[15:07] <zumbi> hi! if at Debian set a subarch name for Freescale i.MX51, i.e., "mx5" would likely break Ubuntu stuff? could you migrate your code base to new subarch name? http://lists.debian.org/debian-arm/2011/04/msg00009.html
[15:10] <ndec> ogra_: hi!
[15:10] <ndec> ogra_: have you had a look at the UCM stuff?
[15:11] <ndec> ogra_: i am sorry I don't have much time to look into this in details...
[15:20] <ogra_> ndec, well, i expected one config file and two PA patches ... somehow it got a lot more, TheMuso will take a look at it tomorrow he promised
[15:25] <ndec> ogra_: there are several config files, because there are several profiles. if you want to take just 1 profile, that should be doable.
[15:27] <ogra_> ndec, well, lets wait for TheMuso, he is our alsa/pulse maintainer, if he has issues i can get him connected with alejandro
[15:28] <ogra_> i just didnt expect that amount of patches, i honestly didnt look at them yet waiting for an experts opinion
[15:59] <GrueMaster> ogra_: None of the omap images are building.  Stuck buildd?
[15:59] <ogra_> GrueMaster, yep, lamont pinged already
[15:59] <GrueMaster> k
[15:59] <ogra_> as long as omap4 is fine ...
[17:19] <sveinse> How is the performance of X and OMAP3 (w/SGX)?
[17:19] <sveinse> We are currently developing Qt against QWS and are looking for more performance. I suspect X is better optimized and more mainstream and thus has more focus
[17:20] <sveinse> One suggestion is to run X and one full screen Qt app
[17:20] <rsalveti> sveinse: with omap3 I'd avoid running with X
[17:21] <rsalveti> when possible, as it's not properly integrated
[17:21] <sveinse> not properly integrated = missing dri drivers?
[17:21] <rsalveti> sveinse: yes
[17:21] <sveinse> alas
[17:21] <rsalveti> currently it's just a hack with some locking with the X11 server to write at the correct place at the framebuffer
[17:22] <sveinse> There are some free desktop SW bundled with the sgx driver. Are they unusable?
[17:22] <sveinse> So X i basically SW redering framebuffer
[17:23] <rsalveti> sveinse: it probably works, but not optimal
[17:23] <rsalveti> not sw rendering, but not respecting the x11 dri architecture
[17:23] <rsalveti> sveinse: talk with michaedw at #linaro, he's working with omap3 and also creating something with Qt and QWS
[17:24] <rsalveti> but using just the framebuffer, and he's getting quite nice performance values
[17:25] <sveinse> rsalveti: Thanks, it helps not to have to learn everything for yourself from scratch!
[17:26] <suihkulokki> sveinse: look at Qt lighthouse: http://qt.nokia.com/developer/learning/online/talks/developerdays2010/tech-talks/project-lighthouse/
[17:30] <sveinse> suihkulokki: AFAIK the biggest problem is translating the Qt graphics primitives into HW accelerated graphics, as it requires insight into the closed SGX/PowerVR module. Right?
[17:30] <rsalveti> sveinse: nops, you would just use gles normally
[17:31] <rsalveti> generally you don't need to touch sgx besides normally using gles
[17:31] <rsalveti> I know michaedw did some optimizations for sgx, but don't think he's touching it directly
[17:32] <davidm> GrueMaster, you about?
[17:32] <GrueMaster> yes
[17:32] <sveinse> ah, so you use lighthouse to bind Qt primitives and the gles lib from imagem/TI
[17:32] <suihkulokki> sveinse: indeed there is nothing about the closed nature you need to care.. the opengl es api is where you work at. the big deal is that traditional qwidgets are painted serially while gpu's are good are doing things parellely
[17:33] <sveinse> hmm, I have obviously lots to learn of how the graphics parts are interconnected
[17:35] <suihkulokki> sveinse: the video on my link is good start =)
[17:35] <sveinse> perhaps a stupid question: could one then make s X driver using the gles to achieve HW accel in X then?
[17:35] <sveinse> Yes, i will right now, thanks
[17:37] <rsalveti> sveinse: well, you could, but generally people are worried only with 2d bits at the xdriver
[17:37] <rsalveti> and for that you could use pvr2d, neon or even other hardware when available
[17:37] <rsalveti> to accelerate EXA
[17:37] <rsalveti> and once you have DRI support it should be fine
[17:40] <sveinse> Generally I'm looking for the solution of Qt apps which gives us the most both in respect of performance and with less development effort
[17:40] <sveinse> rsalveti: BTW: Have you taken a look at 4.03.00.02 ? Anything interesting new stuff here?
[17:41] <sveinse> (for OMAP3)
[17:42] <rsalveti> sveinse: still need to test that with the xserver, planned for this week
[17:43] <sveinse> One thing though is that it now support active power management in OMAP35x/37x, see http://processors.wiki.ti.com/index.php/RN_4_03_00_02
[17:44] <sveinse> Nice for mobile / low-power devices
[17:52] <zumbi> tcl2010rocks
[17:52] <zumbi> :P
[17:55]  * rsalveti lunch
[20:50] <press> Hi,
[20:50] <press> I'm an amateur-journalist wanting to write an article about Ubuntu on ARM.
[20:50] <press> Who would be the best person to help me / answer my questions?
[20:50] <press> Thought this was a good place to ask...
[20:53] <rsalveti> press: sure, just ask your question
[20:53] <press> OK, my name is Hans BTW.
[20:54] <rsalveti> press: http://irclogs.ubuntu.com/2011/04/01/%23ubuntu-classroom.html#t17:31
[20:54] <press> thnx
[20:55] <rsalveti> we had one session at ubuntu-classroom with davidm last week, that can also help you answering some questions
[20:55] <press> Gonna read that first
[20:57] <rsalveti> press: sure, then just ask later your questions
[20:58] <rsalveti> as it's an irc, don't expect them to be answered at the same moment
[20:58] <rsalveti> press: as you're using a web irc, you can check the channel logs later
[20:58] <rsalveti> but we're usually around
[21:20] <press> OK, thanks very much.
[21:21] <press> I'm pretty much a n00b when it comes to ARM/Linux development.
[21:21] <press> So will read some more 'old' #Ubuntu-ARM logs tomorrow, make up some nice questions and come back later.
[21:23] <rsalveti> press: cool