alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - Logs https://freenode.irclog.whitequark.org/panfrost - <daniels> avoiding X is a huge feature
kaspter has quit [Remote host closed the connection]
kaspter has joined #panfrost
kaspter has quit [Ping timeout: 246 seconds]
kaspter has joined #panfrost
stikonas has quit [Ping timeout: 252 seconds]
kaspter has quit [Ping timeout: 265 seconds]
kaspter has joined #panfrost
camus has joined #panfrost
megi has joined #panfrost
kaspter has quit [Ping timeout: 265 seconds]
camus is now known as kaspter
vstehle has quit [Ping timeout: 265 seconds]
megi has quit [Ping timeout: 276 seconds]
chewitt has quit [Ping timeout: 265 seconds]
kaspter has quit [Ping timeout: 265 seconds]
kaspter has joined #panfrost
camus has joined #panfrost
kaspter has quit [Ping timeout: 268 seconds]
camus is now known as kaspter
kaspter has quit [Remote host closed the connection]
kaspter has joined #panfrost
davidlt has joined #panfrost
rhyskidd_ has joined #panfrost
rhyskidd has quit [Ping timeout: 268 seconds]
rhyskidd_ is now known as rhyskidd
kaspter has quit [Remote host closed the connection]
kaspter has joined #panfrost
vstehle has joined #panfrost
chewitt has joined #panfrost
vstehle has quit [Ping timeout: 240 seconds]
vstehle has joined #panfrost
fysa has joined #panfrost
NeuroScr has quit [Quit: NeuroScr]
fysa has quit [Ping timeout: 276 seconds]
davidlt has quit [Ping timeout: 268 seconds]
guillaume_g has joined #panfrost
yann has quit [Ping timeout: 240 seconds]
pH5 has joined #panfrost
yann has joined #panfrost
stikonas has joined #panfrost
karolherbst has quit [Ping timeout: 276 seconds]
karolherbst has joined #panfrost
fysa has joined #panfrost
raster has joined #panfrost
davidlt has joined #panfrost
fysa has quit [Ping timeout: 276 seconds]
jernej has quit [Remote host closed the connection]
jernej has joined #panfrost
eballetbo[m] has quit [*.net *.split]
eballetbo[m] has joined #panfrost
phh has quit [*.net *.split]
phh has joined #panfrost
robmur01 has joined #panfrost
megi has joined #panfrost
cwabbott has quit [Quit: cwabbott]
cwabbott has joined #panfrost
cwabbott has quit [Read error: Connection reset by peer]
cwabbott has joined #panfrost
<tomeu> alyssa: is this expected?
<tomeu> *** panfrost_choose_tile_size: 0 exp_w 0 exp_h 0 best_w 16 best_h 16 width 32 height 32
<tomeu> I'm getting a hierarchy_mask of 0x0 most of the time
<tomeu> btw:
<tomeu> - best_h = MAX2(best_w, util_next_power_of_two(height / 63));
<tomeu> + best_h = MAX2(best_h, util_next_power_of_two(height / 63));
fysa has joined #panfrost
<tomeu> alyssa: lvrp16: narmstrong: chewitt: with those changes, even t820 is all green \o/
<narmstrong> \o/
<narmstrong> Huge news tomeu !
<tomeu> alyssa: though in a rerun, I got some failures on t720: https://gitlab.freedesktop.org/tomeu/mesa/-/jobs/1048057
<tomeu> with no code changes
<tomeu> I see some sched timeouts, wonder if the t720 is so slow that some jobs reach the kernel's limit
<tomeu> then they cause a GPU reset, making simultaneous tests to fail
<tomeu> narmstrong: big surprise indeed :)
<tomeu> robher: what do you think? I guess in real life it's good to have a reset and then a retry, but for running deqp concurrently, it's a problem
fysa has quit [Ping timeout: 276 seconds]
<tomeu> alyssa: that may be another reason to undust --rerun :/
<tomeu> with so few failures, that shouldn't affect much total run time
<robmur01> do you have OPPs for H6?
<tomeu> 2019-12-02T15:50:52 /deqp/lava-deqp-runner.sh: 17: cannot create /sys/devices/platform/*.gpu/devfreq/devfreq0/governor: Directory nonexistent
<tomeu> I think the OPPs weren't merged in mainline
<robmur01> the default GPU clock may be something utterly rubbish (e.g. that's the case on RK3328, where lima doesn't even have devfreq at all so it stays there)
<robher> tomeu: maybe a module param or something to set the timeout?
<tomeu> robmur01: ah, good thing to check
<tomeu> robher: ah, I like that
<tomeu> 2019-12-02T16:06:10 [ 3.319613] panfrost 1800000.gpu: clock rate = 432000000
<tomeu> 2019-12-02T16:06:10 [ 3.324951] panfrost 1800000.gpu: bus_clock rate = 100000000
<tomeu> on s912:
<robmur01> hmm, that's a lot better than RK3328's 150MHz, at least :/
<tomeu> 2019-12-02T15:50:28 [ 2.606308] panfrost d00c0000.gpu: clock rate = 666666666
<tomeu> wow
<tomeu> ah, the rk3399 has only 500MHz, and it's really fast to run the deqp tests
<tomeu> lots of faults in a subsequent run, again without code changes: https://gitlab.freedesktop.org/tomeu/mesa/-/jobs/1048222
<robher> tomeu: But how do parallel jobs really change things? The timeout should start when the job is given to the h/w. So it's only a question of 1 to 3 job slots...
<tomeu> maybe related to gpu resets at the wrong time
<tomeu> running 8 deqp instances in parallel must be a good way to stress test :)
<tomeu> robher: if we have another job queued in the hw, what happens to it?
<robmur01> FWIW, I can tell you T860MP1 at 40MHz is still considerably faster than T720MP1 at 40MHz ;)
<robmur01> the latter can't even manage 1fps on most of glmark :D
<robher> tomeu: we only have the *_NEXT registers, but I don't think we use those yet.
<tomeu> ah, it's true, we didn't merge them because they made mesa/panfrost slower
<tomeu> ok, that's good info
<tomeu> which explains why the latest faults only failed one test, and not 8
<alyssa> tomeu: Typo, whoops.
<alyssa> tomeu: I don't understand the 0xFF specialness.
<alyssa> tomeu: We should definitely revisit the _NEXT registers
<alyssa> IIRC there's some stuff related to bbrezillon's job pipelining series that needs to be finished up and then we should get a good boost from _NEXT
davidlt has quit [Ping timeout: 246 seconds]
pH5 has quit [Quit: bye]
<chewitt> tomeu: alyssa: t820 all green is awesome news! .. things have come a long way since https://imgur.com/a/HwJ0bwP (12 December 2018)
<chewitt> next stop bifrost!?
karolherbst has quit [Ping timeout: 265 seconds]
yann has quit [Ping timeout: 276 seconds]
chewitt has quit [Quit: Zzz..]
<bbrezillon> alyssa: I don't remember what was missing
<bbrezillon> didn't I address all the job inter-deps issues?
pH5 has joined #panfrost
karolherbst has joined #panfrost
davidlt has joined #panfrost
raster has quit [Quit: Gettin' stinky!]
yann has joined #panfrost
adjtm_ has quit [*.net *.split]
_whitelogger has quit [*.net *.split]
warpme_ has quit [*.net *.split]
narmstrong has quit [*.net *.split]
bshah has quit [*.net *.split]
vstehle has quit [Ping timeout: 246 seconds]