Dear Zakir, I think that No. of cores on CPU and GPU are really incomparable thing. However you can compare the total power of CPU cluster and GPU for example in FLOP per second. And then, for the certain application you can measure an efficiency of the parallel algorithm for CPU and GPU both, or even speedup relative to FLOP/s power. As well as for common case of CPU parallelization your will be able then print a graph of dependencies of the efficiency on comp. power of CPU (GPU) system and compare it.
Dear Andrey I want to figure out the exact efficiency of Parallel Programming using GPU. I am using Intel(R) HD Graphics 4600 and the CPU is Intel(R) Core(TM) i5 4200M 2.5 GHz (4CPU's). Would ypu Explain the Execution Time Difference of these please?