu> we have some tests on IGT that should fail if the bifrost page table format isn't working correctly
<tomeu> yeah, for now I'm only interested in sanity testing the kernel
<tomeu> though I guess the clear job in IGT might need sometweaking for bifrost, not sure
<HdkR> Ah. I'm not running a newer kernel on the board
<HdkR> `Linux odroid 4.9.99 #2 SMP PREEMPT Tue Nov 27 13:45:34 KST 2018 aarch64 aarch64 aarch64 GNU/Linux`
<chewitt> I have 5.1 running on bifrost .. mali_kbase breaks with 5.2 and someone didn't feed me with patches for that yet
<chewitt> I have G52 ane G31 devices
<tomeu> chewitt: I was thinking of running 5.2-rc2 with panfrost (no mali_kbase or DDK at all)
<tomeu> and just running the IGT tests
<chewitt> what are the tests .. binaries or scripts?
<chewitt> I have no idea what IGT test are :)
<tomeu> that should have a good chance of passing
<tomeu> this will be harder, we might be missing some changes to the cmdstream: https://gitlab.freedesktop.org/drm/igt-gpu-tools/blob/master/tests/panfrost_submit.c
<chewitt> I think @HdkR is more likely to be helpful on the request :)
<chewitt> but if you do decide some hardware is required, it's easily organised
<tomeu> chewitt: which of these have midgard and which have bifrost? https://kernelci.org/soc/amlogic/
<tomeu> maybe baylibre would let me submit jobs there
<chewitt> anything with g12a in the name is G31, g12b is G52, and gxm is T820
fysa has joined #panfrost
<tomeu> awesome
<tomeu> I think it would be cool to have the kernel side of things sorted out on bifrost sooner rather than later
<tomeu> so we don't have superfluous delay once things start to work on userspace
fysa has quit [Ping timeout: 248 seconds]
<chewitt> that would be good
<HdkR> I wouldn't be able to get back to trying to push a newer kernel on to this device until next weekend
<HdkR> Not sure how many spoons I want to burn fussing with kernel instead of just working on the bifrost compiler
<tomeu> np, have asked in lkml for help with testing
<chewitt> narmstrong ^
<narmstrong> I can prepare a kernel for the N2 based on v5.2-rc1
paulk-leonov has quit [Max SendQ exceeded]
<tomeu> narmstrong: you will need this patch for bifrost: https://patchwork.freedesktop.org/patch/304731/
yann has joined #panfrost
paulk-leonov has joined #panfrost
jolan has joined #panfrost
yann|work has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
stikonas has joined #panfrost
paulk-leonov has joined #panfrost
stikonas has quit [Remote host closed the connection]
chewitt has quit [Quit: Zzz..]
paulk-leonov has quit [Max SendQ exceeded]
chewitt has joined #panfrost
paulk-leonov has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
yann|work has left #panfrost [#panfrost]
paulk-leonov has joined #panfrost
paulk-leonov has quit [Remote host closed the connection]
fysa has joined #panfrost
paulk-leonov has joined #panfrost
fysa has quit [Ping timeout: 248 seconds]
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
<tomeu> and my pine h64 has just arrived :)
<HdkR> woo
<HdkR> All the boards
afaerber has joined #panfrost
rhyskidd has joined #panfrost
afaerber has quit [Quit: Leaving]
afaerber has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
fysa has joined #panfrost
paulk-leonov has quit [Max SendQ exceeded]
paulk-leonov has joined #panfrost
fysa has quit [Ping timeout: 246 seconds]
fysa has joined #panfrost
fysa has quit [Ping timeout: 246 seconds]
<Lyude> narmstrong: if you prepare that kernel let me know, I want to try to start slowly getting myself into the habit of working on Panfrost again
<narmstrong> Lyude: ok
<HdkR> Lyude: Watch out, bad habits might over your life
<HdkR> might take over
<Lyude> H
<Lyude> HdkR: hm?
<HdkR> Lyude: Panfrost might take over like a bad habit if you do it too much. Highly addictive :P
<Lyude> hehe
<Lyude> HdkR: anything major with the bifrost work happen yet?
<alyssa> HdkR: it's true
<HdkR> Lyude: Mostly cleaning things up, making it more like what it needs to be as an actual MIR for emitting
<HdkR> Only did 6 hours last weekend
<HdkR> Mostly worked on cleaning up IO representation
<HdkR> I think UBO/Uniform loading is now creating sane MIR?
<HdkR> Found a bug in RA that I haven't fixed
<HdkR> Was around csel according to notes
<Lyude> HdkR: are we still using simple RA btw, or have we started taking advantage of passthrough registers and the like?
<HdkR> Haven't done that at all
<HdkR> Would probably need a pass that works on moving consumers and producers closer to one another and then an additional pass that runs post-bundling or something for choosing the temp registers
<HdkR> Haven't thought too hard on it :P
fysa has joined #panfrost
<tomeu> may be a good time to start a panfrost anonymous chapter
Elpaulo has quit [Quit: Elpaulo]
tgall_foo has joined #panfrost
<HdkR> I've got a Pandiction. I have to get my Pan fix every week
fysa has quit [Ping timeout: 272 seconds]
yann has quit [Ping timeout: 248 seconds]
BenG83 has joined #panfrost
fysa has joined #panfrost
fysa has quit [Ping timeout: 258 seconds]
<robclark> is it a pandemic?
<alyssa> robclark: Pretty sure you're patient 0
<robclark> :-P
pH5 has joined #panfrost
stikonas has joined #panfrost
fysa has joined #panfrost
afaerber has quit [Quit: Leaving]
Elpaulo has joined #panfrost
adjtm has joined #panfrost
adjtm_ has quit [Ping timeout: 248 seconds]
BenG83 has quit [Remote host closed the connection]
BenG83 has joined #panfrost
BenG83 has quit [Quit: Leaving]
herbmilleriw has quit [Ping timeout: 248 seconds]
herbmilleriw has joined #panfrost
NeuroScr has quit [Quit: NeuroScr]
NeuroScr has joined #panfrost
<cyrozap> It's causing a pandemonium!
<alyssa> cyrozap: True. Also, haven't seen you 'round these parts ina while!
<alyssa> HdkR: Have you tried nir_scheduler? It's in one of anholt's branches
<alyssa> The "moving closer to another" ought to be handled by it, I think....?
<alyssa> Or maybe it was about moving instructions that cause stalls _farther_ away. Depends on your config. Couldn't get it to work optimally myself but you may have better luck.
<alyssa> Lyude: FWIW, on Midgard I was doing simple RA for up until this week (rewrite patches on the list). It's really fine for a while; it's not critical for a long while.
<Lyude> alyssa:oh yeah I'm aware!
<Lyude> i'm not expecting us to get anything fancy, was just curious what's been going on with bf :)
<Lyude> *bifrost
<Lyude> i just relized bf is very ambiguous ;)
<alyssa> We've been over this, he's Midgard
<alyssa> (Kevin. My bf.)
<Lyude> hehe
stikonas has quit [Read error: Connection reset by peer]
jeez_ has joined #panfrost
<HdkR> =O
* HdkR claps for robclark
<robclark> HdkR: well, my mom has a chromebook now.. I figured sooner or later I'd have to know how to help my parents setup printers :-P
<HdkR> haha