| tjaalton | could someone explain why increase in DPI makes the fonts larger? | 06:53 |
|---|---|---|
| tjaalton | when the fontsize isn't changed | 06:54 |
| tjaalton | I had to change the size to 9 on all my setups (laptop, 24" monitor, 30" monitor) | 06:55 |
| tjaalton | AIUI the fonts should appear _smaller_ when DPI increases | 06:56 |
| wgrant | tjaalton: GTK sees the DPI increase, which means something the same pixel-size will appear smaller on the monitor. | 07:13 |
| wgrant | Thus it compensates, increasing the pixel-size of the font, for the same pt size. | 07:13 |
| superm1 | so is something wrong with it's algorithm for compensation then? | 07:13 |
| wgrant | I don't think so. | 07:14 |
| wgrant | I think we're all just so used to the DPI being forced. | 07:14 |
| wgrant | And GNOME fonts are in reality really big by default. | 07:14 |
| tjaalton | sigh.. | 07:16 |
| hyperair | hi. i'm curious about the current status of the intel driver in jaunty. does anybody know? | 10:09 |
| hyperair | i just tested a livecd i downloaded two days ago and it seems that there's still the performance regression | 10:09 |
| tjaalton | what regression? | 10:11 |
| ion_ | *Of course* the fonts’ pixel size has to be bigger if the resolution of the screen is higher. | 10:11 |
| seb128 | tjaalton: has the vblank issue been fixed in jaunty? | 10:11 |
| hyperair | tjaalton: from 60fps to 30fps in a game | 10:11 |
| tjaalton | seb128: yes | 10:11 |
| ion_ | Also, if the DPI is calculated wrong, add a DisplaySize line to xorg.conf | 10:11 |
| tjaalton | seb128: by not using vblank | 10:12 |
| seb128 | tjaalton: ok, I should drop the drirc workaround then thanks ;-) | 10:12 |
| tjaalton | hyperair: what chip? | 10:12 |
| hyperair | tjaalton: it's still playable, just that everything moves extra-slowly | 10:12 |
| seb128 | tjaalton: I was just wondering if that was the issue some users were complaining about but seems it's not | 10:12 |
| hyperair | tjaalton: i965 | 10:12 |
| tjaalton | hyperair: what game, and compared to what version? | 10:12 |
| tjaalton | I've got a i965 and am pleased with it | 10:13 |
| hyperair | tjaalton: touhou 07 -- perfect cherry blossom. comparing intrepid fully updated to jaunty | 10:13 |
| hyperair | yeah if you don't play games you don't notice anything | 10:13 |
| tjaalton | right | 10:13 |
| tjaalton | I don't :) | 10:13 |
| hyperair | yep | 10:13 |
| hyperair | see | 10:13 |
| tjaalton | except on my ps3 | 10:13 |
| hyperair | hahah | 10:13 |
| hyperair | for some very strange reason, performance with compiz on/off is the same | 10:14 |
| hyperair | =\ | 10:14 |
| hyperair | i tried running metacity, making sure the compositer is off and everything | 10:14 |
| seb128 | who plays games under linux anyway | 10:14 |
| tjaalton | seb128: I'm not sure if there are regressions jaunty->jaunty, those are what interest the most atm | 10:14 |
| hyperair | me damnit! | 10:14 |
| hyperair | tjaalton: uxa doesn't exactly seem to help either | 10:14 |
| hyperair | tjaalton: i noticed that DRI2 is disabled with EXA though | 10:14 |
| tjaalton | ok | 10:14 |
| tjaalton | DRI2 requires UXA | 10:15 |
| tjaalton | on intel | 10:15 |
| hyperair | but it was running with EXA on the earlier jaunty images =\ | 10:15 |
| hyperair | UXA wouldn't even start | 10:15 |
| tjaalton | it still is | 10:15 |
| hyperair | wait what? | 10:16 |
| hyperair | DRI2 doesn't run when EXA is running right? | 10:16 |
| tjaalton | exa is used by default, not uxa | 10:16 |
| hyperair | that's the current situation | 10:16 |
| hyperair | correcct? | 10:16 |
| tjaalton | uh | 10:16 |
| tjaalton | dri2 isn't used | 10:16 |
