Graphics woes.

[eno]

Once again it seems that I am having graphics card problems. My X1900GT seems to be suffering from something terminal, during a flight artifacts start to appear and strobe randomly over the screen especially in external views where they seem to radiate from the aircraft. In extreme cases the screen goes black and can only be recovered by a restart. The artifacts have begun to appear at seemingly random times into a flight ...... 5mins ....... 2hrs and under all conditions. During "normal" use ie. not gaming there don't seem to be any problems as yet.

If this could be put down to lack of power or over heating then .... whoopee!! I can do something about it.

Suggestions as to the cause would be helpful because I don't want to have to splash out on another interim card before the proper DX10 hardware arrives. (That'll be a whole new system though).

Full specs.
Dell Dimension 4700 (dont worry there isn't much of that left apart from the case and MOBO)
3Ghz Pentium P4 with HT.
Hiper typeR PSU
1GB RAM
X1900GT 256MB PCI-E graphics card (sort of)

[john_uk]

which Hyper PSU do you have?
as isnt their severaly differant kinds of typeR psu's??

[DizZa]

You could try downloading an overclocking utility and back down on the core and memory clocks to see if the artifacts deminish..?

[ctjoyce]

which Hyper PSU do you have?
as isnt their severaly differant kinds of typeR psu's??

Its the type R, or to you Americans, the HPU 580W.

Cheers
Cameron

[eno]

So Cameron ...... what's your verdict on my predicament?

[ctjoyce]

Personally I would grab ATi tool. If your Temps are over 50*C idle, its probably heat related, however ATi tool also has an artifacting test. So if your card passes this test, and your card is running say 40*C then its probably something in FS.

Cheers
Cameron

[Nick N]

Once again it seems that I am having graphics card problems. My X1900GT seems to be suffering from something terminal, during a flight artifacts start to appear and strobe randomly over the screen especially in external views where they seem to radiate from the aircraft. In extreme cases the screen goes black and can only be recovered by a restart. The artifacts have begun to appear at seemingly random times into a flight ...... 5mins ....... 2hrs and under all conditions. During "normal" use ie. not gaming there don't seem to be any problems as yet.

If this could be put down to lack of power or over heating then .... whoopee!! I can do something about it.

Suggestions as to the cause would be helpful because I don't want to have to splash out on another interim card before the proper DX10 hardware arrives. (That'll be a whole new system though).

Full specs.
Dell Dimension 4700 (dont worry there isn't much of that left apart from the case and MOBO)
3Ghz Pentium P4 with HT.
Hiper typeR PSU
1GB RAM
X1900GT 256MB PCI-E graphics card (sort of)

Check the ATi driver settings and see if "Adaptive anti-aliasing" is enabled. Its usually a checkbox in the drivers. If it is, uncheck it, save the driver settings and see if the problem goes away.

[eno]

Adaptive AA is not checked ......... it's beginning to look temp related ..... Idle GPU temp is 58*.

[ctjoyce]

Two words, Zalman VF900-Cu

Cheers
Cameron

[Nick N]

Adaptive AA is not checked ......... it's beginning to look temp related ..... Idle GPU temp is 58*.

Ok wel lthat eliminates the adaptive AA system not working right

My question is, if you are not overclocking how can it be a heat issue unless the tower/HSF/card fans and sinks are plugged with microfine dust


Hmmmmmmmmm

[eno]

That's what I can't understand ...... I've taken the side off the case and that's only lowered it by 1*.

[Nick N]

That's what I can't understand ...... I've taken the side off the case and that's only lowered it by 1*.

Well, unless you have a defective card showing up then what you describe very well could be power or heat related.


Some things to try... run 3DMark 03 in a loop. I suggest 03 because of the load it puts on the card and system. Set it to run multiple passes of the benchmark (like 3-5 loops over and over) so the card heats up really good. I know it would be boaring to sit there and watch but if the errors start appearing I would say you have a heat or power issue. 58c idle.. hmmmm thats not horrible but if its not warm (summer) where you are then its does sound a bit high to me. Low current PSU will warm things up too.


Sorry you are in this situation. In order to find out if you are dealing with a PSU you would have to get one to test it.


Also, ATi Tool has a GPU temp monitor built into the software you can enable or disable. It will display the low/average/high temp of the GPU during use. It may cause game stutters too when it polls the card for a reading. You would only want to use the program (unless you overclock) for checking the temps around the time the errors start showing up. If it is overheating, the PSU is up to the task and the room is not warm, I think you may be dealing with a defective card. There is absolutely no reason for you to change a heat sink if you are not overclocking. The card would not all of a sudden start spittting out errors due to heat after running fine if the system can breath and you are not clocking the card.


The only other possibility I can think of is some type of driver corruption. Perhaps a flush and install of the latest version from the ATi web site may shed some light.

[eno]

Both the card and PSU are less than a year old ..... so if the problem is with either of them they will be going back to the manufacturer

It's tray tools that's showing the temp as an average of 58* ...... however I haven't run it with any load today .

I've updated the drivers ..... I'll give it a try tomorrow with 3D 03 .... and FS.

[eno]

3D-03 on repeat resulted in a max temp of 89* and no problems, the temps also dropped fairly quickly after the tests.

[Nick N]

3D-03 on repeat resulted in a max temp of 89* and no problems, the temps also dropped fairly quickly after the tests.

If the temps remained good and there were no errors during repeated 3Dm runs, unless you have a card or PSU that has a intermittant bug, it sounds like some type of corruption in the sim or a texture/scenery that triggers some type of error.

Now 89c is actually up there but the 1900 series is also known for running warm too. I would still suspect the possibilty of heat factoring in with an 89c reading but I know 3DM03 will run that card much hotter than MSFS