chibong: hey can any1 tell me how to get full 3d acceleration on an ati radeon 9600 with dual head. Ive tried installing sum drivers but it never works
chibong: hello
chibong: #ati
TobiasTheCommie: heh, damn, both RROutputOrder and noRandR are radeonhd specific... frell :(
TobiasTheCommie: weee, just need to get rid of the stupid flickering, and then i'm in business
TobiasTheCommie: well, that and xv
TobiasTheCommie: airlied: if you want me to test something on the card (pal, or anything else) just let me know...
mcgreg: xv? arent we stil quite far away from xv? :)#
MrCooper: yeah, it can only be done using the 3D engine on Avivo based cards
libv: so technically, it should already be possible on r5xx
libv: with dri already working
MrCooper: DRI != 3D
libv: well, but doesn't this mean that we have sufficient information to do this already?
TobiasTheCommie: mcgreg: hence why i just dropped it in there like that, i have no expectation of getting that working right now
MrCooper: libv: I don't think much more is known about the 3D engine than it's 'similar' to the previous R300 ones
MrCooper: so docs or a lot of REing are still needed
libv: ah, ok, so we do not have the relevant information yet then :(
glisse: MrCooper: after taking a look i would say that only fragprog instruction are different
glisse: MrCooper: it looks like r500 frag unit are an evolution of previous one
MrCooper: k
TobiasTheCommie: what is a frag unit?
agd5f: TobiasTheCommie: fragment shader
TobiasTheCommie: agd5f: ah, thanks
TobiasTheCommie: :D
damentz: hey airlied
damentz: sorry to keep buggin you daily, but the 6.7 series plays games way faster than the 6.6
damentz: if the 6.6 series worked fine with resolution changing, even with an incompatible laptop panel, isn't there a way to implement the old method?
damentz: airlied: i don't even know if your the guy that does the mode changing stuff
airlied: agd5f: you know anything about a cursor update bug on radeon 9600?
agd5f: airlied: need more detail. there were some hangs related to cursor updates and engine idling
agd5f: but MrCooper fixed that by using the indirect regs
agd5f: ages ago
airlied: agd5f: "when the cursor is moved from display 1 to display 2, its shape simply "carries over" to display 2 and is not updated properly according to the display 2 context."
airlied: probably explains what the customer is seeing..
damentz: airlied: strange, i also don't get video playback with xv with the 6.7 driver on my mobility card
damentz: it just has a blue screen
agd5f: airlied: not sure on that one. if they are using rotation there was a bug there that I fixed a while back
agd5f: damentz: what format? if it planar, XPRESS chips apparenly don't support planar formats in the overlay so when planar support was added, XPRESS chips broke.
agd5f: damentz the other possibility is that the overlay is just sourced to the other head if you have dualhead or a clone setup
agd5f: xvattr -a XV_CRTC -v 0/1 depending which crtc you want it on
damentzlap_: agd5f: well, i switched back to 6.6 and the video is accelerated
damentzlap_: i was using 6.7.197 when it broke
damentzlap_: or didn't play, in this case
damentzlap_: i can try it again
agd5f: damentzlap_: it could be either the of the issues I mentioned
damentzlap_: and i only have one output
damentzlap_: lvds
damentzlap_: no monitor attached
damentzlap_: oh
damentzlap_: Host/Kernel/OS "damentz-lap" running Linux 2.6.24-rc7-slh-smp-2 i686 [ sidux 2007-04 - Έρως - kde-full - (200711210243) ]
damentzlap_: CPU Info Intel Pentium 4 Mobile 512 KB cache flags( sse2 ) clocked at [ 1196.148 MHz ]
damentzlap_: Videocard ATI Radeon Mobility M7 LW [Radeon Mobility 7500] X.Org 1.4.0.90 [ 1400x1050@60.0hz ]
damentzlap_: Network cards Intel 82801CAM (ICH3) PRO/100 VE (LOM) Ethernet Controller, at port: 2000
damentzlap_: Intersil Prism 2.5 Wavelan chipset
damentzlap_: Processes 82 | Uptime 3:43 | Memory 131.4/503.6MB | HDD ATA HITACHI_DK23EA-6 Size 60GB (39%used) | Client Konversation 1.0.1 | Infobash v3.01
damentzlap_: if that helps any
agd5f: damentzlap_: does xvattr -a XV_CRTC -v 0/1 help?
agd5f: also if you are using mplayer, does mplayer -vf format yuy2 show a correct picture?
