What are the basic CFG File and Modules Needed for NDK Stack and OpenMP projects?

I don't understand the CFG files and required modules well. What are the required components to build a project to do the following?

1) create the NDK stack for TCP/IP packet sending over ethernet

2)multi-core FFT process data from DDR3 using OpenMP

The processing task is relatively simple. Using an array stored in DDR3, each core is given it's slice of the original data and runs an FFT and outputs to another buffer in DDR3. 

I have gotten the processing to work correctly (data-wise) based on the MCSDK image processing demo, but I have much slower times than I see in my stripped down processing project (based on the OpenMP HelloWorld project). My original benchmark for a N=2048 fft was ~30,000 cycles, and using the image_processing project, it jumps up to about ~130,000 cycles. I suppose there's a problem with caching and settings of that sort.

Thanks,

Ryan