[geeks] cray vs. g4 8-)

[geeks at sunhelp.org]

mrbill at mrbill.net writes:

>A note I just found about the Cray-1:

>Max. performance 160 Mflops; 64-bit arithmetic; max. attainable sustained 
>performance 150 Mflops. There are codes for matrix multiplication and the 
>solution of equations which get close to this. Maximum scalar rate is 80 
>mips. It is easy to attain over 100 Mflops for certain problems, even using 
>Fortran. 

>Wow, and my G4 did ~1500 Mflops when running a fractal benchmark program
>the other day.

Run something big and cache-hostile and see what happens.

The iteration sequence for, say, the mandelbrot set, fits in the
*registers*, never mind the cache.

-------- David Fischer --------- dave at cca.org --------- www.cca.org --------
-------- If it wasn't for good luck, I wouldn't have no luck at all. -------