[00:23] <DanaG> dang: powervr drivers have no omapfb_dri.so (or such) for real GLX.
[00:25] <DanaG> Say, what does GMA500 get?  If gma500 gets GLX, then why does OMAP SGX not get GLX?
[01:50] <Vic_Sage_The_Que> anybody here
[01:50] <Vic_Sage_The_Que> ??
[01:54] <Vic_Sage_The_Que> how is the performance of an ARM processor vs an X86 for things like GIMP
[01:57] <Baybal> gimp it depends on version
[01:57] <Baybal> before gegl it was mostly integer based
[01:57] <Baybal> now its mostly float math based
[01:58] <Baybal> arm is not a best float point performer on the market
[01:58] <Baybal> although A8 knocks down atoms in integer ops
[01:59] <robclark> fwiw, cortex-a9 is better at fp... but integer and single precision float has more potential for simd..
[01:59] <Vic_Sage_The_Que> well i was looking at getting a tegra based tablet for design work on the go
[02:00] <Vic_Sage_The_Que> so i'd be running things like inkscape, scribus, karbon14
[02:01] <Vic_Sage_The_Que> would these perform ok??
[02:02] <robclark> on tegra, no idea
[02:04]  * prpplague pokes robclark in the eye with a sharp stick
[02:04] <robclark> ouch
[02:04] <Vic_Sage_The_Que> http://www.nvidia.com/object/tegra_250.html
[02:05] <Vic_Sage_The_Que> its a cortex-a9
[02:06] <robclark> Vic_Sage_The_Que: so, I guess the answer is maybe ;-)
[02:06] <robclark> I've not played with one, so I can't really say for sure
[02:07] <Vic_Sage_The_Que> has anybody done something like this I was looking at getting a notion ink adam and dual booting android and ubuntu arm
[02:08] <robclark> are they even available already?
[02:08] <Vic_Sage_The_Que> theres devices using similar hardware
[02:10] <Vic_Sage_The_Que> toshiba folio 100
[10:49] <ogra_cmpc> OMG ! they sell the toshiba AC100 around the corner here
[10:49]  * ogra_cmpc just found out 
[10:51] <ogra_cmpc> NCommander, persia, anyone intrested ? its 380 euro (tegra2, 512M) and qwertz kbd ... i could bring one to UDS
[11:21] <NCommander> ogra_cmpc: what's an AC100?
[11:21] <NCommander> oh wait
[11:21] <NCommander> YES!
[11:25] <kblin> nice.. apart from the qwertz keyboard ;)
[11:25]  * kblin shakes his head
[11:25] <NCommander> kblin: that can be fixed with a keymap
[11:25] <NCommander> <- *dvorak keyboard user (at times)*
[11:26] <kblin> NCommander: but it's irritating.. I'm doing that at work already
[11:26] <NCommander> kblin: it could be worse
[11:26] <kblin> true, could be windows
[11:27] <NCommander> ogra_cmpc: do you have a picture (I'm pretty sure I want one, but I can't find a pic of it here)
[11:27] <kblin> with it's broken version of focus-follows-mouse
[11:27] <kblin> engadget had one
[11:27] <kblin> first google hit for me
[11:28] <kblin> too bad I already reached my gadget allowance for this year..
[11:28] <NCommander> kblin: w.r.t. to Android, remapping is trivial if it can be rooted
[11:28] <NCommander> and if the kernel doesn't suck, it should run Ubuntu just fine.
[11:28] <armin76> linux support on the tegra2 is kinda awful
[11:29] <kblin> I shouldn't have told my $significant_other that the dockstar I got was a computer
[11:31] <ogra_cmpc> NCommander, only in a herman news article http://www.heise.de/mobil/artikel/10-Zoll-Netbook-Toshiba-AC100-mit-Nvidia-Tegra-250-und-Android-2-1-1072519.html
[11:32] <ogra_cmpc> *german
[11:32] <ogra_cmpc> try google translate... at the bottom there is a picture series
[11:33] <ogra_cmpc> kblin, well, its a german version ... so it has a german kbd
[11:33] <kblin> ogra_cmpc: sure. it's an uphill fight for me all the time to get a decent programming keyboard that has the right size enter key
[11:34] <ogra_cmpc> heh
[11:35] <kblin> the other differences are irellevant, but the shape of the enter key and thus the positition of the \ key, that can't be fixed by keymaps
[11:36] <ogra_cmpc> well, i find the price quite high, the kbd looks ok to me
[11:36] <NCommander> ogra_cmpc: its not hard to change the keymap.
[11:36] <ogra_cmpc> but i usually use german kbd mapping anyway
[11:37] <ogra_cmpc> and 512M are quite low
[11:39]  * kblin shrugs
[11:39] <armin76> low, heh
[11:39] <ogra_cmpc> compared to the panda ...
