The first indications it was going to be a rough go, were my experiences with the Abit IS7-E. The board did quite well, and with my (OEM) P4 Malay L310A735 2.4C, I easily attained 3.6GHz (1.550Vcore, 3:2 divider, 300FSB) right out of the box. This was achieved with OCZ GolD3700 populating all four DIMM's on the IS7-E. I then discovered whenever I exceeded certain FSB speeds (usually very high e.g. 334FSB) clearing the BIOS via the jumper less method wouldn't work. No matter how many times I hit Reset, and Insert (simultaneously) she wouldn't clear. I initially thought I fried the CPU, however; knowing Intel uses an internal overheat protection circuit, I simply powered down, cleared the CMOS, and tried again. Although I did have to pull the plug to do it.
There were several scenarios where had I to employ this method with the IS7-E whenever I exceeded 309FSB. Then my Corsair Twin-X4000 arrived in 1GB-kit (2x512MB) thank you Robert)) and I excitedly tossed it in the IS7-E's DIMMs 1&3 to begin testing. I upped the FSB to 260MHz (520MHzDDR / 1:1 /2.8VDIMM / 2.4C @ 3.0GHz) everything was fine. I then pushed the memory a little further, but the anomic VDIMM of the IS7-E (2.8V) prevented me from achieving any spectacular overclocks. My only other option was to try out some bandwidth alternatives using dividers. Here the errors compounded. Although the OCZ GolD3700 would work at any divider on the IS7-E, the Corsair wouldn't? Each time any ratio other than 1:1 was used, the motherboard would not post, and in each circumstance I was unable to re-set BIOS values using the jumper less method.

Again I had to "pull the plug" and clear CMOS. So I re-flashed from the Shipping BIOS, to revision 15 (at that time the highest). No such luck. I therefore assumed the memory was to blame, and wrote to Corsair. They proceeeded to send me two 512MB-kits of their Twin-X4000 (4x256MB modules total). This was per my request, as I thought the anomaly was specific to 512MB size modules. Armed with Corsair's 256MB size sticks of Twin-X4000, I experienced the same anomalies, being unable to use any divider other then 1:1. I wrote Corsair again who informed me I was not alone in this, and they had better luck testing the Twin-X4000 on the IC7-G.

By this time the Abit IC7-G arrived. I now had two different sizes of Corsair Twin-X4000 memory (256MB & 512MB) with which to verify the scope of this anomaly, seemingly specific to Corsair memory. I'd been having good luck with OCZ GolD3700 on Abit IS7-E, attaining 3.6GHz and populating all four DIMM's. When the IC7-G arrived I thought for sure the higher quality NB would allow me to reach beyond 3.6GHz, which it did, but only by 96MHz with any stability. I'd been able to post into 4.0GHz with the IC7-G and OCZ GolD3700, but where the Windows logo should pop-up, I only experienced black screen. I tried a different HDD thinking I'd have better luck slowing an older Maxtor Diamond Max Plus ATA133 to PIO Mode-4, instead of the Maxtor Diamond Max Plus 9 SATA150 (a phenomenal drive). But above 3.7GHz the black screen would return, and at 3.7GHz Windows was unstable running Sandra, or even running the snapshot software to verify WCPUID. The highest I was able to reach was 3.696GHz with any stability, and I've not been able to repeat that result. I still wasn't sure at that point if it was the low VDIMM or the NB-cooling holding me back with IC7-G.

Looking back, I believe the NB on the IS7-E may actually be higher quality Silicon? I'm still perplexed about my temps, as well. They were better on the IS7-E prior to my lapping the P4's IHS, which did show a marked drop when tested on the Epox. This is why methodology is so important. Once the variables change, it's next to impossible to make accurate inferences. Using the Thermalright SLK900U (a phenomenal product, look for their SP-94 Heat Pipe) and the Vantec Tornado (helicopter engine) 92mm/119CFM HSF, the temps never went beyond 36C on the IS7-E, even while overclocked to 3.6GHz (see below);



Oddly enough what I originally thought was my handicap (IS7-E) so far has been the best Overclocker of all the boards where stability at 300FSB is concerned. The IC7-G seems to have a little more FSB headroom, but no VDIMM to test the theory. In so far as the IS7-E the speed is surprising considering I populated four DIMM's. That alone can be problematic with some motherboards. In fact my next project is to volt-mod either, or both the Abit mobo's.