Phoronix IRC Log: 2008-08-23
Kano: michaellarabel: in your review of 8-8 you write that you cp the default file over the config file. thats not needed, just delete it.
Kano: btw. i updated the 2.6.27 patches and now 8-8 is installed by my script with dkms active and 2.6.27 support
Kano: i even got a "new" ati ,card, the 2nd rv410, just this time a "pro" edition *g*
maligor: Kano, where's the patch found?
Kano: i made it myself
Kano: just like the 2.6.27 fix for 8-7
Kano: it is not that hard
maligor: last build of 2.6.27-rc4 didn't boot so I haven't gotten to attempting it yet
maligor: or well, it did boot but crashes at a specific point
roberth: okay the ati driver is beginning to seriously piss me off
Kano: will check 2.6.27-rc4-git soon
maligor: roberth, how so?
roberth: after i upgraded the driver, the driver has switched the default display adapter
roberth: which means that the monitor which is connected to the rigth dvi output on my graphics card, is now teh secondary display
roberth: and my the tv is the primary
roberth: so now I'm sitting here with 1386x768 resolution on my monitor which should have been on my tv, and the tv has 1680x1050
roberth: and I cant switch back:(
roberth: since if i connect the monitor the left dvi output i only get out of range...
maligor: you can try resetting amdpcsdb
roberth: how do I do that?
maligor: stop X, copy /etc/ati/amdpcsdb.default over /etc/ati/amdpcsdb
Kano: maligor: just delete the 2nd file you mentioned
maligor: or that
roberth: that helped! thanks
Kano: the export symbol bug is solved in 2.6.27-rc4, but they forgot a tab, well only optical ;)
maligor: Kano, what? rcu?
Kano: maligor: well i posted something in the 2.6.27-rc3 thead, that is solved
Kano: the paravirt thing
Kano: hmm the new rc4 kernel is absolutely crap...
Kano: the one export symbol is fixed but now my nv agp card does not init and fglrx locks up
redeeman: and so because some crappy external drivers doesent work
redeeman: the kernel is absolute crap
redeeman: it's astounding that anyone can come to such conclusions
cxo: redeeman, that's what you call the "layman"
redeeman: a more correct term is "disturbed individual"
cxo: there was someone here last night trying to get his fglrx working on preemp-rt
cxo: if you know anything about preemp-rt, you would know, that's like stabbing yourself and seeing if bleed
roberth: does catalyst 8.8 for with kde 4.1 for anyone here?
Deanjo: *sigh* when the hell is ati going to fix the bloody tearing in video playback in any of their drivers for the 690g?
roberth: Deanjo: well the issue is that it isnt possible to use vsync of XV
Deanjo: easier solution involves slapping a nvidia card in
roberth: yes and then you are leaved with crappy 2d performance unless you use an older card
Deanjo: Hey, 8800 GT's working fine here, but for the 690g even a el cheapo 7200GS stomps on the ati
roberth: i had an 8800gt i had, had so slow 2d performance that i switched to and ati card
roberth: because i could stand tearing
Deanjo: Never had issues with 2d truthfully on a 8800GT appreciable to make a switch
Deanjo: Never had a issue with tearing ever
Deanjo: on the 8800GT's
Deanjo: From composite to HDMI
roberth: i dont see have you cant notice it, since its well known issue
roberth: and I can without any doubt notice it
Deanjo: 1080p no prob with no tearing
roberth: 720p i fucked for me with nvidia because of the crappy 2d performance
Deanjo: 720p no prob even on a 7200GS
roberth: when i pause the video and play it again, the audio jumps like 10 seconds forward, which is a typical sign of crappy 2d performance
roberth: for the nv driver has almost as good 2d performance, and doesnt have accelerated 2d rendering haha
Deanjo: And ATI does on x264?
Deanjo: ummm no
roberth: at least it has normal performance as it should be since it has decent 2d performance
roberth: but the tearing is there
Deanjo: Not here at all
roberth: you must have a seriously fucked up setup
Deanjo: Not really, got many systems setup all perform nicely
roberth: well its works fine here, and nvidia.....works like for many others....crappy 2d performance
Deanjo: Not saying that some have issues.
