[NTLUG:Discuss] Alphaserver 2100 -- Digital +3y, AMD v. IBM (Intel irrelevant), lesser chips

[Bryan J. Smith]

Comcast wrote:  
> So I'm imagining that back in the day (10 years ago) these things
> were the cream of the crop?

- Digital Alpha:  3 years ahead of anyone

The Digital Alpha was the _only_ 64-bit chip available for a long, long
time.  Digital didn't fart around with a 32-bit version at all.  In
fact, unlike any other RISC processor, the Alpha does not have any 8 or
16-bit instructions at all (except for 8/16 load/stores in the 21164+).

The 21064 was _the_ 64-bit chip for a good 3 years, and the 500MHz
21164A was killing when everyone else finally "caught up."  Digital's
16-bit EV4/EV5/EV6 switching interconnect was also the I/O master -- as
Alpha ran the majority of high-end multimedia content and other
specialized storage systems for the longest time.

- MIPS:  Still going strong

MIPS did not hit 64-bit until the R4000 in the mid-'90s.  The R4000 is a
really good chip though, I'll readily admit.  Despite the existance of
the improved R5000, 8000, 10000 and 12000 series, Sony just revamped
their own R4300 design with 128-bit SIMD for the Playstation2, to great
effect.  MIPS is a consortium of fabless design companies, who typically
use TSMC and others (I'm not a fan of TSMC, they screwed up one of our
SRAM designs awhile back, costing use 4 months of lead time).

- IBM:  Knocked out, now back with a vengence

IBM had a _total_failure_ with their original PowerPC 620 design back in
'97-98.  So IBM created the 64-bit Power4 without any 32-bit PowerPC
compatibility.  It wasn't until about 2 years ago when they announced
the PowerPC 970 that they re-established 32-bit PowerPC compatibility. 
It is what the Apple G5 uses.  IBM, of course, has the leading fabs of
any company -- even over Intel.

- Sun:  Phasing itself out of the game

Sun's UltraSPARC (SPARC v9) finally hit 64-bit in '96 I believe.  It's
actually a good design.  Sun uses TI as their foundary.  They are still
stuck around 0.15um feature sizes, so the UltraSPARC III cannot keep
up.  It looks like Sun is getting out of the IC design business soon. 
The UltraSPARC IV is a multi-die chip based on the US3, and they've laid
off about 2,200 people in their semiconductor engineering division. 
They're probably going to phase out to x86-64 over the next 5 years.

- Future:  AMD v. IBM

The future is going to be built on two major platforms.
 - AMD x86-64 (aka Intel E64T)
 - IBM Power4/5

AMD has clearly taken over Intel's spot.  x86-64 is more than just an
ISA, but AMD's HyperTransport and on-board I/O MMU design is years ahead
of what Intel has planned for it's aging Pentium series.  The
"extensions" Intel offers in the Prescott core are just compatibility,
and not very competitive -- especially from an I/O-throughput
standpoint.  IA-64 Itanium III still doesn't offer the
bang-for-the-buck, and no IA-64 probably ever will, even at native
execution sadly enough.

The other challenger is IBM Power5.  The current Power4 is an excellent
example of what to expect out of Power5.  IBM will release a 32-bit
PowerPC compatible version of its Power5, just like they did with the
Power4 in the PowerPC 970 (aka Apple G5).  Where Power kicks x86-64's
but is in efficiency.  It's half as complex as Athlon64/Opteron or
P4/Xeon-E64T.  This is especially ideal for both "Blade" servers as well
as portables.

If IBM can ever get its 0.09um yield issues resolved, *NO* PC portable
will be able to compete with a 64-bit PowerPC portable.  The Athlon and
P4 mobile processors have to "slow down" to under 1GHz when on battery,
to keep consumption under 20W -- they normally use over 50-70W at full
2+GHz speeds.  Even the now 18 month old PowerPC 970 at 1.2-1.4GHz
(which is equivalent to more than a 2GHz P4) only used 19W.  IBM held
off on releasing the part, and concentrated on 1.8-2.2GHz (using 35-45W)
parts for Apple G5 systems and its own workstations instead, waiting on
the 0.09 versions to come out with faster speeds (1.6GHz+) and lower
power (close to 15W).  But, again, yield problems are holding the
release up.

IBM is opening up PowerPC licensure quite liberally.  So it may be
difficult for AMD and Intel to hold onto marketshare by the end of the
decade.  Especially if Linux catches on -- which is what IBM is hoping
for.  Intel's IA-64 has become irrelevant, as HyperTransport is
available in an optical flavor that scales far better for Opteron than
Intel can do with IA-64.

- Other Players:  MIPS, XScale, Transmeta and others

Embedded systems outsell desktops/servers 3:1, so there is still plenty
of opportunity for others.

MIPS will still be around.  MIPS is just nice, and has the greatest set
of cores and peripherials out there -- from hand-helds to networking
hardware.

Intel's XScale is based on Digital's StrongARM, the #1 reason why Intel
bought Digital's fabs and licensed their technology in 1997.  Intel
needed a replacement for its aging i960, and the StrongARM was it. 
XScale is a superscalar microcontroller (yes, that's normally an
oxymoron for those of us in the industry), and Intel will continue to
make large cash from it.

Transmeta has its IP and its innovative 128-bit VLIW (very large
instruction word) technology.  They will probably be bought out
eventually by someone, maybe Intel as its redefines itself in the coming
years.

And there are others.  Lots of 8-bit microcontrollers still sell every
year!  My former employer, Theseus Logic, is saying afloat with such
sales of their clockless NCL technology, and their main product is the
NCL08.

- Intel?  Lots of fabs, lots of money ...

Intel has a massive set of fabs, and lots of cash, so they can easily
reposition themselves.  Heck, maybe Intel will simply license AMD
designs (and not just the x86-64 ISA like it did before).  Who knows?


-- 
Bryan J. Smith, E.I. -- b.j.smith at ieee.org