Argus-G: Comprehensive, Low-Cost Error Detection for GPGPU Cores

Abstract

We have developed and evaluated Argus-G, an error detection scheme for general purpose GPU (GPGPU) cores. Argus-G is a natural extension of the Argus error detection scheme for CPU cores, and we demonstrate how to modify Argus such that it is compatible with GPGPU cores. Using an RTL prototype, we experimentally show that Argus-G can detect the vast majority of injected errors at relatively low performance, area, and power costs.