02:15 orbea: interesting...if I 'find . -name ".git*" | xargs rm -rf' in the envytools src directory, cmake configure breaks....
02:16 orbea: its just .git and several .gitignore ....
07:25 MarkedOne: Hello... I have problem with running bumblebee together with nouveau. I am getting this error. [ERROR]Cannot access secondary GPU - error: [XORG] (EE) NOUVEAU(0): [drm] failed to set drm interface version.
07:26 MarkedOne: I cant find solution myself.. I am googling for 2 days now.. Please help :)
07:44 karolherbst_work: don't use bumeblebee
07:44 karolherbst_work: that's your solution
07:44 karolherbst_work: MarkedOne: https://nouveau.freedesktop.org/wiki/Optimus/
07:45 karolherbst_work: also bumblebee has high pcie overhead and you get worse performance
07:48 MarkedOne: karolherbst_work: Thank you. I am will read that article right now. Just quick question. Am I able to use my dedicated graphic card for rendering in blender without bumblebee?
08:30 karolherbst_work: MarkedOne: if you use nouveau, yes. But performance might be not that good depending ont he gpu
08:33 MarkedOne: ehm... i am little bit lost :D
08:49 MarkedOne: Ehm... I am studying about Optimus and DRI is mentioned... when I `cat /var/log/Xorg.0.log | grep DRI` I have found DRI2... DRI3 is newer... am i able to activate/install it?
08:58 nailyk: Hi. Sorry to annoy you one more time but i'm confused. I finnaly rebuild my kernel and nouveau (out of kernel tree).
08:59 nailyk: Each time i load nouveau dmesg report this: [drm] Initialized nouveau 1.3.1 20120801 for 0000:0a:00.0 on minor 0.
09:00 nailyk: (this time i load it with insmod) How can y check the version of nouveau loaded into the kernel? (build from linux-4.6 branch)
10:49 nailyk: i guess a found a solution: i change the version in nouveau_drm.h and give another try.
11:35 pmoreau: MarkedOne: You can force DRI3 by adding the `Option "DRI" "3"` in a X config file, but AFAICT, Nouveau should be using DRI3 by default. Probably the version of the Intel DDX that you are using is still using DRI2 by default.
11:37 pmoreau: nailyk: Right, or add a custom printk in `nouveau_drm_load()`, as you prefer.
11:38 Tom^: pmoreau: https://gist.github.com/gulafaran/f4749bf0ebabec9d25f64d1c0a8bd6d2
11:38 Tom^: idk
11:39 pmoreau: Weird
11:39 Tom^: i wonder what happends if i force it hm
11:40 Tom^: unimaginable performance increases is what im taking for granted!
11:40 pmoreau: Indeed, I do get the same messages…
11:40 pmoreau: Thanks to your beast of a card :-D
11:41 Tom^: =D
11:41 pmoreau: From the nouveau manpage: "Under glamor acceleration DRI3 is always enabled if supported, as glamor currently does not support DRI2. Default: 2 on exa, 3 on glamor."
11:42 Tom^: well isnt glamor dropped from the nouveau ddx?
11:42 Tom^: in favor of modesetting i guess
11:43 MarkedOne: pmoreau: DRI3 is probably not for my kernel 3.16
11:44 Tom^: MarkedOne: you really should get a kernel from this decade before trying to figure out issues in here.
11:44 karolherbst_work: :D
11:44 karolherbst_work: 3.16 sounds like debian 6 or something
11:45 pmoreau: Tom^ https://cgit.freedesktop.org/nouveau/xf86-video-nouveau/commit/?id=1f8b2b348526d94e9bde4a285f331a4934c11078 ;-)
11:45 Tom^: pmoreau: see im not senile yet! :)
11:45 pmoreau: Never claimed that! :-D
11:46 pmoreau: Time to restart X
11:46 MarkedOne: Tom^: I am running Debian Jessie and it is default :D
11:46 MarkedOne: how old is it? :D
11:47 Tom^: MarkedOne: yea unless you get atleast something like 4.5 or 4.6 the devs in here wont be that useful to you, issues you are seeing could just aswell been fixed 2 years ago.
11:47 Tom^: MarkedOne: and they dont remember them all :p
11:47 Tom^: MarkedOne: 3.16 is from 2012 if my googling is correct
11:48 MarkedOne: oh hell.. really? And everybody is telling me.. that there is no need to update kernel
11:49 MarkedOne: okey thank you btw :D
11:50 Tom^: MarkedOne: you probably want to update your mesa aswell, incase thats very old too.
11:50 karolherbst_work: :D
11:50 karolherbst_work: random debian user: "there is no need to update _random_package_name_here" :p
11:50 urjaman: well, if it works, you dont need to touch it ;P
11:50 karolherbst_work: sure
11:50 Tom^: unless its moldy, it aint stable!
11:51 karolherbst_work: and under the hood, you have spyware all over your machine
11:51 MarkedOne: oh... i wish for command that install all updates for debian >D
11:52 urjaman: but yeah I like arch, i have 4.6.4 here without any messing about it
11:52 karolherbst_work: MarkedOne: you mean besides apt?
