Co-Designing an OpenMP GPU Runtime and Optimizations for Near-Zero Overhead Execution
Authors:
- 1 May 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 504-514
Abstract
GPU accelerators are ubiquitous in modern HPC systems. To program them, users have the choice between vendor-specific, native programming models, such as CUDA, which provide simple parallelism semantics with minimal runtime support, or portable alternatives, such as OpenMP, which offer rich parallel semantics and feature an extensive runtime library to support execution. While the operations of such a runtime can easily limit performance and drain resources, it was to some degree regarded an unavoidable overhead. In this work we present a co-design methodology for optimizing applications using a specifically crafted OpenMP GPU runtime such that most use cases induce near-zero overhead. Specifically, our approach exposes runtime semantics and state to the compiler such that optimization effectively eliminating abstractions and runtime state from the final binary. With the help of user provided assumptions we can further optimize common patterns that otherwise increase resource consumption. We evaluated our prototype build on top of the LLVM/OpenMP GPU offloading infrastructure with multiple HPC proxy applications and benchmarks. Comparison of CUDA, the original OpenMP runtime, and our co-designed alternative show that, by our approach, performance is significantly improved and resource consumption is significantly lowered. Oftentimes we can closely match the CUDA implementation without sacrificing the versatility and portability of OpenMP.Keywords
Funding Information
- LLNL (DE-AC52-07NA27344,LLNL-CONF-826728)
This publication has 20 references indexed in Scilit:
- Compiler Optimizations for OpenMP2018
- Efficient Fork-Join on GPUs Through Warp Specialization2017
- LLVM Compiler Implementation for Explicit Parallelization and SIMD Vectorization2017
- On the Performance of Parallel Tasking Runtimes for an Irregular Fast Multipole Method Application2017
- LLVM Framework and IR Extensions for Parallelization, SIMD Vectorization and Offloading2016
- Offloading Support for OpenMP in Clang and LLVM2016
- Grid: A next generation data parallel C++ QCD library2016
- Integrating GPU support for OpenMP offloading directives into Clang2015
- OpenMC: A state-of-the-art Monte Carlo code for research and developmentAnnals of Nuclear Energy, 2015
- Spectral neighbor analysis method for automated generation of quantum-accurate interatomic potentialsJournal of Computational Physics, 2015