So I've recently got a new cooler ( H110I GT) and managed a 4.7Ghz stable overclock on stock voltage, I tried 4.8 but eventually got a failed core, I boosted the voltage up to 1.4 and am currently at a stable 4.8Ghz. Max socket temp was 48C while max core temp was 35c. Am I wrong in thinking I've gotten extreamly lucky in terms of the silicone lottery? - Chip was bought before the 9xxx series chips were released.
Honestly, most of the FX chips are golden as far as OCing goes. I took a 6300 from 3.5 to 4.5 and it's totally stable. If you managed to get an 8350 to 4.7 on stock voltage... Jeez. Tell me if you manage to get any higher than 4.8. I was thinking of getting an 8350 if I could get it to 5.0
How stable is what you're measuring as 'stable' - does it pass hours and hours of benchmarks and stress tests? Max core temp of 35c..... proof please. Without a sub-zero amient temperature there is no way in Westeros that a FX chip after stress tests/benchmarks can be that low.
They are safe to 1.55v, so 1.46v is not greedy at all. When I first pushed to 5.1GHz my chip was gaming stable at 1.47v but very quickly needed 1.52v to remain stable during extended (72+ hours) use in multiple scenerios.
I know they are safe to quite a high voltage. What i mean by greedy in terms of their power consumption is the amount of power in to performance out just hits a very steep exponential curve.
lol my 8320 on the max would go to 4.6 ghz :P, on stock voltage i managed a 4.2ghz max not bad for a 8320 IMO, but dang 4.7 on stock voltage is a nice overclock. i dont know if a 970 series mobo are any good for overclocking.
I don't have any screen shots of it at 4.7Ghz, as its at 5Ghz stable now. Although I overclocked by bumping up the bus frequency rather than the normal multiplier method due to it being better on the single cores.
Northbridge changes the frequency of other things as well, compared to just cpu multiplier. FSB overclocking usually yields better performance per clock than just multiplier ocing.