Quote Originally Posted by JF-AMD View Post
I broke the argument down to say that in the end power at the wall is the best way to measure this. That has always been my position.

The rest of it IS confusing. Power at the wall is pretty straightforward, wouldn't you agree that it is a better way to compare?
No, it is not true either, because otherwise, you discard the level of performance at a given power watt usage.
It is much more complexe than this when you have a dynamic architecture.
It is OK to have high peak of power when you need a high peak of processing FLOPS or MIPS.
Soon, you ll tell us that your platform has better idle, so basically, your platform is better when you don't need it?
If you don't need it, don't buy it ...

Basically, Nehalem and westmere are scaling in Cores and core performance, provide the best today's avaivalble performance when you need it, and it is capable to step back to awesome idle power, all of this, at a maximize power usage for each level of load.
The problematic of choosing a server is not about comparing 2 or 4 numbers into 2 different PDF ... pretending it is as simple as this is "funny"