Deanjo: But like any other vender, intel/AMD/nvidia/etc a vocal few will always make things seem bigger then they are
roberth: few? okay seriosly....
roberth: its the opposite
Deanjo: Hell I had a guy the other day demanding to talk to engineering the other day demanding a replacement machine because of the G84 issue. It was "a well known fact of defect". Problem was, the guy was running a intel 3100 graphics.
Deanjo: Funny thing is there was 80 more people joining the bandwagon.
roberth: what kind of issue exactly?
Deanjo: With everything from ATI - nvidia - intel graphics all claiming that "it was a well known issue"
roberth: what issue?
Deanjo: G84's melting on Dell laptops
Deanjo: Hell my company ain't even Dell
roberth: well hardware issues is a pain
roberth: specially on laptops where you can repair it your self
Deanjo: Point is people many times "see what they should see" and not even realize that it has nothing to do with them
Deanjo: and will jump a bandwagon blindly
Deanjo: Especially when you have shoddy sites *cough* inquirer *cough* that glorify and exaggerate issues to the point were reality no longer applies
Deanjo: Any developer will tell you bugs that effect 100% of your audience are much easier to solve then issues that effect a vocal minority.
Deanjo: For example the email addy for nvidia for bugs has been around for years but yet it wasn't until just recently people actually figured out that "Hey I read the bloody documentation there is a proper way of reporting issues." To many times people scream and yell but give useless information usually in a unofficial channel.
michaellarabel: Deanjo, but NVIDIA's email address for reporting bugs is still crap. Especially when they provide tracking numbers but no tracking system.
roberth: me when using nvidia: http://img.4chan.org/b/src/1219539178644.jpg
Deanjo: As oppose to a bugtracker that is never looked at michaellarabel? ATI has many outstanding bugs that have been around for eons
roberth: michaellarabel: i hate the way amd/ati is releasing the driver, in the state its now
michaellarabel: Deanjo, ATI/AMD does actively look at the bug tracker, I will tell you that.
roberth: they should rather release a welltested driver more rarely rather than push out driver full of regressions etc every month
michaellarabel: roberth, for more well-tested drivers use the FireGL releases.
roberth: this month amd/ati has at leats filled some the gaps when it comes to features for the driver
michaellarabel: roberth, they're not done filling gaps yet
roberth: michaellarabel: i dont have a firegl card...
Deanjo: michaellarabel, you remember the "Big wine/cedega" improvements that were to be had in 2006 on ATI? Well it improved for a few driver releases and then fell back into a state obscurity.
roberth: michaellarabel: what outstanding features do they miss?
michaellarabel: roberth, their driver code is all integrated. But they release new "FireGL releases" quarterly that go through greater QA then the monthly ones
michaellarabel: roberth, I won't comment on that but you can look in the forums to find areas where users have been looking for improvements?
Deanjo: Shouldn't those be marked as a beta then?
roberth: Deanjo: issues etc back at that time is not related, in that time ati had a quite different view on linux support
Deanjo: roberth, that announcement was AFTER amd said "make it so"
michaellarabel: but with the 2006 release still on their old architecture
Deanjo: It was right after AMD's acquirement of ATI
Deanjo: Right old arch, but even with the new arch it still exists
roberth: Deanjo: well, this new linux support didnt occour that early
roberth: ati is way much better on linux today than before
roberth: michaellarabel: the issue with tearing under video playback with ati is cause by the lach of vsync on xvideo right?
roberth: and if amd doesnt fix that before christmass....
Deanjo: new linux support still hasn't improved the cedega/wine situation though roberth. That was one of AMD's big announcement's when they took over ATI
roberth: what improvements with wine/cedega?
Deanjo: Like 5 days after the acquirement
Deanjo: Getting it to run not sucking ass for 1
Deanjo: Hell just getting it to run was a improvement
michaellarabel: WINE/Cedega will be improved soon
michaellarabel: Part of their problem with WINE is the current lack of automated regression tests for WINE
Deanjo: What about Pro/E and maya being limited to 16-bit?
roberth: michaellarabel: ah so catalyst 8.5 is the latest "stable" driver from amd, so thats why there has just been so many bugs since that release
roberth: michaellarabel: please write and article about this on phoronix, a lot of people dont know this
cxo: hey Deanjo i just installed vmware workstation 6.04
cxo: there is no direct 3d support with it
cxo: directx9 reports only direct draw support