topic Hello Matthew, in Intel® Integrated Performance Primitives

Function for matrix multiplication of complex values

Matthew_S_ — Fri, 11 Jul 2014 12:52:52 GMT

I'm fairly new to IPP and I have what seems like a silly question. Does IPP provide a function for multiplying two matrices of complex values (32fc format)? A colleague and I scoured the Small Matrices and Realistic Rendering and Signal Processing reference manuals yesterday and could not find it. Is this operation not supported (in a single function call) in IPP? If not, what about MKL?

Hello Matthew,

Ying_H_Intel — Thu, 17 Jul 2014 08:36:34 GMT

Hello Matthew,

Yes, MKL should support the complex type matrix multiply, (real, 32fc and double 64fc float type). and also have good optimization for small matrix.

like cgemm() and zgemm(). You can find fortran sample or C sample in MKL install directory/example/blas and cblas

Best Regards,

Ying

Thanks for the reply. I guess

Matthew_S_ — Thu, 17 Jul 2014 12:43:20 GMT

Thanks for the reply. I guess MKL is better suited to my problem space than IPP.