Testing a second hand Graphics Card?

[calza]

As per the title really. Hoping to pick up a graphics card tomorrow that is second hand - and wanting to test it out make sure there is nothing untoward.

Only idea i have is run some kind of stress program and make sure the temperature is ok and nothing unusual happens on the screen. Is this as good a way as any?

If so what program and what type / length of test would work best?

Cheers!

 

[ak47jar3d]

Second-hand? I would say download Heaven Benchmark 4.0 and let it loop for three hours. If there's is anything wrong with it, it will be very evident. http://unigine.com/products/heaven/download/

 

[clutchc]

If you think the card may have been abused, run Furmark for the default test. Watch the temps. If it passes, there's no sense abusing the card any further. One thing I like to do when I get a used card, is replace the thermal paste between heat sync and GPU. If it has been run hard for over a year, there's a good chance the old stuff may have lost some of its thermal conductivity.
http://www.ozone3d.net/benchmarks/fur/

 

[calza]

I've been told it was ran in SLI (760).

I could run furmark (since i have it installed), then just game myself for an evening and go from there on the assumption if it hasn't broke it's as tested as it can be?

 

[clutchc]

I've been told it was ran in SLI (760).

I could run furmark (since i have it installed), then just game myself for an evening and go from there on the assumption if it hasn't broke it's as tested as it can be?

That may be the best thing to do. Run it through the normal paces and see if any artifacts shows up, or any other issues arise. If the card happened to be the #2 card in the SLI configuration, it probably got the least use of the two cards.
I wouldn't run an extended test with Furmark. It can be brutal. Personally, I usually just run one of the preset benchmarks.

 

[calza]

I actually think I might be able to get both. Chances are one has had less load put through it then?

'Burn-in Benchmark 1920x1080 15min' - is that what you would recommend?

 

[clutchc]

I actually think I might be able to get both. Chances are one has had less load put through it then?

'Burn-in Benchmark 1920x1080 15min' - is that what you would recommend?

Not me personally, I usually just run the "benchmark preset: 1080". I don't like stressing my cards that much. I'd rather do it gaming.

The top card in an SLI/CF setup always gets more use. During 2D work and SLI/CF-unaware gaming, the bottom card is always at idle. Of course, the user may have swapped the cards around to balance the wear and tear... like I did with dual card builds in the past.