3 Feb
2007
3 Feb
'07
8:38 p.m.
Chuck Guzis wrote:
The same sort of behavior exhibits itself
on some of the x86 CPUs--where:
l1:
mov al,ds:[si]
dec cx
mov es:[di],al
jnz l1
runs about as fast as rep movsb (and, in cases, somewhat faster). On
those same CPUs, replacing the dec cx/jnz with a loop instruction
will actually slow things down.
Could you clarify which 808x CPUs in particular you're referencing in
the above? There isn't a single case from 808x to 486 where that code
will beat rep movsb. Were you referring to Pentiums and higher?
--
Jim Leonard (trixter at oldskool.org) http://www.oldskool.org/
Help our electronic games project: http://www.mobygames.com/
Or check out some trippy MindCandy at http://www.mindcandydvd.com/
A child borne of the home computer wars: http://trixter.wordpress.com/