[x264-devel] floating point or 64-bit

[Måns Rullgård]

I need to do some simple calculations in the rate control which either
require floating point, or a rearranging that will overflow 32 bits.
The question is thus, which is more costly, switching between mmx
etc. and floating point, or using 64-bit integer math?  It's a matter
of a few statements per MB, and a few more per frame.  Are there other
reasons to avoid floating point, e.g. running on hardware without an
FPU?

-- 
Måns Rullgård
mru at mru.ath.cx

-- 
This is the x264-devel mailing-list
If you are in trouble, please contact <postmaster at videolan.org>