>From what I we understand from the Hammer's release notes
>is that one 32bit core will be fed zero's while the system
>uses 32bit code. 64bit will be codemorphed to feed the
>dual 32 bit cores during 64bit use. I
Hello ? Are you sure you know what you are saying here ? If I understand you correctly, this doesnt sound remotely like what I've been hearing so far (but then again, Im no authority on the matter). Are you saying you'd need a dual core (sledge)hammer to execute 64 bit code ? Because that IS bogus, im positive. Dual core hammers arent even confirmed. THe hammer architecture, with the dual integrated memory controller is ideally suited to build multi core chips, but that is certrainly not the most important feature of its architecture. AS for the 64 bit, a single core clawhammer is perfectly capable of executing x86-64 code. And 32 bit code will definately not cut the performance in half by just using half the cpu, come on. What might be true is that 64 bit registers will contain 32 zero's when running 32 bit code, which seems logical. But thats only the registers (or rather, some registers I presume).
On a more general note, I agree with IIB that Hammer and Itanium are not really direct competitors. Hammer will take P4 Xeons head to head instead, using the 64 bit extentions as a free "extra". Hammer might be an indirect threat to itanium however. Intel currently has no 64 capable product to compete in the "low end" server market with hammer. Itanium is not likely to become low end any time soon. Low end means high volume, so if hammer succeeds, software support will follow, forcing intel to follow AMD x86-64 approach to be able to compete. And THAT would undermine Itaniums IA64 migration.
= The views stated herein are my personal views, and not necessarily the views of my wife. =