| tjaalton | because uxa isn't used | 10:16 |
| hyperair | except when UXA is used | 10:16 |
| hyperair | yeah | 10:16 |
| hyperair | but i stared at the logs in one of the earlier jaunty builds | 10:16 |
| hyperair | it said DRI2 was up | 10:16 |
| hyperair | and EXA was up | 10:16 |
| tjaalton | can't be | 10:17 |
| hyperair | nope, i'm very sure about this | 10:17 |
| hyperair | i lost the image though damnit | 10:17 |
| hyperair | because i turned on wobbly windows | 10:17 |
| hyperair | and moved glxgears around | 10:17 |
| hyperair | and LIBGL_ALWAYS_INDIRECT was not set | 10:17 |
| tjaalton | (WW) intel(0): DRI2 requires UXA | 10:17 |
| hyperair | yaeh now it says that | 10:17 |
| hyperair | earlier it didn't | 10:18 |
| hyperair | when i turned on UXA via xorg.conf, it gave me artifacts and wouldn't even start | 10:19 |
| tjaalton | well, uninteresting at this point | 10:20 |
| hyperair | yeah | 10:20 |
| tjaalton | if it used dri2 or not | 10:20 |
| hyperair | but right now there's still the whole slowdown issue | 10:20 |
| tjaalton | I'd say the chances to fix it for jaunty are nil | 10:21 |
| hyperair | sigh | 10:21 |
| hyperair | that's pretty sad =\ | 10:21 |
| hyperair | tjaalton: this is rather interesting. i'm now in a jaunty liveusb | 10:37 |
| hyperair | ubuntu@ubuntu:~$ sudo chmod 0666 /dev/dri/card0 && glxgears | 10:38 |
| hyperair | 4656 frames in 5.0 seconds = 931.029 FPS | 10:38 |
| tjaalton | so? | 10:38 |
| hyperair | otherwise it's 100+ | 10:38 |
| hyperair | which means there's an issue regarding stuff accessing dri isn't it? | 10:38 |
| * hyperair pokes tjaalton | 10:39 | |
| tjaalton | I'm awake | 10:39 |
| hyperair | hmm lemme try that game again | 10:39 |
| tjaalton | is the livecd used in video group? | 10:39 |
| tjaalton | user | 10:39 |
| hyperair | nope | 10:40 |
| hyperair | it isn't | 10:40 |
| tjaalton | there you go | 10:40 |
| hyperair | a | 10:40 |
| hyperair | h | 10:40 |
| hyperair | but why isn't it? | 10:40 |
| hyperair | that'll be a bug right? | 10:40 |
| tjaalton | dunno | 10:40 |
| hyperair | hmm | 10:40 |
| hyperair | meh. looks like only glxgears showed any improvement | 11:01 |
| Ng | hyperair: so you're on 965 and only seeing slowdown in heavy 3d apps? | 11:10 |
| hyperair | Ng: yep | 11:10 |
| Ng | hyperair: but not compiz? | 11:10 |
| hyperair | Ng: that's right | 11:10 |
| hyperair | Ng: compiz looks fine | 11:10 |
| Ng | this is all very weird, how can tjaalton have no problems, you only have slowdowns with heavy 3d stuff and my CPU only has to think about doing something and compiz chugs like crazy | 11:11 |
| hyperair | Ng: very weird indeed. | 11:11 |
| hyperair | Ng: well if compiz chugs when your cpu thinks of doing something, then you're probably not using hardware accel? | 11:12 |
| Ng | hyperair: I'm pretty sure I am. every way I know to test for that tells me I am | 11:12 |
| hyperair | glxinfo -l | grep render? | 11:12 |
| Ng | direct rendering: Yes | 11:12 |
| Ng | OpenGL renderer string: Mesa DRI Intel(R) 965GM GEM 20090114 x86/MMX/SSE2 | 11:12 |
| hyperair | hm strange you've got the x86/MMX/SSE2 string | 11:13 |
| hyperair | i guess it doesn't show up on x86_64 | 11:13 |
| hyperair | are you in the video group? | 11:13 |
| Ng | hyperair: yup | 11:15 |
| hyperair | Ng: then i have no idea heheh | 11:15 |
| hyperair | i think i'll be sitting in intrepid until a suitably fast intel driver comes along | 11:16 |
| Ng | good luck with that | 11:16 |
| Ng | I reckon we're in for another year of intel pain | 11:16 |
| hyperair | ouch | 11:16 |
| hyperair | that'll hurt =( | 11:16 |
| hyperair | and i got intel so that i wouldn't hvae to experience nvidia/ati pain | 11:16 |
