Boosting beyond static scheduling in a superscalar processor

摘要

This paper describes a superscalar processor that combines the best qualities of static and dynamic instruction scheduling to increase the performance of non-numerical applications. The architecture performs all instruction scheduling statically to take advantage of the compiler's ability to efficiently schedule operations across many basic blocks. Since the conditional branches in non-numerical code are highly data dependent, the architecture introduces the concept of boosted instructions, instructions that are committed conditionally upon the result of later branch instructions. Boosting effectively removes the dependencies caused by branches and makes the scheduling of side-effect instructions as simple as those that are side-effect free. For efficiency, boosting is supported in the hardware by shadow structures that temporarily hold the side effects of boosted instructions until the conditional branches that the boosted instructions depend upon are executed. When the branch condition is determined, the buffered side effects are either committed or squashed. The limited static scheduler in our evaluation system shows that a 1.6-times speedup over scalar code is achievable by boosting instructions above only a single conditional branch. This performance is similar to the performance of a pure dynamic scheduler.