Argh! ATI!

Solas · Post by **Solas** » Tue Feb 24, 2009 8:01 pm

I didn't know about file virtualisation when the mouse posed problems.
a mouse is much cheaper to replace than a good graphics card.

OpenGL and Aero windows or popups cannot overlap.
the system stutters or slows down while Aero turns itself off.
on my machine it was annoying.

turning off UAC only comes into effect after reboot.
try any Oolite installation after. I'll be optimistic till then

Screet · Post by **Screet** » Tue Feb 24, 2009 8:26 pm

Solas wrote:
OpenGL and Aero windows or popups cannot overlap.
the system stutters or slows down while Aero turns itself off.
on my machine it was annoying.

turning off UAC only comes into effect after reboot.
try any Oolite installation after. I'll be optimistic till then

I did reboot, but I didn't uninstall the C:\oolite installation and reinstalled it. Wonder what that would be good for doing it another time.

Concerning the Aero, I don't get stuttering. Firefox turns it off anyway

For reporting that the X800XTPE outperforms my 4870x2 in OpenGL I was banned again from the forum and my messages deleted. Therefore, I cannot report that bug you found. If you want it visible, please create an account and report it - it's easy and quickly done! I'm run out of email addresses to use.

The threads in which the people complained about poor OpenGL performance have been locked. Not sure if it's their companies idea to lock all threads on page 2, but IIRC that wasn't the case when I joined their forum.

After all these experiences with them there it looks to me as if ATI is trying to cover up some hardware issue, afraid of having to pay back a huge amount of money to customers. A few comments from the moderator do hint that way, also he surely didn't want it to sound that way.

Honestly, I did even try to help a dissatisfied customer and their tech support on a problem another user had which I did not have with my card and I did ask all the time what I could do to help them to track down that problem. No answers, but messages were deleted and my accounts banned.

Add to this that I've got one of those Phenoms which they did slow down after selling the processors. Some speed tests on the internet reported as much as 30% decrease in processor power although noone ever was able to produce that bug AMD feared to arise. I wonder who would accept this with a car or something like that. I turned off that slowdown mechanism in BIOS (thanks to Asus for providing that option against AMDs advice!).

It looks like the vendor of my card could be willing to take it back in order to allow me the purchase of a nVidia 295. Guess that's the way I go. I asked nVidia germany if to have a look at those threads and consider wether they would want to help the users in the US who search a lawyer now, and wether they could do tell me where to find a good one or help here in the same way. Didn't get a response yet, hope they will. If I knew about these problems, I had never purchased that card, thus nVidia definitely lost me as a possible customer because of false advertising by ATI.

Screet

Screet · Post by **Screet** » Tue Feb 24, 2009 8:39 pm

Interesting. Some user complained about his 4870x2 making extreme rattling noises with the fan after installing the 9.2 drivers. He said that the noise went away when he returned to 9.1 drivers.

I asked him to check if the problem persists after reinstallation of the 9.2 drivers and turning on manual fan control, as he also did report much higher temperature on the card in 2D mode and I figured that this could be the autmatic fan control overreacting to temperature changes.

That entire thread was then deleted.

Some Linux guy did also report problems which probably should be checked for possible oolite impact:

Code: Select all


Hi,

I am a owner of a AIT Radeon 4870 using Linux operating system.

The current drivers cause many trouble with some game like EVE-Online because of some errornous limit in the drivers.

First problem, the GL_MAX_VARYING_FLOATS_ARB limit is too high. On my card, I got a limit of 68, so 17 varyings. With 17, some game, incluing Wine fail to compile the shaders because of some unusable varyings. Lowering the valur to 13 (GL_MAX_VARYING_FLOATS_ARB = 52) fix this first problem but we must hardcode it into the game / directx wrapper instead of using the standard detection code. So this force anyone wanting working GLSL 1.3 to edit there game source code to fixe it.

First fix : You can just report good working value.

