Clustering case statements for indirect branch predictors

TL;DR

This paper introduces an efficient O(nlogn) algorithm for compiling switch statements into jump tables, optimizing them for hardware branch predictor performance by limiting table size.

Contribution

It improves existing algorithms by reducing complexity from O(n^2) to O(nlogn) and incorporates a size bound for jump tables to enhance predictor efficiency.

Findings

Reduced algorithm complexity from O(n^2) to O(nlogn)

Implemented size bounds for jump tables

Enhanced hardware branch prediction compatibility

Abstract

We present an O(nlogn) algorithm to compile a switch statement into jump tables. To generate jump tables that can be efficiently predicted by current hardware branch predictors, we added an upper bound on the number of entries for each table. This modification of the previously best known algorithm reduces the complexity from O(n^2) to O(nlogn).