Assessing One-to-One Parallelism Levels Mapping for OpenMP Offloading to GPUs
Authors:
Open Access
- 4 February 2017
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 68-73
Abstract
The proliferation of accelerators in modern clusters makes efficient coprocessor programming a key requirement if application codes are to achieve high levels of performance with acceptable energy consumption on such platforms. This has led to considerable effort to provide suitable programming models for these accelerators, especially within the OpenMP community. While OpenMP 4.5 offers a rich set of directives, clauses and runtime calls to fully utilize accelerators, an efficient implementation of OpenMP 4.5 for GPUs remains a non-trivial task, given their multiple levels of thread parallelism. In this paper, we describe a new implementation of the corresponding features of OpenMP 4.5 for GPUs based on a one-to-one mapping of its loop hierarchy parallelism to the GPU thread hierarchy. We assess the impact of this mapping, in particular the use of GPU warps to handle innermost loop execution, on the performance of GPU execution via a set of benchmarks that include a version of the NAS parallel benchmarks specifically developed for this research; we also used the Matrix-Matrix multiplication, Jacobi, Gauss and Laplacian kernels.Keywords
This publication has 12 references indexed in Scilit:
- Integrating GPU support for OpenMP offloading directives into Clang2015
- Coordinating GPU Threads for OpenMP 4.0 in LLVM2015
- Compiling a High-Level Directive-Based Programming Model for GPGPUs2014
- OpenMP Task Scheduling Analysis via OpenMP Runtime API and Tool Visualization2014
- Extending the OpenSHMEM Analyzer to Perform Synchronization and Multi-valued Analysis2014
- Experiences Developing the OpenUH Compiler and Runtime InfrastructureInternational Journal of Parallel Programming, 2012
- OpenMP for Accelerators2011
- Enabling Locality-Aware Computations in OpenMPScientific Programming, 2010
- OpenUH: an optimizing, portable OpenMP compilerConcurrency and Computation: Practice and Experience, 2007
- Dragon: an Open64-based interactive program analysis tool for large applications2004