Overlocked 5970, Furmark FPS inconsistent.

[Oatmealz]

Okay on stock 5970 725/1000:

Furmark stability test , 640x480 NO AA: I get 329 MAX FPS..

Overclocked to 900/1225 w/ increased GPU Voltage (not memory): I get 368 MAX.

Overclocked to 906/1251 w/ increased GPU Voltage (not memory): I get 360 MAX.

The question is: Is Furmark FPS reliable for resting overclock or should I stick with 3d Vantage?

Thanks in advance.

 

[kokin]

Why would you only look at max FPS? The goal of benchmarking is the average FPS. As long as that is consistent, that's all you need.

Max/Min FPS could have only been present for a second and never reach that high/low point ever again for the whole test. So again, as long as you see an increase in avg FPS, then keep going at it.

 

[Oatmealz]

Sorry for the reply. Thanks, PSU just died on me and had to get a new one.

 

[vvhocare5]

Also, there are several protection mechanisms built into the 5000 series cards. As one clocks higher and higher it seems as if the card slows down - there is a wall.

You didnt mention heat levels on your card - 900/1200 is pretty high on that card on air.

Many overclockers up the voltage/clocks and see what is happening to the various benchmarks - not just one. As soon as they flatten they have reached the cards limit and either stop or back off a little. People on air will see those limits much sooner - air simply cannot get rid of that much heat fast enough - even with that fan at 100%

 

[lashton]

640 X 480 why would you do that resolution NO ONE else benches at this resolution, try 1240 X 1024, and what you want is average NOT max, besides thats not really an oveclock its like 2%

 

[overshocked]

Graphics cards drop in performance if you push them past their limit.

 

[cjl]

Furmark isn't a good benchmark. Try something else.