Hi all,

Because of the recent interest in CM vs. OCT performance, I've posted
some old timings for a CPU-intensive benchmark at:

   http://www.allegro.com/sieve_timings.html

The program was the Sieve of Eratosthenes, calculating primes in
the range 2..8190 in a loop (100 times).

The bottom line:

   CM  times ranged from 16,543 to 49,584 CPU milliseconds;
   OCT times ranged from  2,658 to 10,741 CPU milliseconds;
   NM  times ranged from    434 to  2,150 CPU milliseconds;

(depending upon language, and optimizations selected)

Or, to put it differently, the spread from the slowest CM language
(and optimization flatgs) to the fastest NM one was a factor of 114!

--
Stan Sieler                                          [log in to unmask]
                                     http://www.allegro.com/sieler.html