00:44 RSpliet: karolherbst: if you were referring to the box in the right-click menu for game -> Properties -> Set launch options... that doesn't work
00:44 karolherbst: RSpliet: well, it does for me
00:44 karolherbst: RSpliet: what exactly did you put in there?
00:45 gnarface: i thought you have to literally call bumblebee in there too
00:45 karolherbst: there _has_ to be a string like %command% in it. Otherwise everything you put in there gets added to the end of the invocation
00:45 RSpliet: ah ok
00:45 RSpliet: that explains
00:45 gnarface: RSpliet: note "%command%" literally, it's parsed by steam before the command gets sent to bash
00:46 karolherbst: so like for doing prime offloading I need to put in this: DRI_PRIME=1 %command%
00:46 gnarface: there's a couple others that are useful but they all work like you'd expect if they were coded by windows monkeys with 0 unix experience
00:48 gnarface: (finding explicit documentation on it is damned near impossible too)
00:52 RSpliet: yeah that's all better
00:52 RSpliet: cheers
00:53 gnarface: hooraay
01:05 RSpliet: Now at least I know neither portal nor dota 2 benefit from the insn scheduler I've been testing so far
01:05 RSpliet: Boo
01:06 gnarface: what type of framerates you getting?
01:07 gnarface: i'm just curious
01:08 RSpliet: On a GT940M
01:09 RSpliet: Portal 51.7fps, DOTA2 ~38fps
01:10 gnarface: ouch, compared to the proprietary drivers
01:10 gnarface: but still, i'm impressed you got it working
01:10 gnarface: that's totally playable
01:10 gnarface: did you try it from big picture mode too?
03:10 rhyskidd: any reviewers for https://github.com/envytools/envytools/pull/106 ?
03:11 rhyskidd: register name comes straight from the limited hw documentation
03:48 imirkin: mwk: did nv4/nv5 not do lines?
03:48 imirkin: (or points)
10:23 mwk: imirkin: yeah, no lines on NV4 or anything newer
10:23 mwk: you can only use lins
12:21 RSpliet: gnarface: this GPU is not to be compared to on-line benchmarks. Wikipedia does not acknowledge the existence of my 940M, considering it's a GM107 w/ DDR3
16:17 imirkin: mwk: what's this "lins" thing?
16:17 imirkin: i just need enough to draw transformed points and lines, preferably with color attrib support. dunno if other stuff can affect them.
16:19 imirkin: [as an aside, NV10+ have lines...]
16:59 mwk: imirkin: NV10 have 3D lines, which work according to different principles
17:00 mwk: what do you mean by color attrib support?
17:00 mwk: anyhow
17:00 imirkin: mwk: yeah, i mean i want GL_LINES
17:00 mwk: a "lin" is just a line when the first endpoint is included, the second is excluded
17:00 mwk: someone thought that funny, apparently
17:00 imirkin: and the coordinates are in raster space? and how are colors assigned?
17:01 mwk: it's a 2D primitive, unfortunately
17:01 mwk: it works like everything in the 2D engine
17:01 imirkin: i see
17:01 mwk: coordinates are integers in the pixel space
17:01 mwk: as for colors, you only get to choose a single solid color
17:01 imirkin: and i assume there's no color interpolation
17:01 imirkin: gotcha
17:01 mwk: subject to the usual 2D rops
17:01 imirkin: right.
17:02 imirkin: could i fake it with a tri with 010 or something?
17:02 mwk: no, that'd be a 0-area triangle
17:02 mwk: not particularly useful
17:02 imirkin: yeah =/
17:02 mwk: but if you do a bit of a calculation
17:02 imirkin: that'd just get culled
17:02 mwk: you can render a line as a quad
17:02 imirkin: yeah
17:02 imirkin: it's not exactly the same, but close enough
17:03 mwk: a) figure if it's a horizontal/vertical line, b) add -0.5 and +0.5 to x or y of both endpoints, to end up with 4 vertices
17:04 mwk: that should more or less be what GL prescribes anyway
17:04 imirkin: yeah, there's funny stuff with the end points
17:04 imirkin: with wide lines
17:04 mwk: but you won't get anywhere with 2D lins, I'm afraid
17:04 mwk: can't you just not support these?
17:04 imirkin: well, that's the current situation :)
17:05 imirkin: i noticed that lines weren't being rendered
17:05 imirkin: and figured i'd look at fixing it
17:05 mwk: I mean, don't support line width != 1
17:05 imirkin: oh
17:05 imirkin: mmmm... probably
17:05 imirkin: there's a max gl line width
17:08 imirkin: so basically i have to add 0.5 * viewport width/height to the endpoints
17:08 imirkin: er
17:08 imirkin: 0.5 / viewport width/height
17:09 mwk: just 0.5
17:09 imirkin: or are all those coords in raster space? i'll ahve to check
17:09 mwk: the coordinates you give to NV4 are in raster anyway
17:09 imirkin: PUSH_DATAp(push, _tnl_get_vertex(ctx, i), vertex_size);
17:09 imirkin: will have to see what that returns
17:10 imirkin: of course.
17:10 imirkin: return vtx->vertex_buf + nr * vtx->vertex_size;
17:10 imirkin: that should have been obvious.
17:10 imirkin: grrr
17:11 imirkin: yeah i think you're right - it's all raster positions
18:08 telliEATSbig_dix: Hell0
20:39 ni291187: bonsoir
23:48 nullspoon: Quick question. Does anyone here have muxless hybrid working with the newer nvidia cards using nouveau? (notably the 1050 ti)
23:51 nullspoon: My system doesn't quite match the steps documented on the optimus page. When I execute xrandr --listproviders, I see my intel, then nothing. When I modprobe nouveau, I see the second provider, but it's name is 'modesetting', not 'nouveau'.
23:54 imirkin: nullspoon: shouldn't matter
23:55 imirkin: some people have had trouble getting the graph unit in GP107's to come to life in some laptops, but that doesn't sound like it's your issue, at least not based on what you've said so far
23:56 nullspoon: Here's another symptom if it helps. I can run lspci and I see my intel as the VGA, and my nvidia as the 3d controller.
23:56 nullspoon: If I run lspci after modprobing nouveau though, my system freezes.
23:57 nullspoon: I'm a little worried I've got bad hardware.
23:57 imirkin: try modprobing nouveau with runpm=0
23:58 imirkin: so that it doesn't auto-suspend the gpu
23:58 imirkin: running lspci causes a wake-up
23:58 nullspoon: ah, thanks. Was googling there. Never see that before. :)
23:59 imirkin: here's a more complete list of params: https://nouveau.freedesktop.org/wiki/KernelModuleParameters/
23:59 imirkin: not 100% accurate, but what is
23:59 nullspoon: True. Well, so far, you're giving me much more than I was able to find in my research.