How to use ippmMul_mama_32f_S2 of IPPM?

[jdc25]

I wanna get the MUL result of two matrices(3x3) ,the code as following:

Ipp32f mSrc[] = {m_kXValue[0], m_kXValue[1], 0,
m_kYValue[0], m_kYValue[1], 0,
0, 0, 1};
const Ipp32f* m[] = {mSrc};
Ipp32f nSrc[] = {1-m_pkPoint[iFrame][0], 1-m_pkPoint[iFrame][1], 0,
m_pkPoint[iFrame][0], m_pkPoint[iFrame][1], 0,
0, 0, 1};
const Ipp32f* n[] = {nSrc};
Ipp32f dst[9] = {0};
Ipp32f* pDst[] = {dst};
ippmMul_mama_32f_S2((const float*)m, 0, 0, 0, 3, 3, (const float*)n, 0, 0, 0,3,3,(float*)pDst, 0, 0, 0, 1);

It is strange that pDst are always Zero! Why?

Thank you for your reply.

[Vladimir_Dudnik]

Hello,

As we can see you set all stride parameters to zero, that's wrong. We also noticed that you use old, pre-IPP 5.0 API (there were changes in IPP matrix functions API, to make it more consistent and easy to use). I recommend you to read IPP manual about how to use IPP functions and probably you can consider to get the latest version of IPP, which is 5.1.1. The main focus of IPP 5.1.1 is to cover the latest Intel micro-architecture innovations found in Intel Core 2 Duo product line.

Regards,
Vladimir