Just to update this;

When a new phenomena arrives, it is difficult to actually understand or assess it because of the lack of experimentation and experience. As time builds, you develop a much clearer picture of the situation.

I noticed I had a non-booting Phenom like some of you were saying, what you would call "dead Phenom", but I didn't return it even after 3 days although I was about to, rather I experimented and repeated it. It was caused by just going past 626MHz on RAM. Voltage didn't matter whatsoever, I've had 5 Phenoms (7-43/7-44/7-45/7-47 etc) now and ran 1.8-2.9VDIMM on them varyingly since around 20th November '07 at low and high VCore and at low and high NB VIDs, no mem. degradation whatsoever nor did those affect it... but speed did. Phenom did 626 4-5-5-15 at 2.2V BIOS/2.18V real fully stable but failed to boot 627 even at up-to 7-8-8-22 2.9V. It did not matter what IMC/CPU volts were fed nor the CPU/RAM temp (up-to 10C idle tested).

The same RAM I swapped in a P35/C2Q setup and it ran 2.2V 676MHz 2.3V 5-5-5-5 SPi 32M with ease 4mins later.

Guess how it all happened?

When choosing high mem speeds on 1:2.66 divider, Phenom would get stuck in a reboot cycle even at low volts. No POST, no matter what you did even after 2 days. By checking the various voltages, I found CPU voltage had dropped to sub 1.19V at whatever setting you chose before that, it didn't matter and that caused a failed bootup -- didn't matter which CPU/NB/HT speeds/volts you ran previously. I usually test under cold and as you'll see documented in the 9500/MSI K9A2 Plat. thread, I had to take the CPU subzero (below -15C) and leave it at those temps for 30mins, then bootup with only one DIMM installed and it would always work fine and restart up thereafter. So "dead CPU" perfectly alive again.

If it were anyone else not trying sub-10C, or even me not trying it, it definitely was like a dead CPU and an RMA was needed.