🚀 go-pugleaf

G. Baldwin (drizzit@eskimo.com) wrote:

: Digital Equipment plans to have the Alpha pushed out past 1 GigaHerz by 
: the year 2000, and even SGI/Mips/Cray plan to be at least a full step 
: ahead of anything Intel puts out (the P6/150 only does 215 SPECfp while a 
: MIPS R10000 does around 600 SPECfp).  True that Intel puts more focus into 
: Integer math while MIPS put its focus in Floating Point, but the fact 
: remains...  now that SGI is running things a lot more over at MIPS, I 
: can't wait to see what they have in store for us with the R12000.

I believe that the 600 SpecFP number is *estimated* not in actual shipping
systems. P6@200 is currently available and shipping.  Furthermore, I haven't
seen anything shipping past the 500 SpecFP level (except the Dec Alphas)
In shipping systems, you get SpecINT95 of 8.1 for the P6@200, and 8.14 for
Power Challenge 10000, SpecFP of 6.8 for the P6, and SpecFP of 10.5 for the
R10000.  Mips doesn't seem to be any faster in integer code, and about ~35%
faster in FP intensive app.  A full step ahead? sure, in SGI's traditional
graphics animation field, the 35% is still very important, but most 
certainly isn't the 280% differential that you seemed to imply above.

We'll definitely see 1+ GHz before the decade is out. Most certainly 
from Dec, but maybe from other cpu families as well. Watch for Intel's
P7 to hit some unbelievable speed also.

: But he is correct by saying "borrow".  When I look at the specs for all 
: workstation RISC based CPUs, the HP-PA is the 2nd slowest (only the PPC620 
: is slower).  The odd thing is that in dollar volume, the HP-PA line sold 
: the most.  Why is it that the worst CPUs always sell the best?  Hmmm... 

cost. support. (value perception)

: Anyways, I fear that both will snag stuff from other chip makers.  I see 
: the design of something close to the 'Visual Instruction Set' that the 
: UltraSparcs use, or some kind of new instruction sets designed to handle 
: specific jobs.

snag stuff from other chip makers? interesting way of putting it. Fact is,
there is a large hodge podge of architectural techniques that various
people have claimed that will make you cpu so much "faster".  Gradually
more and more are used to get more performance, because over time, the
transistor budget allows its implementation. However, in some markets,
(X86 specifically) some will never be implmented because they'll break 
backward compatibility.  Throw out backward compatibility, and you'll
get better and better performance chips, but no one will buy them because
they're only 20 to 30% faster than the next chip, and it doesn't have
any real software base. 

:                        Greg Baldwin (drizzit@eskimo.com)
:      Amiga junkie and user since 1987    Computer Science & DTV Student
:      Commodore64 fan since about 1983    http://www.eskimo.com/~drizzit
:                Tyranical EFNet #Amiga Channel Operator "Drizzit"