11:52 pmoreau: Interesting: the Xorg.log only talks about DRI2, but running `LIBGL_DEBUG=verbose | grep libgl` says I am using DRI3… o.O
11:52 karolherbst_work: pmoreau: maybe you should check when your xorg.log was created
11:52 urjaman: yeah i dont think Xorg.log is ever really indicative of everything thats going on :P
11:53 MarkedOne: debian has something like stable.. that contain all programs in specific version when shipped... only security updates can pass
11:53 pmoreau: 5 mins ago, and it correctly reference modesetting and that nouveau was removed
11:56 urjaman: anyways if one wants newer stuff from debian, it's testing or unstable , but i have no experience with running either ...
11:58 Tom^: pmoreau: https://gist.github.com/gulafaran/65392999492adbacf1773a16600b400f forcing ftw :p now to find if things break.
11:58 karolherbst_work: urjaman: it is like stable, just faster and more stable
11:58 karolherbst_work: :D
11:59 urjaman: yeah i guessed :P
11:59 pmoreau: Tom^: :-D
12:00 karolherbst_work: Tom^: well, you might have to reclock for proper 2d accell now :p
12:00 Tom^: karolherbst_work: i do at start of x anyways, nouveauctl -p0a & :D
12:01 karolherbst_work: like buntu, which just switch to modesetting ddx for intel gpus, because it hadn't a release in years and they are too scared pushing stuff around, which hasn't a version number like 3.0.0
12:01 karolherbst_work: Tom^: 0a is for the weak
12:01 Tom^: i dont wanna ruin my wallet
12:01 karolherbst_work: :D
12:01 karolherbst_work: your own fault
12:02 Tom^: so mod4 + f12 for -p0f and -b2 , mod4 + f11 to dropdown back to 0a
12:02 karolherbst_work: but maybe your LEDs just draw too much power, so you have to keep it low with your actual gpu
12:02 karolherbst_work: but maybe you should just buy a bigger psu
12:02 karolherbst_work: if you are so scared of breaking it
12:05 Tom^: nothing broke, and i did get a tiny gain in points in unigine heaven. but its so tiny it might aswell just been a mesa update
12:05 Tom^: oh well
12:06 Tom^: http://i.imgur.com/lM49rMO.png
12:11 karolherbst_work: how bad
12:11 Tom^: atleast the placebo effect is making it better
12:29 karolherbst_work: Tom^: it sure is like 20% slower now :p
12:29 Tom^: lies
12:30 karolherbst_work: I guess the tearing is also rather bad
12:31 Tom^: funnily enough i think im seeing less tearing
12:31 Tom^: mpv doesnt seem to tear even tho compton isnt running atm
12:46 dcomp: 12:45 < pmoreau> Tom^ https://cgit.freedesktop.org/nouveau/xf86-video-nouveau/commit/?id=1f8b2b348526d94e9bde4a285f331a4934c11078 ;-)
12:47 Tom^: ?
12:48 dcomp: Sorry CoK (cat on keyboard). I think she's hungry
12:53 pmoreau: :-)
13:16 karolherbst_work: dcomp: if your car really did that, I would be worried, because their computer skills doesn't seem to be that random
13:17 urjaman: car? :D
13:17 Tom^: karolherbst_work: v8 powered cat?
13:19 karolherbst_work: mhh
13:19 karolherbst_work: I mean cat
13:19 karolherbst_work: :D
15:04 kloofy: http://developer.amd.com/wordpress/media/2013/07/AMD_Sea_Islands_Instruction_Set_Architecture.pdf page 23 describes similar behavior as in the NVIDIA patent for amd hw too, totally different terms, and me still does not know how to use this feature
15:04 kloofy: but this one should be the correct method that they describe on page23
17:51 kloofy: cumbersome, basically i want to do so, that cache has address 1232, where i provide 123 only cause it's the only address which matches tag 123 it returns me 2, it seems that it does not work so, and i am just a loon again
21:09 t-ask: hi, how could I determine the current GPU clockrate?
21:14 Tom^: t-ask: i guess one way is checking the pstate file, if there is one. cat /sys/kernel/debug/dri/0/pstate as root , assuming you have kernel 4.5 or newer.
21:15 t-ask: Tom^: nice that was it. I couldn't remeber the path :)
21:16 t-ask: I assume with the latest kernel mod of Karol it should govern itself, or?
21:17 Tom^: the stable_reclocking_kepler_v5 branch? no.
21:17 t-ask: I think I compiled the newer one
21:19 t-ask: stable_reclocking_kepler_v5_4.6
21:20 Tom^: thats just for kernel 4.6
21:20 Tom^: but no that still requires you setting it yourself in the pstate and boost file.
21:20 Tom^: ive keybinded 0a to mod4 + F11 and 0f to mod4 + F12. so its not that much of a hassle imo
21:27 t-ask: so the non 4.6 one is for 4.7 already?
21:28 Tom^: yep
21:28 t-ask: I'm just looking at the git tree, ok, seems like it branched from master_4.7
21:30 t-ask: the pstate 0d and 0f seem to be the same
21:30 appl: dx11 already supported?
21:30 Tom^: t-ask: yea i have that aswell
21:30 Tom^: appl: eh what?
21:30 appl: in the driver
21:31 Tom^: appl: no dx level is supported in the driver
21:31 appl: bad
21:31 appl: d3d11
21:33 t-ask: I wonder that my i5 CPU runs only with around 130% with Guild Wars 2
21:33 t-ask: usually GW2 is very CPU heavy, well maybe they change something
22:13 t-ask: ok, it seems to be normal, considering that Windows users tell me they have 40%