topic Re: Multiplying a complex vector with a real scalar in Intel® Integrated Performance Primitives

Multiplying a complex vector with a real scalar

fa — Tue, 19 Oct 2004 18:57:10 GMT

To multiply a complex vector with a real scalar I have used the ippsMulC_64fc function, and represented the real scalar as a complex one with zero imaginary part.

This seems unnessesary, so I tried instead to use ippsMulC_64f function and instead treat the complex vector as a real one of twice the size. The two approaches give the same results but strangely the complex routine is faster.Is there any better way to multiply complex vectors with real scalars?

The code used is of the form:

Ipp64fc *in, *out, val;

stat=ippsMulC_64fc(in, val, out,length);

stat=ippsMulC_64f((double*)in, val.re,(double*)out,2*length);

Message Edited by fa@maths.lth.se on 10-19-2004 04:57 AM

Message Edited by fa@maths.lth.se on 10-19-2004 05:04 AM

Re: Multiplying a complex vector with a real scalar

Vladimir_Dudnik — Fri, 22 Oct 2004 00:02:35 GMT

Hi,

thanks for the question, we found problem in optimized code of ippsMulC_64f and ippsMulC_64fc function, their performance will be improved in the next version. We also think about adding of new function, ippsMul_64f64fc in the next version.

Regards,

Vladimir

Re: Multiplying a complex vector with a real scalar

snirgaz — Mon, 15 Aug 2011 07:54:56 GMT

Hi,

Regrading ippsMul_64f64fc.

Is this feature already implemented?

Thanks

Snir

Re: Multiplying a complex vector with a real scalar

Vladimir_Dudnik — Thu, 18 Aug 2011 11:06:03 GMT

Hi Snir,

as I can see from product documentation only 32f32fc variant of ippsMul function is availbale.

Regards,
Vladimir