BAD parhelia image problems, reputation damaging!

[103er-Fan]

Originally posted by luni
Agh damn, I thought this was a problem with my setup only, maybe some cables or something, now I see its well known and unable to be fixed .

Hey, just wait, maybe Matrox can find a solution!

[Maggi]

I also see this problem when running PowerDVDs internal system tests, which use several video overlay modes, hence it's not a 3D only problem ...

[Peter Aragon]

Hmm, sounds like there will roll some heads over in Canada, to let that bug go past the final quality tests and into the RTM product.

But I'm confident Matrox will resolve this professionally, not like IBM did with their dud hard-drives, how were those drives called again? (not to mention Fujitsu)
I hope they will come with an official answer soon, this can certainly have impact on their reputation when left unanswered too long.

Peter

[DGhost]

I do have faith that they will correct this issue. to clarify several points in this, they reproduced the bug quite a while ago. My conversation with a customer support person was over two weeks ago (a majority of the conversation occured on 9/20, i can provide specific times too if people want). it was during this where several points came to light...

first was that the problem had already been reproduced, but only for D3D games running in a window...

second is that they were confident that a fix could be achieved through drivers...

third is that (initially) it was of low priority...

forth is that the customer support staff claimed a fix would be released within 2 driver releases, and that Tech support would be advised of the status and plan...

notes on #4... it would appear that these plans changed. as haig has stated the official stance (at least the one that was publically stated) is different from what i was told in the email. so take all of this with a grain of salt.

the distortions are not random. it does deal specifically with what is on screen. and it appears to be a problem when too much information is thrown at the screen at once. kinda like the tearing that happens in 3d games if the frame rate gets too high. except this is much, much lower level.

[Wombat]

With mine, it isn't random. It seems whenever the 3D engine is kicked on, I will get the distortion. Even if I say, fire up UT2003 or something, and minimize it, the desktop is corrupted. It's getting consistantly worse with my card. I'll be really pissed if they don't fix/replace it. It sure looks like a hardware problem though, I'll be mildly surprised if they find a software fix.

[K6-III]

A bios fix may do it, however....

[DGhost]

Originally posted by K6-III
A bios fix may do it, however....

if a software fix cannot be found for this a bios fix probably will not be of any benefit. the bios for the video card only controls it until windows boots. since this problem seems to occur in windows and windows only, a bios update probably cannot fix it.

my fear is that this is a hardware design problem... it looks and smells like one... the fear i have with it is that any software fix that they might be able to create would probably have a fairly negative impact on performance.

[leech]

A BIOS fix would be nice. On my Parhelia, vesa mode bootloaders (like the Lilo one that Mandrake 9.0 loads up with) has an extremely messed up pallete. Psychadelics are one thing, but this is just hard to read. Other things that use the vesa mode are messed up too.

Another weird little thing I've noticed, whenever booting into linux.... it must have some sort of frame buffer problems, because it actually will show a scrambled "Windows XP is shutting down" picture before it re-initializes the screen to bring up whatever display manager I have. I even had it one time, show a Debian picture that hadn't been loaded for over an hour, in Mandrake. (I had Debian, then installed Mandrake 9.0 over it, to test it out, and the scrambled 1600x1200 desktop had some of that Debian wallpaper left in it. Now talk about ODD.) Even after a complete shutdown of my computer, it still retains some of that framebuffer.

I'm not even sure I know what disruption you're talking about, though if it is what I'm thinking, then yes I get it too, but I honestly thought it was just me being tired as all hell, because that's the only time I really see any sort of horizontal disruptions. I did see it VERY faintly when running the PowerDVD tests.

Leech

[Wombat]

he bios for the video card only controls it until windows boots

Not true. Hit your textbooks.

[DGhost]

hehe, i gots no textbooks... i always figured that most of the stuff that the bios does is or can control is either overridden by or passed to the windows drivers to control... guess i was wrong...

OOC, what way does it have for affecting something like that?

[Wombat]

One of the ways BIOS is deciphered is Built-In Operating System. A pretty good explanation. It turns the card in to something the drivers can talk to. It builts the lower functions of the card into an abstraction for the rest of the computer to handle. What exactly it has control of depends on the design though, and that I can't say.

From the way cards seem to be failing, it seems that Matrox probably decided to just ship parts they new were low-quality, or they have really bad QC methods in place. A BIOS/driver fix is a possibility, since I don't know the root cause of the problem, but it seems like a noise issue that is unlikely to be fixed this way.

[xortam]

BIOS stands for Basic I/O Subsystem as I recall. The rest of your description is correct. BIOS is just low level software (many times implemented as firmware) that provides low-level functionality by talking directly to various HW via I/O ports and such.

[az]

I HOPE for them that IF this is a hardware problem, they do NOT call it "normal behaviour" - because that might make RMAs impossible or very hard, and would absolutely DESTROY their High Quality reputation.

AZ

[thop]

if this is unfixable and a hardware problem i see problems for matrox. because they kinda sold you a flaky card and RMA isnt possible since all parhelias have that problem (or so it seems). and if they make new parhelias and recall the old that will cost serious serious money, if not brake matrox' neck.
either way, it will do harm to their reputation.

[d_constructed]

Genlocking missing?

I'm not sure about the exact problem since no-one here owns a digital camera and can grab a shot of the problem, or grab a simple screenshot and FAKE the problem in Photoshop to explain what's going on.

I'm guessing it's a missing genlock between the outputs, expensive 3dlabs WILDCATS have this, when using multiscreen OpenGL you don't get weird banding on the second screen, ever.
Has anyone tried to switch output to the main one which might be better? I guess the secondary and tertiary are non-genlocked slaves that have this issue.
Genlock requires new hardware, so no BIOS-fix in the world help, but I'm probably wrong anyways since I base all this on hear-say.

Or it could be another non-fixable hardware issue, I got nvidia cards and they are full og them The DXTC1 problem in GF1 wasn't fized until the GF4 for example. Are there any gfx cards out there without any hw issue? Don't think so, but since you can buy a 2x faster card every year, we need to think that gfx cards are subscriptions, a service, that gets better every year.