Low-Cost Software-Implemented Error Detection Technique

Abstract

In this paper, a software behavior-based technique is presented to detect control-flow error. The analysis of a key point leads to introduce the proposed technique: effective reduction of the overheads of control-flow checking statements through finding the best sequence of signatures for assigning to consecutive basic-blocks. To evaluate the proposed technique, a functional full-system simulator is used, and several well-known benchmarks are implemented on a quad-core shared memory processor. The experimental results, with regarding to both detection coverage and overheads, demonstrate that on average about 94% of the control-flow errors can be detected by the proposed technique, more efficiently.