Dataflow Virtual Machine Profiling

In the Dataflow model instructions are executed as soon as their input operands are ready, allowing the natural exploitation of instruction level parallelism (ILP), which makes it extremely useful for increasing applications´ performance on multicore machines. However, the lack of accurate information on the parallel code can make it more difficult for programmers to perform code analysis and optimization. Thus, the aim of this work is to propose and implement a profiling mechanism for Dataflow runtime environments. To validate the profiling tool implemented, an analysis of the overhead is presented and, also, how the data generated can be used to optimize the code.