Hi guys. I have a strange problem in 3dmark 11 and my card Evga Gtx 680. If of course this is a problem. Listen very careful.

I have a pc:
3770k 3.5
8 gb ram
Be Quiet Dark Power Pro 1000W

Windows 7 Ultimate x64
301 WHQL drivers



On my friends card Msi Gtx 680, in my pc ( THE SAME PC ), i was have always from 9320 to 9340 points. Just always even on 10 loops. On my Evga Gtx 680 i have between 9264 to 9388 points ( sometimes 0,5 fps to 1 fps different on the same tests ). Dont worry about that? My card is fine even on that differences on scores?
I close all background programs. Both card tested on my pc. But i dont know why my card making from 9264 points to 9388 points in for example 10 loops. While Msi Gtx 680 is making always from 9320 to 9340 points.
Something wrong with my Evga Gtx 680 ? HELP ME , thx. I want to decide to RMA or not.
 
 
 
This is my second after RMA Evga Gtx 680: ( For me its too much 100 + points difference in loops, while on Msi Gtx 680 is only  20+-.                            GPU score is different i dont know why.
 
 
I upload a screens , look down.

 
 
Sorry again for my bad Polish:/  I dont know what to do now.

Again?  Last time it was your friend's asus card.

I do not see the results as outside of expectations. It's very close and many things can affect it in this manner...disk access, different programs in the background (or services) accessing computer resources during the test....

*shrug* Seems fine to me
 

I would consider anything between a specified range of 500ish points as ok.

I think your expectations of stability of scores is unrealistic.
 

First, what version 680 do you have and which model MSI card does your amigo have?
 
You and your friends PCs are exactly the same?, same memory and timings, CPU and GPU clocks, software, drivers, PCIe slot population, everything? That is really unusual and kind of creepy
 
In any event, I can sneeze and get a 100 point difference, not intending to be negative, but it sounds like you might be making a bigger deal out of this than it really is.  The situation you described is definitely nothing to justify an RMA, "An EVGA Return Merchandise Authorization (RMA) allows customers to return defective products to EVGA for a replacement under warranty."
 
A 100 point difference in a benchmark does not imply your card is defective in any way and means absoutely nothing in regards to the proper operation of the card.  

MSI makes more than one model of GTX 680 as does EVGA, what specific models are you referring to here?

Card of a friend is:
Msi Gtx 680
 
http://www.msi.com/produc..a/N680GTX-PM2D2GD5.html
 
 
And now look at my scores  on my Evga gtx 680  :

 
and this is loop from beggining of the day:
 

YES IT is MY model :
http://www.evga.com/produ...-KR&family=GeForce

Right forget the benchmarks for a second.
 
Are you having any problems playing games?
 
If its a no, then you are fine.

It is normal, get over it. 
 
As mentioned before, the clock generator in the computer doesn't produce the same frequency every time due to occilations, so each run will have a slight variation to the score.  In fact, we buy rounded up frequencies, no one buys a 3.49GHz quad-core or a 3.51GHz quad-core, they just buy 3.5GHz.  Same with your RAM and graphics frequencies, and internal QPI and PCIe frequencies, they aren't the same all the time.  The variation you're seeing in score pattern accounts for less than 2% and you're making a big deal about it.
 
The answer to your question is "Clock Generator".  Look it up, calm down and enjoy your hardware.  When we benchmark, we just pick one to represent our efforts. If you want to do an analysis of the variation, go ahead, but it isn't new.

If you don't like that answer, I suppose you can keep asking.  But there won't be anything you can do in BIOS or in Windows to force the score to be exactly the same everytime you run the benchmark. 

It's nothing man, chalk it up to the way the benchmarking software works and move on.

Again, and not meant as a negative comment, but you were making an issue out of something that almost any other computer enthusiast here would consider 100% normal.

Stop worrying and enjoy your rig

Quoted from above to save time

lehpron
It is normal, get over it.

As mentioned before, the clock generator in the computer doesn't produce the same frequency every time due to occilations, so each run will have a slight variation to the score. In fact, we buy rounded up frequencies, no one buys a 3.49GHz quad-core or a 3.51GHz quad-core, they just buy 3.5GHz. Same with your RAM and graphics frequencies, and internal QPI and PCIe frequencies, they aren't the same all the time.  The variation you're seeing in score pattern accounts for less than 2% and you're making a big deal about it.

The answer to your question is "Clock Generator". Look it up, calm down and enjoy your hardware. When we benchmark, we just pick one to represent our efforts. If you want to do an analysis of the variation, go ahead, but it isn't new.

If you don't like that answer, I suppose you can keep asking. But there won't be anything you can do in BIOS or in Windows to force the score to be exactly the same everytime you run the benchmark.

wdflyer
It's nothing man, chalk it up to the way the benchmarking software works and move on.

Again, and not meant as a negative comment, but you were making an issue out of something that almost any other computer enthusiast here would consider 100% normal.

Stop worrying and enjoy your rig

Wait a bit - I am sure you will see similar opinions show up in here as well.


I have had one of our rigs score as low as 9617 and as high as 11848, why?, because variations are normal.

Your obsession with the benchmarking numbers is a tad excessive my friend, in the end they are just numbers and only good for bragging rights. What matters most are things like real world performance measured in, as an example, FPS while playing high end, system demanding games with all the graphics settings maxed, IMO.