Ball-Larus Path Profiling Across Multiple Loop iterations

TL;DR

This paper introduces a novel method for path profiling across multiple loop iterations that extends Ball-Larus path profiling to cyclic paths, enabling efficient analysis of execution patterns in Java benchmarks.

Contribution

It presents a new data structure-based approach that allows profiling concatenated paths of up to k Ball-Larus acyclic paths, overcoming previous limitations with minimal performance impact.

Findings

Faster than traditional Ball-Larus profiling in many cases

Produces compact representations of cyclic paths for large k

Effective on diverse Java benchmarks

Abstract

Identifying the hottest paths in the control flow graph of a routine can direct optimizations to portions of the code where most resources are consumed. This powerful methodology, called path profiling, was introduced by Ball and Larus in the mid 90s and has received considerable attention in the last 15 years for its practical relevance. A shortcoming of Ball-Larus path profiling was the inability to profile cyclic paths, making it difficult to mine interesting execution patterns that span multiple loop iterations. Previous results, based on rather complex algorithms, have attempted to circumvent this limitation at the price of significant performance losses already for a small number of iterations. In this paper, we present a new approach to multiple iterations path profiling, based on data structures built on top of the original Ball-Larus numbering technique. Our approach allows it to profile all executed paths obtained as a concatenation of up to k Ball-Larus acyclic paths, where k is a user-defined parameter. An extensive experimental investigation on a large variety of Java benchmarks on the Jikes RVM shows that, surprisingly, our approach can be even faster than Ball-Larus due to fewer operations on smaller hash tables, producing compact representations of cyclic paths even for large values of k.