[11:39] <kblin> 60% or my ARM devices have 128MB only
[11:40] <ogra_cmpc> tegra also has no neon
[11:41] <ogra_cmpc> and given it comes with android the question is can you get a normal linux kernel source for it to run ubuntu
[11:41] <armin76> well, if you're comparing it with the omap4, then its going to suck for sure :)
[11:41] <ogra_cmpc> right
[11:42] <ogra_cmpc> and for that i find it a bit expensive
[11:42] <ogra_cmpc> though SATA makes it intresting again
[11:42] <armin76> oh, it has sata? the dev board doesn't
[11:43] <ogra_cmpc> the tech data i find everywhere for it says 8G SATA SSD
[11:43] <ogra_cmpc> might indeed be attached through USB or something simlary weird
[11:43] <armin76> probably
[11:44] <armin76> the ethernet on the dev board is usb too
[11:45] <armin76> and not sure if you've ever seen this on other platforms, but the nvidia linux development kit uses ubuntu and a daemon for something :D
[11:46] <ogra_cmpc> omap4 is developed natively on ubuntu ;) doesnt need a daemon or kit
[11:47] <armin76> yeah well, nvidia is "special" regarding linux
[11:48] <armin76> the bootloader used is still the one for android, fastboot
[11:52] <NCommander> armin76: if it is unlocked, or not hard to unlock, that's trivial to get around.
[11:54] <armin76> NCommander: you'd first need to check it the tegra kernel supports the toshiba :)
[12:00] <NCommander> armin76: heh
[12:04] <armin76> which is unlikely
[12:04] <armin76> http://nv-tegra.nvidia.com/gitweb/?p=linux-2.6.git;a=tree
[12:07] <armin76> sorry, this one: http://nv-tegra.nvidia.com/gitweb/?p=linux-2.6.git;a=tree;h=refs/heads/android-tegra-2.6.32;hb=android-tegra-2.6.32
[13:46] <rcn-ee> cwillu, ping.. (btrfs)
[16:00] <rsavoye> Not real exciting, but here's a shot of Gnash doping YouTube on the Sharp NetWalker: http://www.gnashdev.org/images/IMGP2581.JPG
[16:00] <rsavoye> this machine runs Jaunty
[16:02] <devilhorns> you're right ... not real exciting ;)
[16:05] <devilhorns> ewww red ... couldn't they have used a better color ? uff
[16:12] <rsavoye> I had too many black netbooks :-)
[16:12] <rsavoye> the netwalker is mostly for armel demos
[16:12] <devilhorns> ahhh :) yea that seems to be the standard color
[16:13] <rsavoye> at least with a red one, I can tell it's not one of my mipsel netbooks
[16:13] <devilhorns> true :)
[16:13] <rsavoye> the netwalker only runs flashlite, so doesn't handle youtube
[16:14] <rsavoye> but with gnash it does :-)
[16:14] <devilhorns> :)
[16:16] <rsavoye> the wonders of free software... now if I could just get maverick running on the netwalker, I'd be happy
[16:20] <devilhorns> hehe
[16:23] <rsavoye> I was bummed to see ubuntu drop i.mx51 support, but oh well...
[16:29] <lool> rsavoye: Note that the BSP isn't getting updated for latest kernels
[16:30] <lool> also, tracking BSPs is painful; if it was just for the upstream kernel, it would be easier
[16:30] <rsavoye> yes, but I don't plan to fix it
[16:30] <rsavoye> I know, I used to write BSPs for a living
[16:30] <rsavoye> I'm trying not to get my hopes up, but I may have just fixed pr 323 and pr484.
[16:31] <lool> I know you're not going to fix it, but you said you were bummed that Ubuntu dropped i.mx51 support  :-)
[16:33] <armin76> ubuntu dropped i.mx51 support?
[16:46] <armin76> lool: ^
[16:51] <lool> armin76: well we dropped the bsp
[16:51] <lool> linux-fsl-imx51
[16:51] <lool> but linaro builds an upstream imx51 flavour
[16:52] <armin76> ah well, thanks
[16:54] <rsavoye> ah, interesting. just released
[17:23]  * cwillu pokes rcn-ee 
[18:01] <rsavoye> yippie, my fixed arm assembler for OpenJDK passes the test suite :-)
[18:02] <rsavoye> it actually gets better results than the interpreted mode
[18:03] <robclark_> rsalveti: JIT?
[18:04] <rsavoye> thumb2 JIT appears to be working as well now
[18:04] <rsavoye> damn gcc optimizer bug :-(
[18:04] <robclark_> nice
[18:05] <rsavoye> working with b20 in icedtea6 HEAD too :-)
[18:05] <rsavoye> al the test that failed with interpreted mode are now just errors, but passing
[18:26] <rsavoye> yep, slightly better results than the arm interpreter, now running the langtools tests
[18:36] <rsavoye> http://www.senecass.com/projects/OpenJDK-ARM/icedtea6-armel-asm/hotspot/
[21:45]  * cwillu_at_work pokes rcn-ee again
[21:57] <rcn-ee> hey cwillu_at_work, was testing out btrfs, do you use differnet bootargs (such as 'rw') or disable fsck in /etc/fstab?
[21:57] <cwillu_at_work> rcn-ee, both
[21:58] <cwillu_at_work> btrfs's fsck doesn't do anything yet, although somebody's working on that
[21:58] <cwillu_at_work> or rather, fsck doesn't _fix_ anything yet, and fails at the same point that btrfs would itself detect an error
[22:01] <cwillu_at_work> well, that's what I do on my beagle images
[22:02] <cwillu_at_work> I've got two desktop's with btrfs roots as well, they're both ro with nothing done to disable fsck
[22:05] <rcn-ee> Thanks cwillu_at_work, i remember you saying the btrfs's fsck wasn't ready, so i wasn't sure which method you did to get it to boot.. (it fails when trying to find the fake btrfs fsck from btrfs-tools)