sghuge has quit [resistance.oftc.net synthon.oftc.net]
balrog- has quit [resistance.oftc.net synthon.oftc.net]
lstrano_ has quit [resistance.oftc.net synthon.oftc.net]
Mangix has quit [resistance.oftc.net synthon.oftc.net]
clever has quit [resistance.oftc.net synthon.oftc.net]
kode54 has quit [resistance.oftc.net synthon.oftc.net]
anholt has quit [resistance.oftc.net synthon.oftc.net]
ecaoctr^ has quit [resistance.oftc.net synthon.oftc.net]
orbea has quit [resistance.oftc.net synthon.oftc.net]
ybogdano has quit [resistance.oftc.net synthon.oftc.net]
stsquad has quit [resistance.oftc.net synthon.oftc.net]
jhugo has quit [resistance.oftc.net synthon.oftc.net]
mareko has quit [resistance.oftc.net synthon.oftc.net]
codingkoopa3218 has quit [resistance.oftc.net synthon.oftc.net]
JPEW has quit [resistance.oftc.net synthon.oftc.net]
cengiz_io has quit [resistance.oftc.net synthon.oftc.net]
cbraga3 has quit [resistance.oftc.net synthon.oftc.net]
sally has quit [resistance.oftc.net synthon.oftc.net]
jhli has quit [resistance.oftc.net synthon.oftc.net]
enzuru has quit [resistance.oftc.net synthon.oftc.net]
tlwoerner has quit [resistance.oftc.net synthon.oftc.net]
mattst88 has quit [resistance.oftc.net synthon.oftc.net]
dianders has quit [resistance.oftc.net synthon.oftc.net]
anarsoul has quit [resistance.oftc.net synthon.oftc.net]
ids1024 has quit [resistance.oftc.net synthon.oftc.net]
Karyon has quit [resistance.oftc.net synthon.oftc.net]
SolarAquarion has quit [resistance.oftc.net synthon.oftc.net]
swivel has quit [resistance.oftc.net synthon.oftc.net]
demarchi has quit [resistance.oftc.net synthon.oftc.net]
kurufu has quit [resistance.oftc.net synthon.oftc.net]
Hazematman has quit [resistance.oftc.net synthon.oftc.net]
Lightsword has quit [resistance.oftc.net synthon.oftc.net]
MoeIcenowy has quit [resistance.oftc.net synthon.oftc.net]
cyrozap has quit [resistance.oftc.net synthon.oftc.net]
nukelet has quit [resistance.oftc.net synthon.oftc.net]
ngcortes has quit [resistance.oftc.net synthon.oftc.net]
vgpu-arthur___ has quit [resistance.oftc.net synthon.oftc.net]
jrayhawk has quit [resistance.oftc.net synthon.oftc.net]
lemes has quit [resistance.oftc.net synthon.oftc.net]
graphitemaster has quit [resistance.oftc.net synthon.oftc.net]
ngcortes has joined #dri-devel
sghuge has joined #dri-devel
lstrano_ has joined #dri-devel
balrog- has joined #dri-devel
Mangix has joined #dri-devel
clever has joined #dri-devel
anholt has joined #dri-devel
kode54 has joined #dri-devel
ecaoctr^ has joined #dri-devel
orbea has joined #dri-devel
ybogdano has joined #dri-devel
stsquad has joined #dri-devel
jhugo has joined #dri-devel
mareko has joined #dri-devel
codingkoopa3218 has joined #dri-devel
JPEW has joined #dri-devel
cengiz_io has joined #dri-devel
cbraga3 has joined #dri-devel
sally has joined #dri-devel
jhli has joined #dri-devel
enzuru has joined #dri-devel
tlwoerner has joined #dri-devel
mattst88 has joined #dri-devel
vgpu-arthur___ has joined #dri-devel
Karyon has joined #dri-devel
ids1024 has joined #dri-devel
dianders has joined #dri-devel
anarsoul has joined #dri-devel
MoeIcenowy has joined #dri-devel
cyrozap has joined #dri-devel
kurufu has joined #dri-devel
Lightsword has joined #dri-devel
demarchi has joined #dri-devel
lemes has joined #dri-devel
