Processor Models For Instruction Scheduling using Constraint Programming

Abstract: Instruction scheduling is one of the most important optimisations performed when producing code in a compiler. The problem consists of finding a minimum length schedule subject to latency and different resource constraints. This is a hard problem, classically approached by heuristic algorithms. In the last decade, research interest has shifted from heuristic to potentially optimal methods. When using optimal methods, a lot of compilation time is spent searching for an optimal solution. This makes it important that the problem definition reflects the reality of the processor. In this work, a constraint programming approach was used to study the impact that the model detail has on performance. Several models of a superscalar processor were embedded in LLVM and evaluated using SPEC CPU2000. The result shows that there is substantial performance to be gained, over 5% for some programs. The stability of the improvement is heavily dependent on the accuracy of the model.