 

Second problem : OpenGL limit (max_fragment/vertex_uniform_components) are too low, GLSL 1.3 require 1024 and my system detect 512 with Catalyst 9.2. This issue cause a bug problem for games using vec4's > 225 causing shader to not compile. Overriding these value don't work (Shader compile but don't show on screen due to drivers not accepting them). This seems to be related to the unified scalar architecture all return a value of 512. change since the hd2**0 cards as the that series up to the current hd4**0 cards

FIX your Linux OpenGL please!

Many people are currently buying NVidia or switching to NVidia because of bad issue like that.You have some windows games trying supporting Linux and they just tell : "You have a problem with shader not rendered? Buy a Nvidia card..."

PS : Sorry for my bad english

Screet

Cmd. Cheyd · Post by **Cmd. Cheyd** » Tue Feb 24, 2009 8:41 pm

Screet wrote:
...thus nVidia definitely lost me as a possible customer because of false advertising by ATI.

Huh?
Nvidia is a competitor to ATI. Why would you stop buying Nvidia's products because of something their competitor screwed up???

I realize English isn't your first language, but you normally do very well. Is this just a translation error?

Screet · Post by **Screet** » Tue Feb 24, 2009 8:52 pm

Cmd. Cheyd wrote:
Screet wrote:
...thus nVidia definitely lost me as a possible customer because of false advertising by ATI.
Huh?
Nvidia is a competitor to ATI. Why would you stop buying Nvidia's products because of something their competitor screwed up???

I realize English isn't your first language, but you normally do very well. Is this just a translation error?

I am very tired, might be

2 Days without sleep...insomnia is to blame.

What I did want to say: Would I have known that the advertisements of ATI were not true for my card, I would have bought nVidia. Thus, nVidia had at least one less customer because he believed ATI ads! Reading the AMD board, there's obviously more people who would have bought nVidia, would they have nowadays knowledge.

Screet

Screet · Post by **Screet** » Tue Feb 24, 2009 10:15 pm

Screet wrote:
Interesting. Some user complained about his 4870x2 making extreme rattling noises with the fan after installing the 9.2 drivers.
[...]That entire thread was then deleted.

Ooops

Seems I've now really some trouble because of the missing sleep. Did find the thread again, just my suggestion to test with manual fan control removed. Still a strange thing to remove such a suggestion, but the thread's still there (or back).

Screet

Solas · Post by **Solas** » Tue Feb 24, 2009 11:32 pm

"try any Oolite installation after." ... I should have said ..
turn off UAC and reboot. then try any existing Oolite installation. or a new one.

I think your card may have X2 specific driver/configuration issue(s).
but it still feels like there's a file virtualisation problem somewhere.

I've posted the ATI star / nebula bug on their forum.
http://forums.amd.com/game/messageview. ... erthread=y
we will see how long it takes them to correct that bug.
still no response to my ticket pre 9.1

Screet · Post by **Screet** » Wed Feb 25, 2009 4:34 pm

Solas wrote:
"try any Oolite installation after." ... I should have said ..
turn off UAC and reboot. then try any existing Oolite installation. or a new one.

I think your card may have X2 specific driver/configuration issue(s).
but it still feels like there's a file virtualisation problem somewhere.

OK, then I tried it just as you suggested.

Today, I finally felt up to it...and tried to install my XP X64 to that machine in order to check if the problem persists with X64 and the ATI driver for it.

Sadly, halfway through the installation, the XP installer always froze and, of course, I had to repair the vista installation afterwards.

No news from Asus yet, however ATI finally requested more information concerning that problem. Sent it to them and am awaiting their reply.

The card's vendor seems to be really willing to take the card back, as they just requested a response from me towards their reply upon the possibilty to return it. Guess I wait a little bit longer to see if ATI or Asus finally can fix that problem, and if they don't, I'm going to use the vendors offer to consider the card malfunctioning and exchange it for a nVidia 295 (with a little bit of money for me to pay as that one is slightly more expensive than the 4870x2). I'm really tempted to go that way at once, as you might guess

Reading the ATI BBS, it seems to me that the 4870x2 card causes massive problems. I'm quite curious as to what's wrong, drivers not being tested well enough or the cards having some other issue.

