AK> RB> Actually, I am doing a considerable amount of "floating
point" math
AK> RB> using fixed point integer math to do 3-D projections.  The reason I
AK> RB> don't want to use floating point is that integer math is faster than
AK> RB> floating point EVEN WITH A COPROCESSOR especially for addition and
AK> RB> subtraction,

AK>Yes, additions are faster in integer math. However, at least on a pentium,
AK>multiplication and (correct) division are faster using
AK>floating point math, since integer mul/div are actually
AK>executed by the FPU with all the necessary conversions
AK>before and after and no pipelining. Whereas with FPU
AK>instructions, you can start a pipelined add/sub/mul
AK>operation every cycle (successive muls every other cycle).

AK>And pentium just follows the common trend. In case of a Power[PC] or other
AK>RISC, you loose at least half of the processor performance
AK>and half of the available registers, if you replace
AK>floating point ops by integer ops. All thease beasts are
AK>designed to execute floating point in parallel with
AK>integer, and multiplication is usually done much faster
AK>using floating point.

Interesting, I didn't know they had moved integer mul/div math over to
the FPU.  I'll keep that in mind when I have to consider the
Pentium at the bottom end of the performance spectrum for my
projects ;-)  For now, a Pentium would JUST ABOUT get rid of the need
for any assembly work.

Merry Christmas,

Rob.
___
 X SLMR 2.1a X I haven't lost my mind...it's backed up on tape somewhere

--- Maximus/2 2.01wb