Hello, i have written a vectorized code using c intrinsics for Turbo product decoder (Used as a FEC ) on altivec unit...i have no idea how to benchmark it.i mean how many cycles, instruction count ,%usage of various altivec sub units consumed during execution of the vectorized code.i went through the apple website ..n got simg4 tools ,but it was for Mac OS .can u suggest me the way around this for Windows ..i have used MPC7447A @1GHz for evaluating the performence of vectorized code. thanx in anticipation. regards, Sonali Agarwal Scientist, DRDO, India
You may want to try looking at the application notes under the 7447 Product page. They have some fast Fourier example code, that, while not being a benchmark, should give you something by which to measure performance.