| hyperair | sigh | 11:16 |
| Ng | they're into bugfixing atm so 9.10 should be good for hardware they care about, but there'll be a flood of old stuff breaking I reckon | 11:16 |
| Ng | (this is all my wild speculation) | 11:16 |
| hyperair | lo | 11:17 |
| hyperair | well i'll just backport stuff i care about from jaunty to intrepid | 11:17 |
| hyperair | =p | 11:17 |
| Ng | I wish I'd done something like run phoronix's 3d benchmark stuff before I upgraded. It'd be very interesting to see over time how it changes | 11:18 |
| Ng | also glxgears is a terrible test ;) | 11:18 |
| hyperair | heheh yeah | 11:18 |
| hyperair | is there any chance of -intel 2.6.2 getting into jaunty? | 15:22 |
| tjaalton | not sure it's compatible | 15:23 |
| hyperair | hmm it isn't? | 15:23 |
| hyperair | i read in some bug report that all it needs is a newer libdrm | 15:23 |
| hyperair | which means two packages upgraded and we'll get it | 15:23 |
| tjaalton | yes, but the headers are in the kernel these days | 15:24 |
| hyperair | yes, but the user was compiling on ubuntu jaunty =\ | 15:24 |
| hyperair | i wonder if the user was using a custom kernel or the jaunty one | 15:24 |
| tjaalton | maybe it's doable | 15:31 |
| hyperair | tjaalton: so there is a chance of it getting in? =D | 15:34 |
| tjaalton | I'd have to check what commits are needed to the headers | 15:34 |
| hyperair | hmm i see | 15:37 |
| quassel208 | is ext4 better then ext3? because i hear my hd sometimes having a hard time | 15:44 |
| hyperair | supposedly | 15:44 |
| hyperair | i haven't really tried it so i can't tell | 15:45 |
| hyperair | but if many others are reaping benefits i probably will reap more, since i've encrypted my whole drive =\ | 15:45 |
| tjaalton | grim-reaping you say? | 15:45 |
| hyperair | lol | 15:46 |
| mnemo | tjaalton: what intel driver version will ship with jaunty? 2.6.1 or 2.6.2 ?? | 16:02 |
| tjaalton | mnemo: I've no idea | 16:14 |
| mnemo | ok | 16:15 |
| tjaalton | read the scrollback | 16:16 |
| hyperair | well if it doesn't, i'll probably stick it in my ppa, hoping against hope that it fixes all the performance regressions | 16:17 |
| hyperair | otherwise i'll just sit around in intrepid and moan and groan to myself. | 16:17 |
| tjaalton | it most likely won't fix any performance problems | 16:17 |
| tjaalton | gone -> | 16:18 |
| hyperair | alright, then i'll sit around in intrepid and moan and groan to myself | 16:18 |
| hyperair | and maybe backport the features from jaunty i find interesting, like notify-osd | 16:18 |
| mnemo | hyperair: what chipset do you have? at least on my G45 UXA works great | 16:22 |
| hyperair | mnemo: do you play games? | 16:23 |
| mnemo | I play chromium at least but its like a pretty simple game | 16:23 |
| hyperair | mnemo: i can't even play one of those scroll-down shooters with a FPS that matches intrepid | 16:23 |
| hyperair | mnemo: x3100 i965 | 16:24 |
| hyperair | both uxa and exa suck | 16:24 |
| hyperair | running metacity each time of course | 16:24 |
| mnemo | im using compiz+UXA on jaunty and while its not super perfy, its fairly stable (on my chipset, ive heard horror stories on other cards though) | 16:25 |
| mnemo | I hope UXA will be rock solid for ubuntu 10.04 at least because I'm most likely going to stay on that for a long time... it's too distracting to be on pre-release all the time, heh | 16:26 |
| hyperair | well i generally get on during the beta | 16:28 |
| hyperair | because around that time, people start asking me to maintain banshee in a new distro version in the ppa. | 16:29 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!