Scalable Graph Analytics on GPU Accelerators

Abstract

Sparse data computations are ubiquitous in science and engineering. Two widely used applications requiring sparse data computations are graph algorithms and linear algebra operations such as Sparse Matrix-Vector Multiplication (SpMV). In contrast to their dense data counterparts, sparse-data computations have less locality and more irregularity in their execution - making them significantly more challenging to optimize. This is especially true for accelerators and many core systems.

In today's talk, I will cover NVIDIA's and the graph community's effort to overcome these challenges and to create a simple to use framework that will enable both programmers and data scientists to get high performance graph algorithms, with high productivity, and an easy to use API that does not require broad HPC knowledge.