Quantifying Overheads in Charm++ and HPX Using Task Bench

Asynchronous Many-Task (AMT) runtime systems take advantage of multi-core architectures with light-weight threads, asynchronous executions, and smart scheduling. In this paper, we present the comparison AMT Charm++ HPX main stream MPI, OpenMP, MPI+OpenMP libraries using Task Bench benchmarks. is a parallel programming language based on C++, supporting stackless tasks as well threads asynchronously along an adaptive system. C++ library for concurrency parallelism, exposing standards conforming API. First, analyze commonalities, differences, advantageous scenarios in detail. Further, to investigate potential overheads introduced by tasking HPX, utilize existing parameterized benchmark, Bench, wherein 15 different were implemented, e.g., MPI + extend adding implementations. We quantify Charm++, where single task multi-task are assigned each core, respectively. also system’s scalability ability hide communication latency.