Wed 7 Nov 2018 11:37 - 12:00 at Studio 2 - Parallelism and Performance Chair(s): Arjun Guha

Modern microprocessors are equipped with single instruction multiple data (SIMD) or vector instruction sets which allow compilers to exploit superword level parallelism (SLP), a type of fine-grained parallelism. Current SLP auto-vectorization techniques use heuristics to discover vectorization opportunities in high-level language code. These heuristics are fragile, local and typically only present one vectorization strategy that is either accepted or rejected by a cost model. We present goSLP, a novel SLP auto-vectorization framework which solves the statement packing problem in a pairwise optimal manner. Using an integer linear programming (ILP) solver, goSLP searches the entire space of statement packing opportunities for a whole function at a time, while limiting total compilation time to a few minutes. Furthermore, goSLP optimally solves the vector permutation selection problem using dynamic programming. We implemented goSLP in the LLVM compiler infrastructure, achieving a geometric mean speedup of 7.58% on SPEC2017fp, 2.42% on SPEC2006fp and 4.07% on NAS benchmarks compared to LLVM’s existing SLP auto-vectorizer.

Wed 7 Nov

splash-2018-OOPSLA
10:30 - 12:00: OOPSLA - Parallelism and Performance at Studio 2
Chair(s): Arjun GuhaUniversity of Massachusetts Amherst
splash-2018-OOPSLA10:30 - 10:52
Talk
Nachshon CohenEPFL, Switzerland
splash-2018-OOPSLA10:52 - 11:15
Talk
Benoit DalozeJKU Linz, Austria, Arie TalTechnion, Stefan MarrUniversity of Kent, Hanspeter MössenböckJKU Linz, Austria, Erez PetrankTechnion
Pre-print
splash-2018-OOPSLA11:15 - 11:37
Talk
Remigius MeierETH Zurich, Switzerland, Armin RigoPyPy.org, Switzerland, Thomas GrossETH Zurich
splash-2018-OOPSLA11:37 - 12:00
Talk