Overclocking Results!
Ok, lots of good news here! Firstly, I'm running with v1.0 of the BIOS (the one it came with) which is most definitely buggy. One particular bug is the CPU multiplier has no effect, so for my E4300 it's always nine. This meant I couldn't test FSB overclocking. The benchmarks are from Sandra 2007.
At stock: CPU=1800, FSB=800, Mem=400 @ 5-5-5-15 1T, Int=16538, Float=11232, Mem=4500Mb/sec @ 94ns
Simple bump: CPU=3000, FSB=1333, Mem=667 @ 5-5-5-15 1T, Int=27579, Float=18697, Mem=6500Mb/sec @ 80ns. Uses 117W idle 135W busy.
This simple bump was done by setting Unlinked and then FSB to 1333 and Mem to 667 to force a 1:1 instead of 5:6 divider (which is slower). This passed Orthos at five hours. No voltage bumps or anything required, though the BIOS complains about an unsupported CPU.
I then cranked her up bit by bit and got this maximum overclock:
CPU @ 1.4v (+0.1125v)=3375, FSB=1500, Mem=750 @ 4-4-4-12 1T
This yields Int=31047, Float=21061, Mem=7577Mb/sec @ 66ns. Uses 152W busy. This has passed Orthos for an hour (it's still running right now). CPU temp is 40C idle and 52C working.
This CPU limit of 3375 is pretty fixed. It wouldn't even go to 3386 no matter what voltages I applied. For reference, this E4300 has an id of Q641A275 which makes it a week 41 CPU which are known to be good overclockers.
According to anandtech, the new v1.11 bios fixes the cpu multiplier not working bug so after its release, I can tell you how far the fsb will go. For reference, it gets warm but not hot at present.
There are some irritations with the BIOS eg; it doesn't recover well currently from bad overclocks and you need to clear the BIOS most times. Still, not at all bad for a v1.0 BIOS.
Any questions?
Niall