Doug, Re: So that myself and others can assign the appropriate weight to your observations, could you roughly describe the extent of your understanding and the means by which you arrived at it?
It's more of a hunch, Doug. I consider the Opteron micro-architecture and apply it to projects that I have worked on. One thing that engineers always try to do is reuse existing code, even if that code is written in verilog logic. It doesn't make sense to rewrite code if you don't have to. Looking at the block diagrams for Opteron, I can see that the micro-architecture remains untouched. The implementation could be changed for all I know, but why? What does it allow AMD to do? Other than adding 64-bits, much of which does not require a complete overhaul of the design, AMD hasn't changed very much. The instruction latencies are still similar to Athlon. The bandwidth is similar, the micro-architecture is the same... so what has changed? It doesn't add up. So either I am not seeing the whole picture (quite possible), or the writer from the Inquirer is making things up (which wouldn't be the first time).
Market Data 
Markets 
AI
