An empirical evaluation of chains of recurrences for array dependence testing

Abstract

Code restructuring compilers rely heavily on program analysis techniques to automatically detect data dependences between program statements. Dependences between statement instances in the iteration space of a loop nest impose ordering constraints that must be preserved in order to produce valid optimized, vectorized, and parallelized loop nests. This paper evaluates a new approach for fast and accurate nonlinear array dependence testing using Chains of Recurrences (CRs). A flow-sensitive loop analysis algorithm is presented for constructing the CR forms of array index expressions. Unlike other approaches, the CR forms are directly integrated into a standard dependence test to solve nonlinear CR-based dependence equations. To study the coverage and performance of the proposed CR-based enhancements of a standard test, we chose the inexact Banerjee test. We implemented a new CR-based Banerjee test in the Polaris compiler and compared the results to the Omega test and Range test on a set of SPEC and LAPACK Benchmark programs. The experimental results suggest that a CR enhancement can dramatically increase the effectiveness of a dependence test without a significant cost increase. More surprisingly, the findings indicate that the enhanced test exceeds the capabilities of the Omega and Range tests for many nonlinear dependence relations detected in the PERFECT Club and LAPACK Benchmark programs.

Computer-Aided Recoding to Create Structured and Analyzable System Models

“…This is in contrast to optimizing compilers or synthesis tools, where data dependence testing is integrated into a fully automatic process. Therefore, advanced dependence analysis techniques, such as the chains-of-recurrences algebra [Engelen et al 2004], which are very effective in compilers [Birch et al 2006], are not suitable for our designer-controlled approach.…”

Discovering Maximum Parallelization Using Advanced Data Dependence Analysis

“…Using Chains of Recurrences as a tool to analyze induction variables is a novel approach to IV analysis, to in turn, benefit dependence testing. Previous work [5] has casted CR-forms for direct application of the Banerjee-BoundsTest, a simple test based on the bounds analysis of indexexpressions. That seminal work was limited to dependence testing between array subscripts of assignments and showed limited parallelization returns.…”