[rescue] Cleaning out, anyone want some FREE stuff?

[Mike Meredith]

On Thu, 13 Jan 2005 10:43:50 -0600 (CST), Jonathan C. Patschke wrote:
> Clock-for-clock, SPARC is generally faster than MIPS by a pretty big
> margin.

Well they say that there's lies, damm lies, and statistics, but it looks
as though that's not the case (ok I'm cheating by not looking at older
CPUs) :-

System                           SPEC CINT2000     CFP2000
IBM p5-550, POWER5 at 1650MHz            1200(.72)     2121(1.29)
Fujitsu PRIMEPOWER900, SPARC64 at 1890   1174(.62)     1510(.80)
Sun Blade 2000, USIII at 1200             642(.53)      953(.80)
Sun V1280, USIII at 1200                  608(.51)      873(.72)
SGI O3200, MIPS R14K at 600               483(.81)      499(.83)

The numbers in brackets are the relevant benchmark divided by the clock
speed of the CPU ... a totally bogus value that's quite revealing. It's
interesting that Sun didn't have values for the larger servers in the
CINT2000/CFP2000 tables I looked at ... just values for CINTrate and CFP
rate.

Now back to tackling a particular big pile of self-assembly furniture :(