So i just got my 680 last night and noticed that unless i enable prefer maximum power in the nivida control panel the card downloads to like ~300 while i am in game. Is this normal? Do we have to enable that or is there something i am missing? Once i enable it, its fine but id rather not run at full clocks all the time lol. 
 
New to the 600 Series not computers ;)

Is it causing any problems when it down clocks, shutter or lag?
Anyhow you wouldn't think it would while gaming, I haven't played much BF3 with these cards yet I'll have to keep an eye out.
If the cards are down clocking maybe turn the settings up. In the nvidia control panel>manage 3d settings>BF3 turn the Anti Aliasing Transparency to multi or super sample X2 - X4 - X8. 
 
EDIT: And turn off post AA in game, 680's don't need Post AA... 
 

  
  
 

Even at 60c don't they down clock a little?

Does the problem happen upon a cold boot into windows? I had the same problem and it would go away after my pc was on for about an hour. I'm currently RMA'ing, another person on this forum had the same thing happen as well.

Ok so it looks like im also effected with the cold boot issue. Can a moderator comment on this to let us know if this is a known issue?

Try uninstalling Precision (and all saved information and profiles)
Reinstall it, and leave -everything- at defaults.
 
Make no adjustments to the power target, memory or fan profile and do NOT enable the OSD function.
Now test again in BF3, after a few miuntes in-game... leave and then open up Precision to view the test results
 
Could be a borked PT value, OSD bug... you have to rule out Precision before moving forward.

I hear it's good to keep them at 69c max temp to avoid down clocking,. which I have already done but must test in games as I have only tested in furmark's burn-in test.

Ok so after  some searching it appears that im not the only one plagued with the issue.
 
http://forums.evga.com...mpage=1&print=true
 
http://forums.evga.com...mpage=1&print=true
 
and a few more on google, strangely its only this card that keeps popping up... hmm

It appears that several other people are now being plagued with the same issue.
 
http://forums.evga.com/tm.aspx?m=1659153
http://forums.evga.com...mpage=1&print=true
http://forums.evga.com...mpage=1&print=true
 
Im trying to compile all the issues so that we can get someone to look into this. If you have a similar issue please let me know! :P
 
My card is a mad overclocker, but having to wait 30minutes to an hour to play kind of negates that lol!
 

I agree, perhaps they need to implement something in precision that takes all this "energy saving" stuff away ;)

Well that sucks! It seems kind of silly to limit what they can and cant do with their technology. Its like owning a sports car and never taking it out of first gear lol.

northfaceseen

Hey,

My 2GB FTW card was downclocking to 324Mhz in 3d apps and my Power % was hitthing around 275% when my card was doing it.  I tried both the 301.24 and 304.48 drivers and both drivers gave me the same problem.  As another poster commented, I noticed my card started to work normal after my system had been running for awhile.  I ordered 2 FTW cards and only 1 card had this problem.

 
I agree 100% with this behavior.  One of my gtx680 4GB FTW cards behaves exactly like this.   I tried  301.24 and 304.48 drivers, swapped the cards in SLI and tested each independently.  One of my cards downclocks to 324Hz under load(even at very low temps) and shows the power spike up to 275%.   Like others have stated, the problem is most apparent when the system is first turned on for at least 30 minutes, but I do see some symptoms even after an hour.
 

Do they still downclock when you manually set the voltage of the card through precision?

hmm, must be an issue, mine stays where i set it. Probably firmware update they will need to release to fix that.

I would just RMA the card because I don't think it will be fixable by the user.  I've already dropped off my RMA.  Obviously there is some component that is related to the power monitoring that is defective.    I'm sure that the downclock to the idle clock while underload is a result of the monitoring software seeing that the power usage is 275%.   I'm fairly certain that the power usage is not accurate.