agd5f: https://bugs.freedesktop.org/show_bug.cgi?id=12744
agd5f: damentzlap_: looks like you have an m7. you said you had an XPRESS chip before
xAFFE: hi folks, I'm currently testing radeonhd and so far it behaves nicely :)
NizzleFoS: hey guys
NizzleFoS: anyone in here?
xAFFE: a bit ;)
NizzleFoS: cool
NizzleFoS: quick question
NizzleFoS: do any of you know if AMD's upcoming open-source support will include older legacy cards, like my All-in-wonder 9600 series?
xAFFE: hmm, thats something I dont know, but have a nice stay and idle, maybe someone will know
NizzleFoS: thanks for your help xAFFE
NizzleFoS: *sigh*
NizzleFoS: lol
NizzleFoS: that's probably one of the few things holding me back from switching to Linux
NizzleFoS: I want to be able to run Beryl/Compiz accelerated
airlied: NizzleFoS: it should be accel al ready with open source on 9600
NizzleFoS: oh?
NizzleFoS: there's support for that?
NizzleFoS: ?
NizzleFoS: \\ a 9600 driver
NizzleFoS: ?
NizzleFoS: perhaps I'm running an old distro then
airlied: the radeon driver should run on a 9600 fine..
NizzleFoS: ah ok
NizzleFoS: I'll give it a shot
NizzleFoS: thanks
damentzlap_: NizzleFoS: what distro?
NizzleFoS: Ubuntu
NizzleFoS: the version right before Fiesty Fawn
damentz_: ya
damentz_: thats pretty old man
NizzleFoS: so 7.04 or 7.02
NizzleFoS: whatever it was
damentz_: 6.10
damentz_: go ahead and get gutsy gibbon
NizzleFoS: naw, wasn't there a 7 something?
NizzleFoS: oh
damentz_: 7.10
NizzleFoS: nm
damentz_: or upgrade at least
NizzleFoS: fiesty fawn was 7.0x
NizzleFoS: right?
damentz_: yes
airlied: Edgy something.
damentz_: 7.04
NizzleFoS: yeah
NizzleFoS: that's the one
damentz_: edgy eft is 6.10
damentz_: i don't even use ubuntu and i know this stuff :/
NizzleFoS: nah, I tried 7.04
NizzleFoS: haha
NizzleFoS: more power to ya
damentz_: i suppose
damentz_: i order every cd
damentz_: and i see if its improved
damentz_: so far each release does something i don't want it to
damentz_: and i like up2date packages all the time, so i stick with debian sid
damentz_: on linux, theres no reason to be using an old version of a distro
damentz_: all you get is out dated software
damentz_: windows is a different matter though, kind of sad
damentz_: people are downgrading
NizzleFoS: yeah
NizzleFoS: Vista's tanking hardcore
damentz_: i think its the fact that people want them to be unsuccessful now
damentz_: since we know what they do when they have a monopoly
damentz_: or the majority percentage of consumers
NizzleFoS: you can sloppy, very easily
damentz_: so because everyone wants them to die so badly
NizzleFoS: *get
damentz_: the devs are probably lacking motivation
damentz_: i couldn't imagine developing windows, in windows
damentz_: sounds too unproductive
NizzleFoS: haha
NizzleFoS: funny you should say that
NizzleFoS: we had a MS game developer come talk to us my senior year of college
NizzleFoS: and he was singing the praises of their programming style
NizzleFoS: I forget what it was called
NizzleFoS: the one where you have the main dispatch loop
NizzleFoS: that runs every x seconds/minutes
damentzlap_: dispatch loop?
NizzleFoS: yeah
NizzleFoS: I think that's what he called it
xAFFF: re, hmm, it seems that the radeonhd drivers support composit? is that possible?
damentzlap_: nope
xAFFF: hmm, thats strange, because I have transparant urxvts. and iirc they use composite to do transparency stuff oO
umtc: i finished setting up a r300 card with tvout, but it has some colored speckles
umtc: it seems like in an area that is too white/bright, you get colored speckles
umtc: (all this refers to the s-video out using xv)
umtc: years ago when i used fglrx, i solved this by telling the driver to reduce the s-video output saturation
umtc: i can turn off the saturation by setting the XV_SATURATION attribute to -1000, and they do disappear but any color at all brings them back
umtc: i'll be sure to pester agd5f about it when he comes back
umtc: seems like a visual representation of a numeric overflow. it is odd that is controlled by the saturation power when the areas that have them have no color
umtc: i don't have that old configuration file, but the setting was called "TVColorAdj" and had to be set to about -30 to eliminate the speckles