Phoronix IRC Log: 2009-05-23
Lightkey: it's bad enough that Phoronix posts an opinion on whether or UT3 comes out and regards that as "news", how could some actual (old) reply from the only one who knows for sure be any worse? :-p
peppo: anyone have some insight on how far away the r600 code is from compositing and xv?
Ivanovic: peppo: xv is really close (should be working with 2.6.30 kernel and latest (git) radeonhd/radeon driver
peppo: Ivanovic, nice
Ivanovic: peppo: though i have not tested it yet
Ivanovic: last time i wanted to try it with building drm drivers right out of the mesa/drm tree and using them xorg was not able to start...
Ivanovic: (since you have to force on dri and select exa to make use of the stuff)
Kano: hi michaellarabel , just saw your latest commit, will test it later
michaellarabel: Kano: k
Kano: currently recording, will not force a 3d crash ;)
maligor: compositing probably works using xrender in it
maligor: I really should slice off some space and test radeonhd git
bkero: I'm wondering why my Vertex SSD isn't getting nearly as impressive readings in dbench
maligor: hmm.. I wonder if that ssd runs linux
bkero: Nearly identical hardware, video card and 200MHz clockspeed difference.
bkero: heard tale that the 30GB Vertex's had 32MB of cache instead of 64, however the case says 64.
maligor: looks like it has JTAG on the board
bkero: The solid/core series had uSB
bkero: Also a terrible JMicron controller
maligor: was looking at this naked ssd pr0n: http://www.pcper.com/article.php?aid=670&type=expert
maligor: it even looks like a standard 20 pin arm jtag
maligor: not sure how useful it'd be to actually run custom firmware on something like that, heh
bkero: It'll be nice when SSDs don't have to emulate block devices.
maligor: they don't
maligor: those are called MTDs
bkero: I'm saying it'll be nice when I can get MTD SSDs
bkero: runs half his VRAM as an MTD
maligor: there's no bus for it
bkero: Just use what turbocache/robson does and stick it on the pci bus
maligor: it's just not terribly useful to have a raw MTD device
maligor: you'd pretty much have to use one of the embedded fs
maligor: like jffs2
bkero: There are plenty in development
bkero: I use one on my beagleboard, which has 256MB of NAND on it as an MTD.
bkero: Linux does the wear leveling, etc.
maligor: yeah, or a wear leveling layer
maligor: you mean LogFS?
bkero: There was logfs, another one too
bkero: It got mentioned on the beagleboards front page, which is where I heard about it
maligor: UBIFS I guess
bkero: That sounds right
Kano: michaellarabel: that was the usual fun... your git is out of date
Kano: could you update it
Kano: i know it would crash somehow
michaellarabel: Kano: updated
Kano: i should have grepped before..
Kano: michaellarabel: it still asks for res
Kano: SKIP_TESTS=vdrift phoronix-test-suite benchmark vdrift
Kano: at which point it should leave
michaellarabel: Kano: It will ask you, but it will skip it during testing
michaellarabel: though, I could make it skip that part too...
Kano: ok, it was not started later
maligor: heh, decided to try fedora 11 on a laptop.. promising.. Installing glibc 2.9.90...
maligor: I guess it is still a preview but still
kristian_: hi all - looking to scale down the image on my tv
kristian_: I'm on a Radeon 3450, Linux Mint (Ubuntu fork)
Kano: michaellarabel: rc7 is out
Kano: i already use rc7, just compiled a new kernel and saw it in the git log
michaellarabel: Kano: I know, already posted about it
Kano: i had some issues with aufs2 + 64 bit, therefore i got a patch, i wanted to test that
Kano: creating new iso now
Kano: 32 bit it looked already stable, but 64 bit crashed aufs
redeeman: hmm i wonder what chunk size bitmap i should use on my new raid
Kano: so, 64 bit with 2.6.30-rc7 works now with patched aufs2..
Kano: anybody wants to try ;)