Video Card Burn/Torture Test?

GreatestOne

Limp Gawd
Joined
May 15, 2005
Messages
488
We all use Orthos or Prime95 for torturing the CPU.... but is there any program more exclusively designed for the GPU? I think there was one maybe in the ATI tool? but it seemed lame... I use rivatuner and i wanna try to squeeze some miles out of my 8800GT in running COD4 and the game was crashing whenever i tried running it at 1680x1050 (their "optimized res was 1280x1024 i think)) so I am suspecting the video because my CPU at 3.2 OC (from 2.66 conroe) was rock solid in all testing through Orthos and Prime. Thanks guys.

P.S. If I cant run 1680, is it ok on a 22" wide to use the laptop res of 1440x900? or 1600x1200?
 
If you have the game, loop the Crysis benchmark. It finds instabilities better than anything else I've used.
 

Beauty... awesome.

But one thing, wont ANY card or even SLI fail pushed with the xtreme burning? Is there any benchmark on how long to run that, ie, we run Orthos for 24 hrs we pretty much say its solid right... i am assuming no card will last 24 hrs on xtreme, so... # of hours maybe?
 
it shouldn't, furmark limitation is that it doesn't work with SLI (at least I don't think it does
 
So u saying I can run furmarks extreme testing for 2 days and my card wont fail?

I don't know but I don't know why it would not, another more productive way to test your card is to fold@home.

I have run furmark for 4 hours before to test a video card. worked perfectly.
 
If I run a benchmark at 1680x1050 and it gives say, 40fps avg, does that mean my rig (or JUST my video card isolated?) can avg that fps with the toughest rendering parameters? I.e., is that equivalent to running Crysis with everything on?
 
If I run a benchmark at 1680x1050 and it gives say, 40fps avg, does that mean my rig (or JUST my video card isolated?) can avg that fps with the toughest rendering parameters? I.e., is that equivalent to running Crysis with everything on?

depends on what benchmark your running. generally benchmarks are worthless at predicting real game playing. For instance a 9800GX2 will beat the GTX280 in a lot of benchmarks, but can't touch it in actual playing. In other words your going to have to try it. I suggest you download the demo version.
 
Furmark isn't very useful for testing your GPU overclocks. It's pretty good at detecting RAM instabilities, but it just catch the core problems in a decent amount of time (<2hrs.). For instance, I can run Furmark at 870 core on my 4870 on Furmark for awhile, but it crashes midway through the first loop in the Crysis benchmark.
 
Does the crysis demo have a benchmark and torture test built in?

And I know all games perform differently with diff cards, oc's, etc., but I assume if it runs ok with crysis then it should be gold for all the others from the last year correct?
 
it shouldn't, furmark limitation is that it doesn't work with SLI (at least I don't think it does

It works with SLI.
I had SLI'ed 7950GT's at 99C for 2 hours with the almighty Fur. :eek:
 
I ran ATitool artifact scanner with most sensitive settings, and Furmark from anywhere between 8-10h to test stability of my overclocked GTX 260. Both held up till the end and only stopped when I manually stopped it. I then looped the Crysis benchmark 99 times without failure.
 
I ran ATitool artifact scanner with most sensitive settings, and Furmark from anywhere between 8-10h to test stability of my overclocked GTX 260. Both held up till the end and only stopped when I manually stopped it. I then looped the Crysis benchmark 99 times without failure.

Yes sir, still using ATI Tool as well.
RT3dhDR works well, but ATITool is perfect for testing
 

+1, and if your looking to test your cooling there is nothing better.

Beauty... awesome.

But one thing, wont ANY card or even SLI fail pushed with the xtreme burning? Is there any benchmark on how long to run that, ie, we run Orthos for 24 hrs we pretty much say its solid right... i am assuming no card will last 24 hrs on xtreme, so... # of hours maybe?

Furmark is not really that good at telling you your max gaming overclock. Furmark will just allow you to push your core clocks way past gaming clock stability so you're going to have a hard time figuring out what is your max stable gaming overclock is. For example I was able to max my core to 885 and that was furmark 15 minute stable in 0-8xaa testing modes. However whenever I tried to game that overclock crashed right away in cod4 or crysis. Then I tried atitool and the max clock I could get 15min stable was 860. Which turned out to be my max gaming clock as well and I had no problem playing through multiple cod4 missions at all at that core clock.
 
Best stability tester I have found is Crysis. For thermal testing, as in testing your cooling solution, Furmark gets things mighty hot. Hotter than Crysis.
 
I ran ATitool artifact scanner with most sensitive settings, and Furmark from anywhere between 8-10h to test stability of my overclocked GTX 260. Both held up till the end and only stopped when I manually stopped it. I then looped the Crysis benchmark 99 times without failure.

It seems like u have the right idea in a system setup (maybe except your cpu).... may i ask what u get on your 3dmark06 score?
 
Furmark is not really that good at telling you your max gaming overclock. Furmark will just allow you to push your core clocks way past gaming clock stability so you're going to have a hard time figuring out what is your max stable gaming overclock is. For example I was able to max my core to 885 and that was furmark 15 minute stable in 0-8xaa testing modes. However whenever I tried to game that overclock crashed right away in cod4 or crysis. Then I tried atitool and the max clock I could get 15min stable was 860. Which turned out to be my max gaming clock as well and I had no problem playing through multiple cod4 missions at all at that core clock.

Thats makes sene to me now.... so u basically test with the ATiTool for 15 min bec thats the avg time it takes for each mission on the relevant FPS games?
 
I don't use furmark for overclocking/stability test but it works great at finding the worst case of (over)heating.
 
Thats makes sene to me now.... so u basically test with the ATiTool for 15 min bec thats the avg time it takes for each mission on the relevant FPS games?

Well I only bothered to run it for 15 minutes, some people prefer to run their atitool tests even longer but 100% load for 15 minutes is a good amount of stress on the core overclock.
 
It seems like u have the right idea in a system setup (maybe except your cpu).... may i ask what u get on your 3dmark06 score?

I'm sorry, I don't have a 3DMark06 score at the moment, but I will give it a run and let you know.

As well, what is wrong with my CPU in my setup?
 
Back
Top