How to do cross correlation in MSP430

Hi,

      I wrote a cross- correlation function in MSP430f2274. But it is really slow if I increases the number of samples (to 500).

      I am wondering if there is a faster way to do accurate cross correlation in MSP430 ? I checked cross-correlation using fft and inverse fft. I am not sure if it is implementable with number of samples over 500.

      Here is the code.

int  xcorr ( int Array1[175], int Array2[175], const int size)
{
    int i,lag=-size + 1;
    int index =0;
    int16_t temp = 0;
    int16_t peak=0;
    
    int16_t xcorr_result[350];       
    for (lag=-size+1; lag<size;lag++)    //loop of calculation
    {
        temp=0;
        for (i=0; i < size ; i++)
        {
            {
                temp = temp+Array1[i]*Array2[lag+i];
            }
        }
        xcorr_result[lag+size-1]=temp;
    }

    
    for (i=0; i < 2 * size-1;i++)              //finding the peak index
    {
        if (xcorr_result[i] >= peak)
        {
            peak = xcorr_result[i];
            index = i-size + 1;
        }
    }

    return index;

}

The first part of doing all the calculations is using the majority of time.

For two arrays of 175 length, using a 8Mhz clock, it takes about 0.6s to complete such a calculation.

Anyone can help ?

Thanks!