Intel buzzwords
Scott Raynor
mentioned everywhere, but nowhere is there a definition for all of them. For
instance: Klamath, Deschutes, Merced, Katmai, Xeon, Coppermine, Cascades,
Tanner, Willamette, Flagstaff, Mendocino, and Celeron. I know what some of
these are, but not all. How are they different? Are they processors?
Instruction sets? CPU architectures? I am just curious as to how all of these
are related.
Tim
David D
Looking at Tom's Hardware is good advice. If you want to save time, here's
as much about it as I know off the top of my head.
"Klamath" and "Deschutes" are code names for processor cores: the .35 and .25
micron cores, respectively. PII processors using the .35 core (233MHz to
300MHz) are Klamath processors; 333MHz and higher PII's are Deschutes processors.
"Celeron," you probably already know--it uses the .25 micron Deschutes core,
but has a smaller amount of L2 cache than the PII. "Covington" is the name
given to Celeron processors with no L2 cache at all; "Mendocino" is the name
for Celerons with 128K of L2 cache. Mendocino-class Celerons will come in two
flavors: Slot 1 and Socket "9". Socket 9 isn't the actual name given to the
socket that this Celeron will use, but since it hasn't been given an "official"
name yet, and the last Socket we had was Socket 8, well...
"Katmai" is a processor family name, like "PII". It will use the Deschutes
25 micron core, at least initially, and will be a Slot 1 processor (initially).
Its main distinction from PII is the implementation of MMX-2 instructions,
which like the 3DNow! instructions of AMD's K6-2 processor, will enable the CPU
to streamline the process by which it handles standard graphics and multimedia
calculations, improving 3D and multimedia performance. Celeron, Pentium II and
Katmai are all aimed at the home and business PC market segment.
"Xeon" is another Deschutes processor, using the same .25 micron core as the
higher PII's. Xeon is given its own processor family code name to distinguish
its target market segment: servers and workstations. Xeon has L2 cache that
runs at full processor clock speed, unlike the PII's 1/2 processor speed cache.
Xeon also has three flavors of cache: 512K, 1MB, and 2MB. It's a bigger
cartridge as well, and to top it off, it uses a Slot 2 connector, while PII's
use the smaller Slot 1 connector.
The successor to Xeon will be Tanner, another family name. I haven't heard
any info. on the distinction between Tanner and Xeon. Presumably Tanner will
be faster; probably it will use a smaller die size, making it a non-Deschutes
processor core. The next step down in core size is supposed to be .18 micron,
achievable through copper technology, although I have heard that Intel will
attempt to produce .18 micron processor cores using current production
technology, before switching over to copper. Tanner may be a .18 micron, non-
copper core processor. I really don't know.
"Coppermine," as the name implies, is supposed to be a copper-based chip,
starting at a .18 micron die size. It will be the successor to the Deschutes
core and will inherit its market segment: home and business PC's. "Cascades"
is supposed to be a processor family using the same .18 copper core as
Coppermine, but aimed at the server and workstation market, as the Xeon is
today. What distinguishes Cascades from Tanner? I don't know.
"Merced" is Big Daddy--the first processor to use Intel's IA64 64-bit
processor architecture. This CPU will compete directly with other 64-bit CPUs,
like Digital's Alpha processor, in the high-end market.
I've heard the terms "Flagstaff" and "Willamette" somewhere, but I can't
remember where right now, or I'd look them up too. Maybe someone else around
here can clarify those codenames for us.
Regards,
D.D.