#
✕ nick:chadmed
#asahi-dev · 2026-06-04 11:46
chadmedthe measure is per mhz, you need to peg one core to its top pstate and then clamp the coremark task to it
#asahi-dev · 2026-06-04 11:43
chadmedthat is the correct way. the fastest cores in the system are by convention always 1024, and the others are normalised relative to that
#asahi-dev · 2026-06-04 10:01
chadmedyeah it sounds reasonable to me, especially if its reproducible
#asahi-dev · 2026-06-04 09:51
chadmedbecause thats what crappy ecores usually come out at :P
#asahi-dev · 2026-06-04 09:50
chadmedblizz is like 74% of avalanche
#asahi-dev · 2026-06-04 09:50
chadmedyeah
#asahi-dev · 2026-06-04 09:49
chadmedit wouldnt surprise me tbh. firestorm and icestorm were pretty close, so were avalanche and blizzard
#asahi-dev · 2026-06-04 09:39
chadmedecore still faster than pcore?
#asahi-dev · 2026-06-04 09:11
chadmedwe might have used coremark instead of dhrystone
#asahi-dev · 2026-06-04 09:10
chadmedthe actual benchmark is kind of immaterial, youre supposed to fix the pcores to 1024 and normalise everything relative to that
#asahi-dev · 2026-06-04 09:09
chadmedistr we didnt use dhrystone for m1 or m2
#asahi-dev · 2026-06-04 07:22
chadmedim not entirely sure all of this was done at arm's behest
#asahi-dev · 2026-06-03 12:17
chadmedchaos_princess: if/when you rebase silence on 7.1 make sure to use the new macaudio and i2s bus keeper settings. the amps will ignore the current ones and you will get interesting results.
#asahi-dev · 2026-06-03 11:51
chadmedsimilarly the dcp driver has a bunch of insane^H^H cool macros and such to deal with the divergent fw abis
#asahi-dev · 2026-06-03 11:50
chadmedyeah, it just has to be reversed and actually expressed
#asahi-dev · 2026-06-03 11:22
chadmed"we moved the fb swap submit callback from d408 to d409 because ummm well ahhh you see mmmm"
#asahi-dev · 2026-06-03 11:21
chadmedthe gpu and display controller have not stayed the same across a single major os release, it is a bit annoying :P
#asahi-dev · 2026-06-03 10:45
chadmedis there anything else in the logs? it looks like it got stuck dealing with a power domain
#asahi-dev · 2026-06-02 20:54
chadmedshould be in the ADT
#asahi-dev · 2026-06-02 20:54
chadmedcheck to see if it has a different AID
#asahi-dev · 2026-06-02 20:51
chadmed(neo notwithstanding)
#asahi-dev · 2026-06-02 20:51
chadmedi suspect it's probably all the same as the m2 machines though
#asahi-dev · 2026-06-02 20:51
chadmedpossibly. we'll need to see if there's new TS parameters and new tunings required
#asahi-dev · 2026-06-02 20:49
chadmedgood news to wake up to :)
#asahi-dev · 2026-06-02 20:48
chadmedchaos_princess: did you need to make any material changes? nco, admac, mca all the same?
#asahi-dev · 2026-06-02 11:35
chadmedyou either need to save and restore the dart state when you reset avd (ew) or figure out how to reset avd without resetting dart or figure out why avd is faulting and fix that
#asahi-dev · 2026-06-02 11:34
chadmedthe driver is trying to talk to memory it thinks the dart has mapped, but the dart has been reset and has not mapped that address
#asahi-dev · 2026-06-02 11:31
chadmedclear m3 memory and reset it, reupload the fw and then kick it off again
#asahi-dev · 2026-06-02 11:28
chadmedwe couldnt even get upstream to accept a dummy fixed regulator :P
#asahi-dev · 2026-06-02 11:28
chadmedthanks i hate it
#asahi-dev · 2026-06-02 11:27
chadmedim surprised that reinitialising the cm3 does not clear all hw faults, iirc (its been three years i am likely incorrect) after the fw is uploaded and the m3 starts running it, everything behind it gets reset
#asahi-dev · 2026-06-02 11:25
chadmeddoing that is likely the issue here :P
#asahi-dev · 2026-06-02 11:20
chadmedhave you tried just resetting avd's cm3 and reinitialising rather than doing a full power cycle?
#asahi-dev · 2026-06-02 11:09
chadmedcan we just like, not do that?
#asahi-dev · 2026-06-02 11:08
chadmedare you resetting it via pmgr for PM ops or something else
#asahi-dev · 2026-06-02 11:01
chadmedhave you published any of this anywhere too? would be good to actually see, backseat debugging is kind of annoying
#asahi-dev · 2026-06-02 10:59
chadmedif the latter, have you expressed its dart in your devicetree?
#asahi-dev · 2026-06-02 10:59
chadmedin m1n1 or with a prototype linux driver?
#asahi-dev · 2026-05-31 10:46
chadmedthe read command is similarly weird, the readback from the display does not look very much like a ddc packet other than the data inside of it
#asahi-dev · 2026-05-31 10:45
chadmedand one of the formats does not have all of the data one would expect
#asahi-dev · 2026-05-31 10:45
chadmedi need to look at this more. this is definitely the thing we want but the write command has two different formats and there's nothing obvious that distinguishes which one should be used. they are different enough that there should be
#asahi-dev · 2026-05-30 11:50
chadmedvcp commands work, had hdr enabled before which was blocking them from doing anything
#asahi-dev · 2026-05-30 10:50
chadmedthe avservice ddc functions abstract away the low level transport which is nice, but idk about trying to shove that into a fake linux i2c device
#asahi-dev · 2026-05-30 10:49
chadmedover displayport too? im curious how that would look with dp aux packets
#asahi-dev · 2026-05-30 10:29
chadmedthe ddc api takes standard mccs/vcp commands and data but obviously the api looks nothing like an i2c device so thats going to be fun
#asahi-dev · 2026-05-30 03:03
chadmedreads work and are accurate, and the monitor acks writes but it just doesnt actually do anything. i know the monitor does actually support it though so maybe the tracer is throwing off timings
#asahi-dev · 2026-05-30 03:02
chadmedunfortunately.... it does not seem to work :P
#asahi-dev · 2026-05-30 03:02
chadmedfound ddc/ci, as suspected it's just a pair of commands in dcpavserv to read/write
#asahi-dev · 2026-05-24 01:45
chadmedlatest u-boot cannot init mtp
#asahi-dev · 2026-05-24 01:45
chadmedhm devicetrees from 7.0.9-2 seem to break mtp on t8112

plus ›