To optimize code effectively, compilers must deal with memory dependencies. However, the state-of-the-art heuristics available in the literature to track memory dependencies are inherently imprecise and computationally expensive. Consequently, the most advanced code transformations that compilers have today are ineffective when applied on real-world programs. The goal of this paper is to solve this conundrum through dynamic disambiguation of pointers. We provide different ways to determine at runtime when two memory locations can overlap. We then produce two versions of a code region: one that is aliasing-free - hence, easy to optimize - and another that is not. Our checks let us safely branch to the optimizable region. We have applied these ideas on Polly-LLVM, a loop optimizer built on top of the LLVM compilation infrastructure. Our experiments indicate that our method is precise, effective and useful: we can disambiguate every pair of pointer in the loop intensive Polybench benchmark suite. The result of this precision is code quality: the binaries we generate are 10% faster than those that Polly-LLVM produces without our optimization, at the -O3 optimization level of LLVM.
Fri 30 OctDisplayed time zone: Eastern Time (US & Canada) change
10:30 - 12:00 | 9. Compilation & Dynamic AnalysisOOPSLA at Grand Station 1 Chair(s): Frank Tip Samsung Research America | ||
10:30 22mTalk | Runtime Pointer Disambiguation OOPSLA Pericles Rafael Alves Federal University of Minas Gerais, Brazil, Fabian Gruber INRIA, France, Johannes Doerfert Saarland University, Alexandros Labrineas INRIA, France, Tobias Grosser ETH Zurich, Switzerland, Fabrice Rastello INRIA, France, Fernando Magno Quintão Pereira Federal University of Minas Gerais, Brazil Link to publication | ||
10:52 22mTalk | Performance Problems You Can Fix: A Dynamic Analysis of Memoization Opportunities OOPSLA Luca Della Toffola ETH Zurich, Switzerland, Michael Pradel TU Darmstadt, Germany, Thomas Gross ETH Zurich, Switzerland DOI | ||
11:15 22mTalk | RAIVE: Runtime Assessment of Floating-Point Instability by Vectorization OOPSLA Wen-Chuan Lee Purdue University, USA, Tao Bao Purdue University, USA, Yunhui Zheng IBM Research, Xiangyu Zhang Purdue University, USA, Keval Vora University of California at Riverside, USA, Rajiv Gupta University of California at Riverside, USA DOI | ||
11:37 22mTalk | Automated Backward Error Analysis for Numerical Code OOPSLA Zhoulai Fu University of California at Davis, USA, Zhaojun Bai University of California at Davis, USA, Zhendong Su University of California at Davis, USA DOI |