Screet

Screet · Post by **Screet** » Wed Feb 25, 2009 6:01 pm

Since I just made some screenshots to depict the problems with the 4870x2 rendering (not the speed issue) and how the X800XTPE performs correctly, I thought you also might be interested to see them. I didn't rescale these images in order to best show the problems. Sorry if this causes any inconvenience while reading!

These screenshots look to me as the ATI forum entry from a Linux user reporting problems which I copied few messages ago could also apply to my problems with the 4870x2 under Vista. Any ideas?

Many, many thanks to photobucket.com for making it so easy to show you these problems!

Edit: Removed italic and switched to image, stupid me hitting the wrong button

Screet

Baza · Post by **Baza** » Wed Feb 25, 2009 6:12 pm

yes yes and yes same for me 3850 x 2 (crossfire). What exactly causes the pink problem??? oolite virtually unplayable...am awaiting 1.73 with baited breath

Screet · Post by **Screet** » Wed Feb 25, 2009 6:20 pm

Baza wrote:
yes yes and yes same for me 3850 x 2 (crossfire). What exactly causes the pink problem??? oolite virtually unplayable...am awaiting 1.73 with baited breath

Sad to hear that. Looks like they really messed up with those x2 cards

The forum is filling with bug reports from such users...

I am really curious if this problem depicted above is the same reported here:
http://forums.amd.com/game/messageview. ... erthread=y

Solas, since they banned my account for writing the bug reports on their forum and yours still seems functional, do you consider it a good idea, and if so, place a link to this forums thread with the images in the AMD forums thread linked above? Maybe that linux guy could say if these images are matching his experience, thus help to solve this more quickly. That would also put his report up to page one again, thus showing the problem to more users, probably receiving more feedback on this topic.

Screet

Solas · Post by **Solas** » Wed Feb 25, 2009 7:38 pm

No .. I cannot confirm a similarity .. maybe an OpenGL wizard can !
that problem is with a 4870 on Linux and there are no screenshots.
other more likely explanations .. buggy or conflicting OXPs ..
a moved / altered Oolite installation or a misconfigured RunOolite.bat ..
an as-yet-unknown problem involving Trumbles.

Screet · Post by **Screet** » Wed Feb 25, 2009 7:41 pm

Solas wrote:
other more likely explanations .. buggy or conflicting OXPs ..
a moved / altered Oolite installation or a misconfigured RunOolite.bat ..
an as-yet-unknown problem involving Trumbles.

I fail to see how an oxp could have bugs visible under Vista, but not running properly under XP X64. As I wrote, I use the same driver versions (just XP X64 and Vista 32 being the difference), both systems use a fresh installation of oolite 1.72.2 test and both systems use exactly the same OXPs.

Maybe there's really a trumble in my machine, annoying me now for barbequing so many of them

Screet

Solas · Post by **Solas** » Wed Feb 25, 2009 7:42 pm

GTX 295 benchmarks better than your 4870X2.
either working properly will not be the Oolite limiting factor.
However ..
GTX 295 and infinite loop error, Return of the BSOD
http://forums.nvidia.com/index.php?showtopic=88809
No HDMI Audio output on Vista 64bit
http://forums.nvidia.com/index.php?showtopic=90314
an infinite loop error ! .. and the warranty period is how long ?

Screet · Post by **Screet** » Wed Feb 25, 2009 7:58 pm

Solas wrote:
an infinite loop error ! .. and the warranty period is how long ?

The error sounds nasty - and I don't understand why that guy has such a short period to return it, probably some US state?

Here in germany, we've got multiple timelines:
14 days return for everything bought from the internet without need to provide any reasons
6 months (IIRC) for damaged products without need to proof that it's caused by the hardware manufacturing, not the buyer
24 months general warranty (which might require to proof that the product is malfunctioning)

Therefore, I'm still within the period where I would not even have to provide the proof, although I got it

Sometimes I'm quiet happy about living here...sometimes. Were it for the nature and temperatures, I would have to leave for norway or canada, I'm afraid...or this region (italin, but in the region I most like, it's almost entirely using german as public language): http://en.wikipedia.org/wiki/Trentino-A ... 3%BCdtirol

Screet