Advanced Micro Devices Inc (AMD)

[sgolds]

HailMary, it is an all too human tendency for people to convince themselves of the rightness of their positions by contriving scenerios which make it so. I see that in the proponents of Itanium - always just a couple of years from being a mainstream solution (even though both Intel and HP seem to be targeting high end servers only, for the next few years).

Does it make sense to you that the cache and die size disadvantage goes away in a few years (at 45nm)? Not to me. There is a fault in their logic, a simple fault actually:

Itanium on 45nm with sufficient cache may compare favorably with the die size of today's processors, but how will it compare with a K9 on 45nm? I posit that there will still be a cost disadvantage for Itanium, just everything gets scaled down as time goes by.

The current price point will be represented by multi-core processors in that time frame, at die sizes similar to today. Multi-core Itaniums will still be big.

As you correctly pointed out, x86 will continue to shed baggage. I'd be very surprised if K9 has Legacy Mode; many of the x86 historic disadvantages will be cut loose. Techniques to improve both floating point and integer will be adopted and the architecture will be more pure 64-bit.

It won't look very much like today's Pentium at all.