ChanServ changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - Logs https://oftc.irclog.whitequark.org/panfrost - <macc24> i have been here before it was popular
WoC has joined #panfrost
cphealy has quit [Read error: Connection reset by peer]
rasterman has quit [Quit: Gettin' stinky!]
WoC has quit [Remote host closed the connection]
WoC has joined #panfrost
nlhowell has joined #panfrost
nlhowell is now known as Guest1478
nlhowell has joined #panfrost
Guest1478 has quit [Ping timeout: 480 seconds]
cphealy has joined #panfrost
leah1 has quit []
leah has joined #panfrost
MajorBiscuit has joined #panfrost
tolszak has joined #panfrost
rasterman has joined #panfrost
JulianGro has quit [Remote host closed the connection]
nlhowell has quit [Ping timeout: 480 seconds]
nlhowell has joined #panfrost
nlhowell is now known as Guest1513
nlhowell has joined #panfrost
Guest1513 has quit [Ping timeout: 480 seconds]
JulianGro has joined #panfrost
atler has quit [Quit: atler]
atler has joined #panfrost
nlhowell has quit [Ping timeout: 480 seconds]
MajorBiscuit has quit [Ping timeout: 480 seconds]
Daanct12 has joined #panfrost
<tjcorley>
Hey, I was wondering if anyone could help me with my issue. I am trying to issue a WRITE_VALUE job to a bitfrost gpu using the kbase driver. I believe that I have setup the job correctly as I have copied the structure defitions and initialization from this commit: https://gitlab.freedesktop.org/mesa/mesa/-/commit/a0d1be30e1b4bc6f9440851c183ea03609b4f253
Danct12 has quit [Ping timeout: 480 seconds]
rellla has joined #panfrost
<tjcorley>
So I believe this means I am messing up filling out the kbase atom structure, and I am mostly confused at what value I would use for the `core_req` field. If I populate it with seemingly any value the event_code I recieve back is `BASE_JD_EVENT_JOB_CONFIG_FAULT`
<tjcorley>
any thoughts would be welcome, thanks :)
nlhowell has joined #panfrost
tris_ is now known as trixie
<robmur01>
I suspect the only way to get any support for doing anything on top of kbase is to get a job in Arm's GPU software group :P
Daanct12 is now known as Danct12
<tjcorley>
haha will go submit my resume I guess :P
WoC has quit [Read error: Connection reset by peer]
<tjcorley>
Yep that helped. Turned out my `base_jd_atom_v2`struct definition was wrong... Guess wherever I copied my mali header definitions was wrong :P Thanks again :)
floof58 has quit [Remote host closed the connection]
floof58 has joined #panfrost
cphealy has quit [Ping timeout: 480 seconds]
tjcorley has quit [Ping timeout: 480 seconds]
tjcorley has joined #panfrost
<HdkR>
:D
tolszak has quit [Ping timeout: 480 seconds]
rcf has quit [Quit: WeeChat 3.2.1]
hl` has joined #panfrost
rcf has joined #panfrost
cphealy has joined #panfrost
<alyssa>
It looks like Valhall has a nice way to implement SSBOs
<alyssa>
(More efficiently than just shoving a pointer into a uniform and doing 64-bit math in the shader, and with partial bounds checking)
<daniels>
progress!
<HdkR>
ooo. Does it support an SSBO binding table or something? :D
<alyssa>
HdkR: I'm not 100% sure on the details, but I think the driver can bind SSBOs as if they were UBOs and have that Just Work
<alyssa>
....maybe that works on Bifrost too..
<alyssa>
no, that specifically does not work on Bifrost, boo
<HdkR>
Interesting. Does the type of access change depending on instruction type/flags then
<alyssa>
tbd
<alyssa>
still have a while to go before implementing SSBOs, but my curiosity got the best of me
<HdkR>
Everyone loves arbitrary memory access
<alyssa>
(And more importantly ... the Midgard code path for SSBOs ought to work fine on Valhall, so it's an optimization at best)