Regression Test Suites Optimization for Application-specific Instruction-set Processors and Their Use for Dependability Analysis

verification, dependability, regression tests

Nowadays, Application-specific Instruction-set Processors (ASIPs) are an important part of embedded systems, especially those which are utilized in Internet of Things (IoT). They can be optimized for high performance, small area or low power consumption by changes in their instruction set or by various changes in their hardware architecture. With fast prototyping of ASIPs, another very important issue arises. It is fast verification of such changes, ideally with optimized test suites, so comprehensive and long-running verification after every change can be avoided. In this paper, a new technique for building optimized regression suites for ASIPs is proposed which is based on the genetic algorithm. Experiments show that using this technique, reduction of the original test suite from first phase verification is significant, the new optimized regression suite remains strong in checking key functionality of an ASIP and a considerable improvement in the verification runtime can be achieved. Moreover, the paper shows a proposal of multi-domain application of the regression suites in evaluating dependability of FPGA-based ASIP prototypes.