graphitemaster has joined #dri-devel
SolarAquarion has joined #dri-devel
swivel has joined #dri-devel
jrayhawk has joined #dri-devel
nukelet has joined #dri-devel
Hazematman has joined #dri-devel
fab has joined #dri-devel
guludo has joined #dri-devel
phasta has quit [Quit: WeeChat 4.6.2]
tlwoerner_ has joined #dri-devel
tlwoerner has quit [Ping timeout: 480 seconds]
APic has joined #dri-devel
fab has quit [Quit: fab]
Calandracas has joined #dri-devel
JRepinc has quit []
JRepinc has joined #dri-devel
nerdopolis has joined #dri-devel
davispuh has joined #dri-devel
Lynne has quit [Read error: Permission denied]
fab has joined #dri-devel
kts has joined #dri-devel
kts has quit []
<tzimmermann>
javierm, thanks for the quick response
<tzimmermann>
i merged the original patch earlier today
<tzimmermann>
apologies for the fix. i didn't occur to me that there are format-helper tests for bt601
rgallaispou has joined #dri-devel
guludo has quit [Ping timeout: 480 seconds]
guludo has joined #dri-devel
fab has quit [Quit: fab]
fab has joined #dri-devel
JRepinc has quit []
JRepinc has joined #dri-devel
fab has quit [Read error: Connection reset by peer]
davispuh has quit [Ping timeout: 480 seconds]
<javierm>
tzimmermann: yeah, I didn't catch it either. I wonder if there could be a way to make that association
<javierm>
because is easy to miss unless you grep for tests
coldfeet has quit [Ping timeout: 480 seconds]
warpme has quit []
coldfeet has joined #dri-devel
<tzimmermann>
idk
feaneron has joined #dri-devel
evadot has quit [Remote host closed the connection]
asrivats_ has joined #dri-devel
coldfeet has quit [Ping timeout: 480 seconds]
Lynne has joined #dri-devel
evadot has joined #dri-devel
coldfeet has joined #dri-devel
warpme has joined #dri-devel
phasta has joined #dri-devel
parthi has joined #dri-devel
kzd has joined #dri-devel
parthiban has quit [Ping timeout: 480 seconds]
coldfeet has quit [Quit: Lost terminal]
asrivats_ has quit [Ping timeout: 480 seconds]
italove8 has joined #dri-devel
phasta has quit [Quit: WeeChat 4.6.2]
digetx has quit [Read error: Connection reset by peer]
digetx has joined #dri-devel
<robclark>
karolherbst: so thinking about it a bit more.. I could more easily support bindless on a7xx.. and if there was a way to support it for compute-only contexts, it wouldn't be too hard on a6xx.. maybe that is the most reasonable way to deal w/ tflite (where, I'm not even sure it won't go above 128 tex and/or img)
Caterpillar has joined #dri-devel
<karolherbst>
robclark: bindless is kinda a weird api, because of the resident call... though I could make all images resident and just roll with it
<karolherbst>
the question is just, how much of a problem would that be to make every single image resident all the time
<karolherbst>
alternatively, I'd do it on every kernel launch
<robclark>
it isn't really clear to me when they become un-resident
<robclark>
but tbh just make them resident immediately.. and we can optimize later if needed
<robclark>
need to get things running first before I can profile ;-)
<robclark>
I may want to intro a bindless_texture_compute_only sort of cap for a6xx (or even a7xx just to avoid exposing bindless in gl).. but we'll see
<karolherbst>
robclark: why would it matter to do it compute only?
