Warren Nagourney wrote:
I don't think that adds so much, because the compiler can (a) just use registers for all temporaries, (b) align variables so they don't have to be shifted, and (c) schedule the preferred slot code so that it is dual-issued with something else.I was very surprised that the spu was so much faster than the ppu since the spu has all of the added code needed to move the data to and from the preferred slot. If you ever find out something official you might post it here.There is some double precision in my program, but it is mostly single. I did this to facilitate conversion to altivec which doesn't support doubles (this was 3 years ago). One hopes that the next generation of the cell will support doubles better (I assume that the shrink to 65 nm will not be different in this respect). Jon --
 |
_______________________________________________ yellowdog-general mailing list [email protected] http://lists.terrasoftsolutions.com/mailman/listinfo/yellowdog-general HINT: to Google archives, try '<keywords> site:terrasoftsolutions.com'
