Optimizing Computational Parallelism

By default, C/C++ does not model computational parallelism, as it always executes any algorithm sequentially. On the other hand, OpenCL™ does model computational parallelism with respect to work groups, but it does not use any additional parallelism within the algorithm description. However, especially when it comes to fully configurable computational engines like FPGAs, it allows more freedom to exploit this computational parallelism.