Abstract

While many dynamic branch prediction schemes have been proposed and studied, few have been compared to static branch prediction. Fewer yet have been implemented side-by-side on the same machine to allow full performance evaluation. The Hewlett-Packard PA-8000 microprocessor implements both a simple dynamic prediction scheme and static prediction, selectable by the application programmer. This paper studies the PA-8000's trade-off between static and dynamic prediction, and the compiler optimizations needed to support an innovative static branch prediction convention while maintaining object code compatibility with earlier revisions of the PA-RISC architecture.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call