I'm seeing reviews of Gigabyte motherboards where they struggle to reach 4.8GHz with a 2500K. I'll admit I haven't been paying too much attention to this generation of cpu's, but as I recall a 2600k is supposed to have a max multiplier of 57, for a max clock of 5.7GHz.
My experience with 1156 Gigabyte boards has been that they can go faster than As.s and other boards. Are the 1155 Gigabyte boards different?
What boards support 5.7GHz clocks on a 2600k?
I suppose you'd have to find someone who is fiddling with both the ASUS and Gigabyte boards. But I've already seen comparison reviews where the result is neck and neck. For instance, if I'm not mistaken, per the ASUS and AsRock Z68 boards. Also, I would think that the Z68 chipset itself is a step up from the P67 and an improvement upon it. I"ve seen less posts on these forums from people who worry about some malfunctin of the ASrock boards as opposed to the ASUS or Gigabyte, but small samples don't tell a lot, and problems can develop from cases where the consumer/enthusiast doesn't know -- or isn't sure -- what he/she is doing.
I'm also skeptical about the hype, only because I'm a conservative over-clocker looking for a "sweet-spot" that's just within a "safe" voltage range and the thermal throttling spec. I have observed profound confusion among the Hot-Dawg OC community over the specs for the "K" chips posted by Intel. Unless I didn't look hard enough, this is the first time -- the first generation of chips I've looked at -- in which Intel seemed to omit two specs: the "safe" voltage range, and the "operable" voltage range. It might seem that the VID range posted by Intel was the operable range, and the same impatient OC community has certainly proven it.
The reviews have been written by people who are simply running benchmarks to see how high the chip will go. There seems to be a tongue-in-cheek acceptance that voltages of 1.4V and even higher are acceptable, as the reviewers push the CPU over 5.0 Ghz. One young lady enthusiast, trying to post her own OC'ing guide to fellow forum members at another over-clocking site, began her dissertation by stating "Step 1: Raise the VCORE to 1.45V." If she wasn't just doing a quick benchmark, I'll take bets from people about the number of months before she either attempts to RMA the chip or simply replaces it.
And all these reviews show voltages for more modest over-clocks with VCORE settings which may themselves be higher than necessary to sustain those clocks as stable.
Meanwhile, I've seen people posting threads on this forum who are worried about their temperatures -- their core temperatures -- breaking 80+C @ 4.7 Ghz. They're voltages also seem to be pushing the limit. I'd never seen a review or heard from anyone who wasn't running their VCORE at 1.37+V to 1.40V at the 4.7Ghz speed, even though I'd seen someone who had reached 5.0 with the 1.40V VCORE, on the same system and board-- which happened to be the ASUS P8Z68 [ . . . ]. The individual enthusiasts posting on forums were raising worries when suddenly they were having trouble with the entire range of temperatures on those CPUs, but again, some in the "community" had advised to "keep your temperatures below 90C." The thermal throttling spec for the K chips is 72.6C.
At the Intel community forum, another individual posted a thread with the same developing problem about temperatures. In the ensuing interaction, he noted he'd been told the 90C myth, and the respondent -- possibly and even probably an Intel employee -- asked with impatient animation: "Who told you that you could run these CPUs at load temperatures of 80C?"
The Nehalem chips were also 32nm silicon, as are the Sandy Bridgers. Since the Nehalem safe-range limit had been spec'd at about 1.37V, I think that's probably the case for the Sandys. I've paid more attention to the "advice" of reviewers or aspiring OC-guide writers who suggest that 1.30V is a good target.
For 24/7 operation with air cooling, I'm optimistic about getting to 4.6. If I can't, I'll console myself that the performance in Giga-flops seemed to peak at 4.43 and dropped off by a single Gigaflop at 4.53. Or better, it's still the fastest thing I've seen at only 3.91 Ghz.
So I'm not so sure that any board supports the speed you mentioned unless it uses subzero, phase-change, liquid nitrogen cooling. Then -- under those circumstances, it may well be any such board. Even so, I think I saw that with such a cooling system, the voltage was still pushed close to 1.6V. But you're not likely of the singular inclination to invest in such cooling, and I'd think you'd worry about electromigration and chip-death from overheating.
At this point, with only my ASUS board as reference but from what I've seen second-hand on this and other forums, the practical limit for 24/7 operation is defined by the CPU, it's voltage and thermal limits.
Others my rebuke my thoughts here, but this is my assessment so far.
Facebook
Twitter