135 replies to this topic

[litwicki22]

Hey guys. I will be very greatful. Can somebody make for me 5 loops from a row in 3dmark 11 in performance mode?
I ask because i have very difference results, between 9391 and 9463 points. Too much variant.
If somebody make 5 loops and paste here scores ok? Please , please. Cheers.

pc:
3770k 3.5 stock
Msi Twin Frozr Gtx 680
8gb ram
301 whql



Please have the same clocks on cpu and gpu on that loops.

Edited by litwicki22, 07 October 2012 - 04:58 AM.

[Knight_72]

Are you that guy I remember that was over-compulsive about the variations in scores ?

Hopefully someone will post their scores so that you can stop freaking out about it

[Waco]

That's like .5% variation. There's nothing wrong with your scores.

[litwicki22]

OK nodoby have Gtx 680 and cant past here some scores:(?eh

[vandreadstriker]

Are you that guy I remember that was over-compulsive about the variations in scores ?

Yes he is..

OK nodoby have Gtx 680 and cant past here some scores:(?eh

Yes I don't have a 680 and so I can't post any scores... But that isn't the problem really.
Pardon me if I'm being rude but the problem is you're freaking out over score differentials that isn't even considered as "Wrong" (if you can consider a score as "wrong" that is, in which case, I don't think so).
Even big companies write something like "within +/- xx% error" or math questions with something like "answers within 0.x difference will be accepted". There's nothing wrong with it, is there?
I'm not saying that you shouldn't worry about anomalies. In fact it's good that you pay attention to them which shows your great attention to detail (very helpful with jobs I tell you). However, you're once again asking for a solution to something that we've told you not to be worried about. We've told on your previous thread that those differentials are to be expected and there's nothing wrong with your card.

Sorry for the rant but it's late here and I'm stressed out over tomorrow..

[Muchoman1]

Once again

With ALL 5 OF MY KEPLER CARDS, THERE IS ONE THING THAT ALWAYS HAPPENS

The speed of the card is very reliant on temps, so running multiple tests one after the other, will result in higher scores initially due to the card starting cooler

If the room warms slightly due to the new heat source (running card) this can lower the boost a few mhz and result in different scores

Maybe a backround process used a little more resource one time, and it is CPU related


If you were having variations of around 1000 points, then worry, if not, there is no problem at all

[El_Capitan]

You have a big problem with your GTX 680. Sell it to me for $300, and buy another one.

What is your GPU Core Clock and Memory Clock, and what is your max Core Clock when it boosts?

Edited by El_Capitan, 07 October 2012 - 08:19 AM.

[Muchoman1]

I just had a look at my local forums 3dmark 11 thread, those scores are very normal


Please accept this, otherwise we will have to let El_Captain confiscate your card

[EuroFight]

You have a big problem with your GTX 680. Sell it to me for $300, and buy another one.

What is your GPU Core Clock and Memory Clock, and what is your max Core Clock when it boosts?

$300? For a card with THAT much variation? You must be insane. $10 would be more like it

[El_Capitan]

$300? For a card with THAT much variation? You must be insane. $10 would be more like it

I see what you did thar!

[_TheAlexO]

P9420 w/ gtx680 and 3570k all stock

[Coors]

I just ran it three times on my desktop and got:

9303
9285
9289


As you can see it's normal for there to be slight variances. Just like an olympic runner isn't always going to have exactly the same time, or a drag car isn't going to always have the same time...these synthetic benchmarks aren't an exact science and 1+1 doesn't always = 2. As others have said your system is performing like it should. Stop worrying about meaningless benchmarks and play some games.