Pegasus: coordinated scheduling for virtualized accelerator-based systems

Abstract

Heterogeneous multi-cores--platforms comprised of both general purpose and accelerator cores--are becoming increasingly common. While applications wish to freely utilize all cores present on such platforms, operating systems continue to view accelerators as specialized devices. The Pegasus system described in this paper uses an alternative approach that offers a uniform resource usage model for all cores on heterogeneous chip multiprocessors. Operating at the hypervisor level, its novel scheduling methods fairly and efficiently share accelerators across multiple virtual machines, thereby making accelerators into first class schedulable entities of choice for many-core applications. Using NVIDIA GPGPUs coupled with x86-based general purpose host cores, a Xen-based implementation of Pegasus demonstrates improved performance for applications by better managing combined platform resources. With moderate virtualization penalties, performance improvements range from 18% to 140% over base GPU driver scheduling when the GPUs are shared.