#asahi-dev · 2023-04-30 23:08
eiln_ChaosPrincess: ty. looking at the error.#asahi-dev · 2023-04-30 23:05
eiln_sorry shoudlve mentioned#asahi-dev · 2023-04-30 23:05
eiln_curl -LJO https://www.dropbox.com/s/lpjap6w0kdlom1h/add.hwx?dl=0 #asahi-dev · 2023-04-30 22:48
eiln_shot in the dark but does any t6000 mind running experiments/ane.py on ane-t6000 branch https://github.com/eiln/m1n1/tree/ane-t6000#asahi-dev · 2023-04-30 07:45
eiln_terrible day for one off errors, i'm was counting on my fingers lol#asahi-dev · 2023-04-30 07:44
eiln_both have the same pds, just in diff locations#asahi-dev · 2023-04-30 07:44
eiln_t6000 has ane_sys_cpu elsewhere, and there's 6 too, so it goes 0xc000 - 0xc028#asahi-dev · 2023-04-30 07:44
eiln_ane_sys_cpu is 0xc000 for t8103, and there's 6 more after to make it 0xc000 - 0xc030#asahi-dev · 2023-04-30 07:44
eiln_oh wait i figured it out#asahi-dev · 2023-04-30 07:29
eiln_so the kext turns on 5 to unlock the mmio region, then the fw turns on the last one... why are they playing games with power domains of all things#asahi-dev · 2023-04-30 07:27
eiln_jannau: received, thank you so much. it looks like it's 6.#asahi-dev · 2023-04-30 07:16
eiln_can you see if the 8th one turns on too?#asahi-dev · 2023-04-30 07:16
eiln_jannau: yeah, the engine region seems to be locked behind 5 then. weird the last two turn on though, whose domain is that?#asahi-dev · 2023-04-30 06:43
eiln_jannau: does it work with 4?#asahi-dev · 2023-04-29 20:58
eiln_and attach the mon.poll() results if some PD_COUNT val works#asahi-dev · 2023-04-29 20:58
eiln_in https://github.com/eiln/m1n1/tree/ane-power branch#asahi-dev · 2023-04-29 20:58
eiln_jannau: do you mind running experiments/ane_t6002_power.py for PD_COUNT 5, 6, or 7#asahi-dev · 2023-04-29 20:58
eiln_but the kext only turns on 5? is the fw doing the last one/two??#asahi-dev · 2023-04-29 20:58
eiln_id guess 7 bc t8103 has 7 pds & everything looks the same#asahi-dev · 2023-04-29 20:58
eiln_given rPMGR_ANE2_SET4_PS there should be at least 6 pds, but likely 7#asahi-dev · 2023-04-29 20:57
eiln_the kext & fw flip flop with the pds for some bizarre reason#asahi-dev · 2023-04-29 08:04
eiln_they can't utilize ane1 to 2x one job bc they have separate L2s. nothing suggests the two can communicate actually. so at most it 2x two jobs simultaenously#asahi-dev · 2023-04-29 08:01
eiln_ig stacking two ane's didn't go as planned#asahi-dev · 2023-04-29 08:00
eiln_its bc of mmio base address lol#asahi-dev · 2023-04-29 07:59
eiln_ane/ane0 fw is rlly similar too, i highly doubt ane0/ane1 had mem sharing features#asahi-dev · 2023-04-29 07:58
eiln_they share a unified -h13 flag#asahi-dev · 2023-04-29 07:58
eiln_interestingly t8103 ane & t6000 ane compile the same microseq#asahi-dev · 2023-04-29 07:51
eiln_ane0/ane1 have near identical fw, maybe overheating?#asahi-dev · 2023-04-29 07:49
eiln_m2 max shoulve had ane1 by trends, but it only has ane0#asahi-dev · 2023-04-29 07:49
eiln_jannau: the nodes are in driver strings & iboot loads the firmware. some last minute mishap ig#asahi-dev · 2023-04-29 07:39
eiln_jannau: received, thank you!#asahi-dev · 2023-04-29 07:38
eiln_nvm it's fine, just need data on ane2 for now#asahi-dev · 2023-04-29 07:36
eiln_jannau: really? it should be in the dt#asahi-dev · 2023-04-29 07:36
eiln_jannau: start/shutdown is enough for now#asahi-dev · 2023-04-29 06:31
eiln_dottedmag: I had no idea, ty