When designing modern real-time systems, which have to deliver results at specified deadlines, knowing the worst-case execution time (WCET) of software components is of utmost importance. Although there has been much research in the field of WCET analysis in the last years, with a focus on improving the accuracy of processor models and WCET-calculation methods, researchers have paid little attention to exploring the impact of the instruction set architecture (ISA) on the time predictability of the code executing on a given real-time processor. In this paper we explore ISA extensions that allow compilers to generate highly time-predictable code. To this end, an existing instruction set has been extended by a number of instructions, and the LLVM compiler framework has been adapted to use these new instructions in its assembler-code generator. The timing behavior of the generated code has been evaluated by means of an instruction-set simulator. The results of the experiments allowed us to identify a promising combination of the newly introduced instructions. The use of these instructions leads to a reduction of the number of branches in the assembler code, thus improving time predictability while still providing competitive worst-case timing.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.