I would try it without overvolting and see how far that gets you. You may or may not need to overvolt. .Leave the Memory Clock Offset be for now and just concentrate on raising the GPU Clock Offset.
In the NVIDIA Contol Panel, go to "Manage 3D settings" and set "Power management mode" to "Prefer maximum performance" and reboot. Now your GPU will be running at it's Base Clock. Increasing GPU Clock Offset from there will increase the Boost Clock. You can use GPU-Z () and run the Render Test (click the little "?" icon beside "Bus Interface") to quickly see how the Base Clock and Boost Clock are behaving without having to load up and run a benchmark or game.
One thing about Precision X is that you won't see a GPU overclock of your card until you change the GPU Clock Offset by 13MHz. In other words, you can only overclock in 13MHz increments. So if you raise the GPU Clock Offset by, say 10MHz, and don't see any change in clockspeed, that's why. Or if you raise it by 20MHz, you'll only see a 13MHz increase.
The maximum GPU temperature spec. for your card is 95C (). I would unlink the Power Target from the Temp Target and set the Temp Target to 85C and the Power Target to it's maximum. Any time the card hits either the Power Target or Temp Target it will begin to throttle down. You can make a custom fanspeed curve so that it ramps up quicker than stock to promote cooling. Increase the GPU Clock Offset 26MHz or 52MHz at a time and run a benchmark like Unigine Valley or Heaven () to test it. Use On-Screen Display for monitoring. When you get to a point where freezing or large amounts of artifacts occur in benchmarks or games, reduce the GPU Clock Offset by 13Mhz and try again.
When you achieve your core overclock, you can return Power management mode to "Adaptive" if you want to which will enable the power saving features.
When you get the core overclock where it's stable, work on the Memory Clock Offset in the same manner.
Hope that helps.
Facebook
Twitter
Instagram