If it's CodeWarrior, maybe try enabling "Relax IEEE floating point" and
see if that helps.
On Oct 30, 2004, at 9:00 AM, Paul Sargent wrote:
... but does anybody think that all of this informative discussion
about fsrp being used before storing a double into a single precision
hole has anything to do with the original posters code?
It isn't actually doing that, it's frsp'ing between a fsubs and a
fmuls, and happens to store the original pre-rounded result prior to
the multiplication.
(See my first post this morning)
Paul
(Trying to drag this back onto topic, and expecting a few 'No, but it
was an interesting topic' responses. That, of course, is fair enough
:) )
_______________________________________________
Do not post admin requests to the list. They will be ignored.
PerfOptimization-dev mailing list
(email@hidden)
Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/perfoptimization-dev/
email@hidden
