FlipSphere: A Software-Based DRAM Error Detection and Correction Library for HPC

Abstract

Proposed exascale systems will present considerable resiliency challenges. In particular, DRAM soft-errors, or bit-flips, are expected to greatly increase due to much higher memory density of these systems. Current hardware-based fault-tolerance methods cannot cope by itself with the expected soft error frequency rate. As a result, additional software is needed to address this challenge. We introduce FlipSphere, a tunable, transparent silent data corruption detection and correction library for HPC applications that is first in its class to use hardware accelerators such as the Intel Xeon Phi MIC to increase application resiliency. FlipSphere provides comprehensive silent data corruption protection for application memory by implementing on-demand page integrity verification coupled with software-based error correcting codes that allow for automatic error recovery in the event of memory failures. We investigate the trade-off of hardware resources for resilience and find that up to 90% of memory may be guarded with error detection at a 40% performance overhead.