01:01AndrewR: imirkin, hi! are you around? Just wanted to say I tested pmpeg stuff a bit more (on pcie nv92, ~latest mesa and kernels 4.2 and 4.12-rc4) and ...it only works relatively good with 1 cpu online. If I add more - it hangs shortly after playing few seconds of video (I mean cpu online/offline via sysfs of nr_cpus boot parameter). Setting irq affinity doesn't help, disabling MSI doesn't help. But sadly even with 1 cpu it hangs on bigger streams , like 1080i at
01:01AndrewR: 23 mbit / second.. Not sure if any of it useful for you ...
01:03imirkin: fwiw i've never seen pmpeg hang
01:03imirkin: i've seen it not be able to keep up, but that's a separate matter
01:04imirkin: AndrewR: on the bright side, your nv43 should be back to working
01:04imirkin: AndrewR: and OSD should work now =]
01:07AndrewR: imirkin, yeah, but I haven't pushed my working machine aside from desk edge for making room for older machine yet. Will do this ...hopefully soon (I had record uptime ~18 days and broke it by trying to find out if Unigine Sanctuary still hangs at high. It hangs ....)
01:07imirkin: not a lot of improvements being done for the nv50 stuff =/
01:08AndrewR: err, not Sanctuary, heaven ... (more demanding). There was fix for gallium few days ago..so I had hope... anyway this is not most used app on this comp :P
01:08imirkin: let's hope :)
01:17AndrewR: http://www.w6rz.net/parkrun1920_23mbps.ts - i used this file for some testing. May be it makes difference (while it shouldn't - as long as it displayed correctly?)
01:19AndrewR: NOUVEAU_PMPEG=1 taskset -c 0 mplayer ~/botva/vid/dreamtime.mpg -vo xvmc -quiet -lavdopts threads=1 (as you can see I tried to pin mplayer to one core, and disabled lavc-level multithreading)
01:19gnarface: wait though
01:19gnarface: doesn't that need to be -vo vdpau?
01:19imirkin: he knows what he's doing.
01:20gnarface: wait then
01:20gnarface: should i be using -vo xvmc instead?
01:22imirkin: what gpu are you on again?
01:23gnarface: i've got a few here
01:23imirkin: ok, well, it may make sense to use xvmc on pre-G98 chips
01:23imirkin: it obviously only does mpeg2
01:24imirkin: (and mpeg1)
01:24gnarface: oh, well i do have that G92
01:24gnarface: and a GK106
01:25gnarface: those are just the two that i have plugged into machines right now
01:26gnarface: i'll have to try it on the G92
01:28gnarface: i'm thinking it's primary performance issue is due to the lack of reclocking though
01:34rhyskidd: any particular quirks I should be aware of getting an mmiotrace on a PRIME setup?
01:34rhyskidd: obviously focused upon the secondary GPU that doesn't have the display directly connected to it
01:35imirkin: ask karol... he's done it a bunch
01:35rhyskidd: ccaione: also, pinging whether you'd been able to get your strap_peek from the GP108
01:59imirkin: skeggsb_: what do you make of https://hastebin.com/moxoyizoju.go ?
01:59imirkin: [when trying to use vdpau on GK208... with DRI3]
07:35pmoreau: imirkin: Not sure whether you still need that information, but here it is anyway: "0002245c: 00000002", "00022448: 00000004", "00612004: 30700f0f" on GP106
07:51ccaione: rhyskidd: what do you need exactly?
10:48ioria: hi, ubuntu trusty (xenial hwe) with new 4.4.0-83-generic kernel, cannot boot with nouveau ( i had to install nvidia 304) ... any help ?
10:54RSpliet: ioria: nope, no help. For that you'd have to go and ask the Ubuntu folks. We're here for upstream (eg. kernel 4.11/4.12, mesa 17.1...) only
10:54ioria: RSpliet, thanx
12:08imirkin: pmoreau: thanks, so that 2245c was a red herring anyways
12:08pmoreau: If you need more peeking, feel free to ask.
12:13imirkin: pmoreau: no, i think we've got it covered. thanks!
12:13pmoreau: Ok, cool
12:14imirkin: actually if you have a nvd9 plugged in perchance, a 612004 peek would be neat
12:14imirkin: hakzsam: you have one of those, no?
12:14hakzsam: but not plugged
12:15imirkin: ah ok. no worries.
12:15imirkin: pretty sure it'll work out ok :)
12:15pmoreau: I am not sure I have an NVD9, and even if I did, it is several thousand kilometers away from me at the moment.
14:25RSpliet: skeggsb_: are you aware of https://bugzilla.redhat.com/show_bug.cgi?id=1457626 ?
14:25imirkin_: since when do you dig around in RH BZ?
14:27RSpliet: imirkin_: normally I don't. I went over the Fedora 26 go/no-go meeting minutes because I was curious why they slipped their release deadline 5 times. Came across that gem that's been around for a month but without attention of anyone relevant
14:27RSpliet: What I failed to check was the "assigned to" - apparently it's Ben. Surely he's aware of it then!
14:28imirkin_: along with 30000 other bugs, i'm sure =/
20:30karolherbst: mhh, I am getting "GL_OUT_OF_MEMORY in glGetTexImage" while the state's being dumped