<karolherbst>
on the GL side it's just another extension
<karolherbst>
I don't think anything uses bindless unless it is requested by the application
<robclark>
so, a6xx has 5 bindless bases.. we use that internally to have separate img/ssbo state per shader stage (which is what I meant by already using bindless internally within driver).. otherwise I'd have to munge galliums per-stage state into a single global state like I did on earlier gens (but earlier gens didn't have gs/tess so it was a bit less painful) a7xx ups the bindless bases to 8 so we have an additional one to use for
<robclark>
gallium api level bindless
parthiban has joined #dri-devel
<robclark>
with compute only context, there is only a single shader stage so the limit on # of bindless descriptor sets isn't a problem
jsa1 has quit [Ping timeout: 480 seconds]
<karolherbst>
mhhhh, I see
<karolherbst>
could turn it into a shader_cap
jsa1 has joined #dri-devel
<karolherbst>
robclark: oh, you also have a limit on bindless descriptors?
<karolherbst>
or is it limitless, but you can only have nr of sets
<robclark>
the size of descriptors is essentially limitless.. but the # of descriptors is fixed
parthi has quit [Ping timeout: 480 seconds]
<karolherbst>
okay
<karolherbst>
I mean.. going full in with bindless would allow me to ditch all the binding code, so there is definitely an advantage of doing so
<karolherbst>
and if I can just keep all images resident at no cost, the better
<karolherbst>
I also want to make buffers more bindless, but that's a bit more problematic
<robclark>
some other drivers might want you to keep the bindful path.. and I guess folks with <=a5xx would like to still be able to use rusticl so you might need to keep both paths :-(
<robclark>
(not sure how well rusticl runs on a5xx, so not sure if that is the end of the world.. but I guess panfrost/asahi/etc?)
<demarchi>
we could have something similar in maintainer-tools, but the bootstrap is not simply copy & paste :(
<sima>
yeah, actual functional tests with containers would be neat in dim, but a lot of work
bolson has joined #dri-devel
<sima>
maybe just one for dim setup? that seems to be the brittle thing where breakage is unnoticed for the longest
epoch101 has joined #dri-devel
<sima>
and also kinda annoying since it makes it harder for new committers to get going
epoch101 has quit []
epoch101 has joined #dri-devel
rgallaispou1 has joined #dri-devel
rgallaispou has quit [Read error: Connection reset by peer]
vliaskov_ has quit [Read error: No route to host]
<alyssa>
karolherbst: fwiw there's a gpu overhead penalty to bindless on m1
<karolherbst>
alyssa: what about indirect?
<karolherbst>
like I could also just manage a table of bound images/textures, but use indirects on the GPU side instead of fixed indicies
<karolherbst>
but I could imagine it has the same overhead as bindless ones on m1?
<alyssa>
different overhead, I guess
<alyssa>
what's wrong with what we have now exactly?
<karolherbst>
apparently there are applications binding the same image like 128 times to different kernel args
<karolherbst>
I mean it's an application bug
<karolherbst>
but I also kinda wanted to optimize the binding model I'm currently using, because rebinding things on each launch is also kinda annoying
<karolherbst>
but maybe that's fine
<karolherbst>
at least the indirect/handle would be uniform always
<alyssa>
i'm not sure this is actually an optimization on all hw
<karolherbst>
me neither
<karolherbst>
I don't have concrete plans to do either, just thinking out loud
<karolherbst>
however, I was considering using bindless to bind more images than the driver/hardware supports
<karolherbst>
like the first 32 could be bound, and if the kernel uses more, they are bindless
<karolherbst>
like only the additional ones
<karolherbst>
some applications depend on a FULL_PROFILE and that means 128 samplers and 64 images
DarkShadow44 has quit [Read error: Connection reset by peer]
DarkShadow44 has joined #dri-devel
<alyssa>
make it the gallium driver's problem to advertise enough images then
<alyssa>
since we already implement that bindless fallback internally
<alyssa>
and juggling both the internal one and the state tracker one is a mess
<alyssa>
as robclark mentioned
JRepinc has quit []
JRepinc has joined #dri-devel
<karolherbst>
It's good enough for me if driver wants to deal with it internally
lemonzest has quit [Quit: WeeChat 4.6.3]
hikiko_ has joined #dri-devel
dsimic is now known as Guest18097
dsimic has joined #dri-devel
hikiko has quit [Ping timeout: 480 seconds]
<robclark>
hmm, I guess I could emulate a higher limit with bindless internally.. an only advertise the higher virtual limits for compute shaders
Guest18097 has quit [Ping timeout: 480 seconds]
feaneron has quit [Ping timeout: 480 seconds]
jsa1 has quit [Ping timeout: 480 seconds]
tzimmermann has quit [Quit: Leaving]
coldfeet has joined #dri-devel
lemonzest has joined #dri-devel
warpme has quit []
feaneron has joined #dri-devel
warpme has joined #dri-devel
guludo has quit [Ping timeout: 480 seconds]
kts has joined #dri-devel
Jeremy_Rand_Talos_ has quit [Remote host closed the connection]
Jeremy_Rand_Talos_ has joined #dri-devel
rgallaispou1 has quit [Read error: Connection reset by peer]
warpme has quit []
JRepinc has quit []
JRepinc has joined #dri-devel
krei-se- has joined #dri-devel
krei-se has quit [Ping timeout: 480 seconds]
lynxeye has quit [Quit: Leaving.]
