ChanServ changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard + Bifrost + Valhall - Logs - I don't know anything about WSI. That's my story and I'm sticking to it.
cphealy has joined #panfrost
<hays> its not an aquarium to eat out of haha
<alyssa> "Some people are vegetarians because they love animals. I'm a vegetarian because I hate plants."
Daanct12 has joined #panfrost
<hays> what is v10 referring to?
<alyssa> the architecture in mali-g610
<alyssa> i.e. rk3588
<alyssa> (v10 is marketed as "Mali-Gxy0")
Leopold_ has quit [Ping timeout: 480 seconds]
Daanct12 has quit [Ping timeout: 480 seconds]
Daanct12 has joined #panfrost
Daaanct12 has joined #panfrost
Daanct12 has quit [Ping timeout: 480 seconds]
<hays> Ahh figured it was something like that.
Daaanct12 has quit [Quit: Leaving]
davidlt has joined #panfrost
davidlt has quit [Ping timeout: 480 seconds]
floof58 is now known as Guest2164
floof58 has joined #panfrost
Guest2164 has quit [Ping timeout: 480 seconds]
guillaume_g has joined #panfrost
davidlt has joined #panfrost
rasterman has joined #panfrost
hanetzer has quit [Remote host closed the connection]
alarumbe has quit [Ping timeout: 480 seconds]
alarumbe has joined #panfrost
MajorBiscuit has joined #panfrost
nlhowell has joined #panfrost
chewitt has joined #panfrost
alarumbe has quit [Ping timeout: 480 seconds]
rasterman has quit [Read error: Connection reset by peer]
rasterman has joined #panfrost
alarumbe has joined #panfrost
greenjustin_ is now known as greenjustin
MajorBiscuit has quit [Ping timeout: 480 seconds]
chewitt has quit [Quit: Zzz..]
davidlt has quit [Ping timeout: 480 seconds]
mensi has joined #panfrost
<mensi> so, I'm still on my (foolish?) quest to try rusticl and I got clinfo as far as actually listing my Mali-T860 but I'm running into the unreachable() at because dest_size=64 and nr_comps=3 so the bitsize check doesn't match any branches... Is that completely wild or explainable? Any debugging hints?
<alyssa> it's not "try"
<alyssa> it's "if you want spend hundreds of hours minimum writing an opencl implementation"
<alyssa> (I say this having spent hundreds of hours on the newer mali and still don't have something passing CTS)
<mensi> So I guess my impression that rusticl generates NIR that then goes to the card is naiive and there's far more code in the middle missing?
<alyssa> well, you need a compiler that ingests NIR and spits out native code
mensi has quit []
<alyssa> Midgard assembly in your case
mensi has joined #panfrost
<alyssa> and by nature of OpenCL being so much more expressive, the NIR that Rusticl emits is much more complicated than the NIR that the GLSL compiler producers
<alyssa> *produces
<alyssa> need to teach the compiler (Midgard in this case) to handle all the new OpenCL shiny
<alyssa> 8-bit integers, 64-bit integers, numerical precision issues, etc
<mensi> I see
<alyssa> i'm also not aware of a single linux/arm64/opencl workload that's faster on the GPU than the CPU
<alyssa> so it's difficult to motivate.
<mensi> oh? are the NEON and whatnot extensions so good or the GPUs so bad?
<robclark> usually the cpu can saturate bandwidth.. tilers get a big leg up when it comes to 3d workloads by just avoiding a lot of memory bandwidth. But compute workloads don't benefit from that. So if it comes down to CPU saturating bandwidth vs GPU saturating bandwidth, it comes down to which draws less power doing so
<mensi> as in geometry / textures / the big data is relatively static for 3D -> you only copy once?
<mensi> actually do these Mali GPUs even have dedicated memory or are they saturating the main memory bus for everything?
<robclark> as in all the memory traffic to/from color/depth/stencil doesn't leave the gpu core
<robclark> no dedicated memory other than caches and tile buffer
<mensi> I see... are these dedicated NPUs like rock chip sticks in newer chips any better in that regard?
<mensi> or is it also just a memory contention thing?
<robclark> I mean, it comes down to what the bottleneck in a given workload.. if it is memory bandwidth, the gpu doesn't have any more than the cpu.. that was the point I was trying to make
<mensi> true, like what I was thinking is all these inference workloads and whatnot sound quite memory intensive if you don't have any considerable dedicated space to keep intermediates
<alyssa> correct
<alyssa> in general don't expect to run serious workloads on not serious hardware :p
<mensi> hehe of course, but I feel bad if I don't at least try anything with all these SBCs littering my desk ;)
<alyssa> Where are all these SBCs coming from?!
<mensi> China? ;)
<alyssa> hmm seems legit
<mensi> I was positively surprised how much better the OSS support for these is these days thanks to you people here :)
<alyssa> I guess that's good :)
<alyssa> usually i only hear when people are unhappy with it so based on my inbox you'd think support is garbage.
<alyssa> (who's jaded? not me! /s )
<mensi> yeah I guess that is one of those traits of humanity, you can share as much goodness as you want, there will always be people with entitlement issues and complaints
<alyssa> well, that, and drivers are infrastructure
<alyssa> most people don't think about them unless they break
<alyssa> which means as a driver developer by the time I talk to a user they're probably quite frustrated :)
<mensi> well, my driver writing experience so far was only for where we didn't really have users ;)
<alyssa> ah..
mensi has quit [Remote host closed the connection]
rasterman has quit [Quit: Gettin' stinky!]