• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

So how does one OC a 680GTX?

Since the thread got locked ... how does one do this exactly? Going to pick up my card from UPS shortly, so I am just getting ready!

Is it with EVGA Precision now? Or does MSI Afterburner still work?

How does it OC with that boost feature? Or should I just let Boost deal with it, since it will run itself to the 'max oc' anyhow (well above stock speeds)?
 
Have you read about the information there is out there , how it all works.
Understanding of it is evolving.
Yes, almost everyone is using EVGA Prescion.
http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_680/30.html
GPUReview's GTX 680 overclocking roundup!


gtx680oc.png
 

That's a great video, explanation of how this card works. It's very cool sounding software. I'm hoping the new models (more affordable hmm) also use these features. The card will essentially run cooler with this dynamic o/c, you could see the voltage drop during heaven at certain points. Probably when load lightens and the fps would be almost freewheeling upwards, but yet clocks aren't dropping, it just needs less power at certain points. At least how I'm understanding it.
 
Bingo.

This is prolly the first card ever to be overclocked with the fan slider.


I agree, at least with the core. The algorithms are going to keep the core in a place that works as a successful overclock, because it's monitoring all variables. Voltage/ TDP / heat . If something goes out of what's probably contained in a internal workable database. The dynamic o/c will pull back.
If it works as described, sound great !
 
As for me it's awesome and I'm waiting for the custom midrange and its performance/price ratio.

Die hard overclockers in the other hand won't be pleased.
 
As for me it's awesome and I'm waiting for the custom midrange and its performance/price ratio.

Die hard overclockers in the other hand won't be pleased.

I don't see why not. Other than change (yes, people do hate change). The variables can still be controlled. Keeping the temps down will help. So if you can deal with noise or find free ways to make your rig cooler things will go smoother. Right now with the 2-6 pin card , the TDP cap is +132%, (or abouts). Maybe there will be a work-around for that ?
If someone needs something so much % above, the now fastest gpu, then they buy a more specialized custom model.
 
Cuz they like to talk about artifacting, voltages, stability and etc. With this OC method you're taking away more than half the fun about OC.

As far as I know you hit the TDP wall way before the Temps one even with the stock cooler. Custom cooling and water will mean only less noise.

Aside that custom PCBs will allow to overclock higher but Nvidia was clear about not letting users get past the safe numbers in every single Kepler card.
 
You can't just slide the sliders up all the way. It will try to clock itself above what it is capable of doing and crash.
 
I have one.

That is not the case at all.

edit: also, there is nothing dynamic about the ram clock either. It is locked in to whatever you set.
 
Last edited:
I have one.

That is not the case at all.

edit: also, there is nothing dynamic about the ram clock either. It is locked in to whatever you set.

Scenario plox. Cuz on paper the GPU Boost will never push the GPU above it's limits and you can set the offset as high as you want.

Memory OC is different. Never said otherwise.
 
If you set the GPU boost too high, it will most certainly try to clock it that high and crash.

I'm not sure where you are getting this information, but it is demonstrably incorrect.
 
So it crashes if you push the OC limit too high?


Yes, it does.

You can push the power limit to the max and it will only come in to effect if it needs the extra power (because that's just setting your max power usage, and not adjusting the clocks), but if you set the clock offset too high, it will, under load, try to clock up to the max that you have set if it needs to, and your card will crash. You still have to find the max it is capable of for a stable overclock.

The memory overclocks exactly like traditional cards do as well. The base is 3Ghz, and the offset you apply to that is always active.
 
Last edited:
Back
Top