Exploring Julia at the Large Scale
Practitioners of high-performance parallel computing have long sought better programming models and languages to ease the task of writing programs for large-scale systems. However, there is an undeniable tension that exists between extreme performance and developer friendliness. While the steadily increasing performance of high-level languages shows promise, and the ubiquity of concurrent programming is on the rise, these advancements have not yet made a significant impact in the HPC community. The Julia language is of particular interest as it boasts single-threaded performance on par with C and Fortran while retaining familiar echoes of MATLAB, Python, and Ruby-like syntax in the language. In this project, we are experimenting with Julia's ability to scale to large machines using benchmarks such as HPCG and custom microbenchmarking. We are also exploring ways to enhance Julia with the Hybrid Runtime Model.