compile with:
g++ -fPIC -shared -o libclFFT.so -I /opt/hsa/include clFFT.cpp fft_kernelstring.cpp -L/opt/hsa/lib -lhsa-runtime64
For each new FFT size, uncoment line 318 of clFFT.cpp, grab the OpenCL kernel,
save it, and create a file clFFT_SIZE.cl. Compile & run.
