high performance computing on graphics processing units: hgpu.org

One of the major challenges faced by the HPC community today is user-friendly and accurate heterogeneous performance modeling. Although performance prediction models exist to fine-tune applications, they are seldom easy-to-use and do not address multiple levels of design space abstraction. Our research aims to bridge the gap between reliable performance model selection and user-friendly analysis. We propose a straightforward and accurate performance prediction suite for multi-GPGPU systems that primarily targets synchronous iterative algorithms using our synchronous iterative GPGPU execution model. The performance modeling suite addresses two levels of system abstraction: low-level where partial details of implementation are present along with system specifications; and high-level where implementation details are minimum and only high-level system specifications are known. The low-level abstraction models use statistical techniques for performance prediction whereas the high-level abstraction models are composed of existing analytical and quantitative models. Our initial validation results yield high prediction accuracy with less than 10% error rate for several tested GPGPU cluster configurations and case studies. The final goal of our research is to offer a reliable and user-friendly performance prediction framework that allows users to select an optimal performance modeling strategy for the given design goals.