New
#71
Proving once again what I have been saying about Prime Blend all along. It's the real deal for stablity testing. 5 minutes of Prime blend and you are far from stable my friend, not even close IMO. 10 hours of Prime Blend and you will have something. It takes a lot more than Vcore adjustments to get a system stable.
Ditto. They're good if you are going for ultra-quiet setup (like a pure HTPC etc) but even at full tilt, the noctua fans are nice and quiet.
nah mate, don't give up. Just take a break - read a bit more and you'll be tinkering again in no time
It's quick and easy to ham fist a working OC - but it takes time to refine.
As for 'stability' - when overclocking, there is no such thing as an absolutely 100% stable system. However you can come pretty damn close
Every overclocker has their own methods and tools that they use. Usually it's a battery of tests. It comes down to personal preference.
Often, you can start with 'loose' RAM timings to dial in your CPU O/C - then once that's sorted, you move on to 'tightening' the RAM. ( potentially eliminates initial RAM or vcore issues)
Generally, apps like IBT, LinX etc are good for 'quick and dirty' vcore testing. Make an adjustment, run for 5 passes - if all good run for 20+ etc.
Then you move onto Prime95's blend (torture test) if that passes the 12-24hr run and you have already passed IBT, LinX etc - then you are pretty much good to go.
But if you fail on Prime, then whatever FFT test you fail on gives an indication where the problem may lie. Low FFT failure is generally vcore/CPU related. Large FFT points towards RAM.
(However it's worth noting that some can pass IBT but fail Prime, or conversely pass prime but fail IBT etc)
The last and most important test is simply real world usage. If you have minimal, or preferably no BSOD's, freezes etc - then you can call it 'stable'