05-05-2009 08:16 PM
I'm new to IPP, so my apologies if I'm missing something obvious.
I don't see complex versions of MatVecMul and DivCRev. It seems a little odd since DivC and so many of the other of the arithmetic functions in the signal processing subset have complex versions. The algorithm I'm hoping to implement with IPP needs these two functions for complex values. Is there something I'm missing?
I could implement MatVecMul by using DotProd, but I assume that would be less efficient than an actual implementation of ippsMatVecMul_64fc. But I don't see how to replace DivCRev for complex values, other than by filling a vector with my scalar constant and using ippsDiv_64fc. But that seems a shame to have to do at runtime. Thanks for your thoughts.
05-07-2009 05:39 AM