I'm testing Kingston's Hyper-X3200 2GB kit, and was shocked when I easily booted into 240FSB (1:1), then 245FSB, then skipped to 255FSB, (all at 2.85V). 255FSB wasn't stable enough for 3D benchmarks, although I did run Sandra getting 6013MBs Buffered. 250SFB was stable, but at that frequency I run into voltage issues with my 3.0E, which increases instability.
245FSB (1:1) is most stable, this is with PAT enabled, running on SPD (PAT enabled in Advanced Chipset menu as well) again at 2.85V. If the Asus P4C800E-dlx could muster a little more VDIMM I can only imagine 260FSB would be easily attained.
i also want to share something about the Socket-478 3.0E Prescott, which needs additional voltage to operate above 245FSB. I've found Prescott, does better overclocking below 1.450Vcore. Anything above this creates more instability then it does increase spped. So at or above 250FSB Prescott becomes a quandry. I still beleive people haven't begun to realize (and probably never will with motherboard makers failing to modify the motherboards CPU power circuitry) it's true overclocking potential.
All rumors of excessive heat have been misdiagnosed, and wrongly associated with "on-die" temps, when in fact, it's the motherboard power circuirts which run to the point of failure (see my article). By now it's obvious no motherboard maker will build a board with the appropriate power circuitry to meet current demands the Socket-478 90nm Prescott requires. There just aren't enough out there, to justify the production cost.
For those whom haven't given up on 478-Prescott, remember this rule. Once you "Manually" adjust Vcore, even if it's at the original default Vcore level, your creating a conflict with a feature known as D-VID. This allows the processor to determine it's own Vcc based on Intel's "On Demand" technology, load, and thermal conditions. The following quote is from Sec 2.8 of the tech document linked above.."...During a D-VID event, the processor load may not be capable of absorbing output capacitor energy when the VID reference is lowered. As a result, reverse current may flow into the AC-DC regulator’s input filter, potentially charging the input filter to a voltage above the over voltage value. Upon detection of this condition, the AC-DC regulator will react by shutting down the ACDC regulator supply voltage..."
Back to the topic of Kingston. This stuff is miraculous...
Dam I can't upload images, nor will my server allow me to "share bandwidth" Oh go here, and scroll down a bit, I've posted my first results at my site.
Reply With Quote
Bookmarks