3DMark06 and RivaTuner? Whats better?

[CrossStealth]

With Rivatuner ive changed my
Core Clock--------ShaderClock
698-----------------1589

This gave me a score of 14818

the lower clock of
681-----------------1551

This gave me a score of 14988

Why is a slower clock getting better results? Should I keep it at the best score? or should I use the fastest successful rate?

I know it doesnt matter in this case but Im overclocking a GTX 285 and running a Q9550 cpu

 

[calinkula]

I would keep the lower speed. A higher score with a lower clock shows that the hardware is spending time on error correction with higher speeds.

Do you have an OC on your Q9550? You may see some better gains overclocking there if you haven't yet.

 

[CrossStealth]

I havent overclocked anything at all yet. To be honest I have no idea how to overclock anything. I read some guides but still shaky about it all. Video card OC seemed to be the easiest so thats where I started. Should I stop this OC and do something else first? All I know is I can OC CPU Graphics card MB and ram but dont know in what order or even how to begin since I dont know a lot of terms used in the OC process.

 

[Silverion77]

Truthfully...thats not like an amazing OC and gives almost 0 results

Also, all 3D mark runs provide different scores in a Standard Deviation from each other. Run 10 different ones and they'll all be different and 170 isn't a whole lot...
On top of that, 3D Mark is a bench and in no means a real world result

In terms of performance gain the order is: CPU (mobo OCs with that), Graphics and Ram (wouldnt really bother with ram)

 

[CrossStealth]

its only partially overclocked I didnt touch the memory clock yet. Im in the process of finding a good number for that now. If you dont test your score with 3DMark06 how do you get your vid card OC numbers? Do you just guess or whats your personal method?

 

[Silverion77]

????

Stress test it is the only proper way to see a stable overclock
Prime for CPU
and ATI Tool or Furmark for GPU

 

[CrossStealth]

i think I misunderstood your last post. So I should ditch 3DMark06 and get ATI tool and continue the OC process in your opinion?

 

[Silverion77]

3D Mark is a bench to see computer performance, but if ur looking for stability you dont use that. U want a graphic stress tester like ATI Tool or Furmark to see if the OC is stable and the card isn't overheating or causing artifacts.
After its stable u can run 3D Mark for a performance score

 

[CrossStealth]

Oh ok I have FurMark since I have Vista 64 and wasnt sure if ATI was compatible or not. I just realized I have no idea how hot this card is supposed to be though.
Whats the max temps this thing should reach after like 15 mins on Furmark?

 

[Silverion77]

idk personally...

Generally heat is only an issue when artifacts pop up. Thats when the card is 1) unstable so drop back the clocks or 2) too hot and u drop back the clocks

 

[jitpublisher]

The difference you have noticed is nothing...nil....don't worry about it. There are many things that can make the benchmark vary greatly that have absolutely nothing to do with the overclock you are trying for....for instance Cool and Quiet with AMD, or Speed Step with Intel.
My benchmark with Cool and Quiet on about 8900, and it varies about 500 points everytime you run it! With it off, slightly over 15,000 everytime.
What programs do you have running in the background? Remember, this benchmark tests more than your just your GPU performance, it also tests CPU performance. Any number of background programs running can change the score a few hundred frames everytime you run it.
To get the best estimates, you should run the benchmark at least 3 times, and get an average score to look at.
And, make sure you don't have anything running in the background when you run the benchmark.