High Performance Accelerated Graph Programming Enabled by Julia

Julia Computing seeks to significantly enhance visualization, data and graph analytics at high performance speeds to provide information for actionable intelligence to DoD and commercial sectors. This SBIR will leverage the Julia computing language's flexibility and modern architecture to exploit parallelism available through a range of modern accelerators including many core architectures (e.g. GPUs and Xeon Phi) connected through shared and distributed memory. The prototyped Julia approach is substantially faster allowing this proposal to focus on work that will lead to production level computational scalability. Julias intent is to go beyond easing the burden of integration, and head towards the elimination of that burden. The dedicated Julia team will explore a linear algebraic approach to graph algorithms that exploits the sparse matrix representation of graphs can provide a variety of benefits. These benefits include syntactic simplicity, easier implementation, and higher performance. One way to employ linear algebra techniques for graph algorithms is to use a broader definition of matrix and vector multiplication. The progress that the sparse matrix community has made with performance can be brought into the world of graph-based computing. Julia is well poised to bring these abstractions to the highest performance of computing.