Replies to post #63974 on Advanced Micro Devices Inc (AMD)

[jhalada]

Klaus,

Merom is definitely a new architecture. Although it may be heavily derived from Pentium M (Banias, Dothan, Yonah), but there are substantial changes. Just the headline ones - AMD64 support and 4 issue ALU make up basically a new core, rather than just a revision of the existing one.

My guess is that Intel will be conservative with other changes, since Merom is a life and death release for Intel. A failure of Merom equals failure for Intel as a whole. There is nothing left to bail Intel out. Getting substantial gains from Netburst core at reasonable performance / Watt ha been increasingly difficult.

Server version of Yonah (Sossoman?) is at best a niche product (Microsoft calls 32 bit in servers "Legacy" now) and Itanium continues its dollar short, day late routine.

Yonah looks like a great core, and alone could keep Intel profitable for another 18 months, but without Merom, there is a steep cliff for Intel beyond that.

Frankly, I am surprised that Intel is even going from 3 to 4 issue. It may turn out to be a brilliant choice if everything falls in the right place, but, IMO, it was an unnecessary risk, considering the consequences of major failure of this core for Intel.

Joe

[wbmw]

Re: Whenever did Intel introduce a new µ-arch in the middle of a node? Nobody with a sane mind would ever do this.

LOL, almost every new micro-architecture is introduced on a stable process node.

- Intel introduced the Pentium Pro on .35u in November 1995 only after the Pentium 120MHz and 133MHz (P54CS) came out on .35u in March and June of 1995, which was a shrink of the P54C core.
- Intel introduced SSE in the Katmai core on .25u in February 1999 on after establishing it on the Deschutes core in September 1998, which was a shrink of the Klamath core.
- Intel introduced the P4 Willamette core in November 2000 on .18u in only after establishing it on Coppermine in October 1999, which was a shrink of Katmai with integrated L2 cache.
- Intel introduced Banias on .13u in early 2003 only after establishing it on the Tualatin core in July 2001, which was a shrink of Coppermine.
- Finally, Intel will have introduced Merom micro-architecture in H2 2006, only after establishing their 65nm process in early 2006 with the release of Yonah, Presler, and Dempsey, which are themselves process shrinks.