An Approach to Analyze Effects of Soft Errors from Dynamic Software Behaviors

The fatal effect of soft errors is that the execution of software is influenced. Dynamic implementation for software-based soft error tolerance method which can protect more types of codes can cover more soft errors. Based on requirement of dynamic soft error tolerance to know dynamic software behaviors, this paper proposes an approach to analyze dynamic software behaviors under the effects of soft errors on hardware. In our analyses, we use a special program model which combines abstract computing on the high level with instructions on the low level. On the high level, programs are divided with the granularity of function. On the low level, every function is implemented by instructions. Those effects of soft errors on hardware to instructions on the low level are passed to the computing results of function on the high level. Backed by the computing results of function on the high level, those instruction errors which can cause wrong program outcome are analyzed in function. Based on those instruction errors in function, program reliability model is built under the effects of soft errors. From the reliability model, we can know different behaviors of different software under the effects of soft errors. Finally, in our fault injection experiments, those experimental results demonstrate our analyses of soft errors to dynamic software behaviors.