MathGroup Archive 1994

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Power Macs

  • To: mathgroup at yoda.physics.unc.edu
  • Subject: Re: Power Macs
  • From: BOSS at pmel.noaa.gov
  • Date: Thu, 30 Jun 1994 10:38:50 -0700 (PDT)

Kiyoshi Yoda writes:

>Somebody mentioned 486DX2/66 is remarkably fast for mma calculation.
>10000! calculation ended up with just 17.5 second using my 486DX2/66.
 
>I was amazed this is faster than SPARC10 result mentioned above!
>Natural extrapolation is that PENTIUM is faster than PowerPC regarding
>at least 10000! computation. Any feedback about this?

I have seen comparisons of Specmark performance of the PENTIUM vs the Motorola 
601.  I am unable to quote numbers but I believe the PENTIUM was slightly 
faster in integer operations at comparable clock speeds.  This explains the 
results for 10000! which is an integer operation.  It is in the floating point 
arena that the 601 really outshines the PENTIUM.  It would be interesting to 
see performance figures for the gamma function, the equivalent of factorial for 
real numbers.

I believe I saw a table of the Specmark statistics in MacWeek.  I will try to 
locate them and forward them to the Math Group.

Ed Boss
NOAA/Pacific Marine Environmental Laboratory  







  • Prev by Date: Nonlinear optimization
  • Next by Date: Transformation of Expressions
  • Previous by thread: Nonlinear optimization
  • Next by thread: Transformation of Expressions