Thanks for the replies.
Hmm.. I suppose it could be the PSU, but I'm hardly asking for much, just to be stable at 2.1 at stock voltage! I also think that core 1 would be having problems as well if it was the PSU, but it's totally stable up to 1.5V (at which point my computer won't boot!!).
Yes, I've run with my memory at a 1/2 divider, and as I mentioned I'm pretty sure it isn't the memory because a simultaneous blend test running on core 1 is totally stable, and that should isolate the RAM as being fine. At any rate it's well within it's specifications. Right now it's at 180. Additionally, all the memory timings i'm using are stable according to the OCZ forums. I've found that running at drastically relaxed timings has no effect at all on the stability of core 0.
I'm aware that there is a +% modifier to the voltage (in fact, i'm using 1.1V + 110% in order to get a 1.184 readout from CPU-Z).
Let me reiterate that below this voltage I cannot boot, and above this voltage (1.1 + 113%) core 0 is not prime95 (FPU stress test) stable.
I've tried using higher LDT voltages, but it has no effect. I've swapped out the RAM with another set of 2x generic PC3200 DDR400 and found no change in stability. The only parameter that seems to have a bearing on the stability (the difference between failing in <1min and stable for 5+ hours) is the CPU voltage.
Starting at 200MHz HTT, the stability of the system is roughly linearly related to the CPU voltage. To be stable at 2.1, V=1.25. At 2.2, V =1.21, At 2.35, V=1.8.
Lastly, I have read the overclocking guides, fairly carefully. That's why I'm so surprised!! My computer is behaving almost exactly contrary to the dogma: "If the CPU is unstable (and the RAM is okay), increment the voltage upwards"