parthiban has quit [Read error: Connection reset by peer]
parthiban has joined #dri-devel
JRepinc has quit []
JRepinc has joined #dri-devel
RSpliet has quit [Quit: Bye bye man, bye bye]
lsntvt_ has quit [Ping timeout: 480 seconds]
RSpliet has joined #dri-devel
epoch101_ has joined #dri-devel
epoch101 has quit [Read error: Connection reset by peer]
<karolherbst>
that would work, yeah
kts has quit [Quit: Konversation terminated!]
epoch101_ has quit []
lsntvt_ has joined #dri-devel
feaneron has quit [Ping timeout: 480 seconds]
epoch101 has joined #dri-devel
guludo has joined #dri-devel
yogesh_m1 has joined #dri-devel
yogesh_mohan has quit [Ping timeout: 480 seconds]
rasterman has quit [Quit: Gettin' stinky!]
feaneron has joined #dri-devel
hikiko has joined #dri-devel
simon-perretta-img has quit [Ping timeout: 480 seconds]
hikiko_ has quit [Ping timeout: 480 seconds]
simon-perretta-img has joined #dri-devel
<alyssa>
is there an env var to bypass the shader cache for fossilize-replay?
<alyssa>
oh MESA_SHADER_CACHE_DISABLE=true works for vk too ok
alanc has quit [Remote host closed the connection]
alanc has joined #dri-devel
<dj-death>
alyssa: I use VK_ENABLE_PIPELINE_CACHE=0
<alyssa>
ooh nice
<alyssa>
thx
JRepinc has quit []
JRepinc has joined #dri-devel
<anholt>
we should probably set one of those in fossil_replay.sh
<alyssa>
anholt: possibly although in this case I was hitting fossilize-replay directly to grab a disassembly of an affected shader
<alyssa>
but of course there was no disasm the second time around because it was hot
<pendingchaos>
last I checked, fossil_replay.sh was faster with pipeline caching
<alyssa>
i might be holding it wrong
<anholt>
hmm, yeah, guess I haven't had trouble with shader status going missing with it due to caches.
simon-perretta-img has quit [Ping timeout: 480 seconds]
jsa1 has joined #dri-devel
fab has joined #dri-devel
sima has quit [Ping timeout: 480 seconds]
illwieckz has quit [Remote host closed the connection]
illwieckz has joined #dri-devel
vgpu-arthur___ has quit []
JRepinc has quit []
fab has quit [Ping timeout: 480 seconds]
JRepinc has joined #dri-devel
fab has joined #dri-devel
Nasina has joined #dri-devel
fab has quit [Ping timeout: 480 seconds]
epoch101 has quit [Ping timeout: 480 seconds]
haaninjo has joined #dri-devel
coldfeet has quit [Quit: Lost terminal]
JRepinc has quit []
warpme has joined #dri-devel
JRepinc has joined #dri-devel
sima has joined #dri-devel
jsa1 has quit [Ping timeout: 480 seconds]
sima has quit [Ping timeout: 480 seconds]
Nasina has quit [Read error: Connection reset by peer]