by Yuxiong He
As an example of multicore-enabling a performance-sensitive application, we decided to convert the Discrete Hedging example to a parallel Cilk++ program (that is, to “Cilkify” the example) and increase performance on multicore systems from the popular computational finance library QuantLib. We chose DiscreteHedging.cpp, an example of using the QuantLib Monte Carlo framework.
As a preview of the results we obtained, here is a graph of the results on a 16-core system, showing that speedup is almost linear with the number of workers, and hence the number of cores used by the parallel Cilk++ program.
