What is the fastest way to get the four scalar floats out of an altivec
vector float?
I believe the only way the processor can do this is to write the vector
out to memory (or at least to cache) and then read it back in. Not
quick.
I currently a procedure similar to the FillVectorFloat described on
Apple's website. Essentially, I form a union between the vector float,
and a vector of four floats, and return the floats.
...and that's how you tend to write it in C.
This process is buried deep inside my inner loops... Can it be sped up?
Best way to speed it up is not to do it. Seems like an odd thing to be
trying speed up...moving data out of the nice fast vector unit into the
scaler unit.
Paul
_______________________________________________
scitech mailing list | email@hidden
Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/scitech
Do not post admin requests to the list. They will be ignored.
