research-article

Specialized Kernels for Optimizing GPU Offload in OpenMP

Authors:

Dhruva Chakrabarti,

Gregory Rodgers,

Carlo Bertolli,

Gheorghe-Teodor Bercea,

Jan-Patrick Lehr,

Ron LiebermanAuthors Info & Claims

SC-W '23: Proceedings of the SC '23 Workshops of The International Conference on High Performance Computing, Network, Storage, and Analysis

Pages 1918 - 1928

https://doi.org/10.1145/3624062.3624605

Published: 12 November 2023 Publication History

Abstract

Programming models for general purpose GPU (GPGPU) computing include grid and non-grid languages. Grid languages like CUDA and HIP map directly to the GPU hardware and can extract high performance from applications. However, this low-level programming approach makes them more difficult to program than non-grid languages such as C, C++, and Fortran with OpenMP target offload. Furthermore, grid languages often have more portability issues than non-grid languages. However, code generated from non-grid languages using automatic compiler and runtime techniques often incur higher overhead while generating GPU kernels for target regions.

This paper discusses compiler and runtime techniques to generate specialized, high-performance kernels for OpenMP target regions in certain common situations. We outline the conditions under which specialized kernels are generated for OpenMP target regions, both with and without reduction clauses. Experimental results on AMD GPUs indicate that a large percentage of OpenMP target regions are amenable to specialization and consequent improvement in performance.

Supplemental Material

MP4 File - Conference presentation recording

Recording of "Specialized Kernels for Optimizing GPU Offload in OpenMP" presentation at the Tenth Workshop on Accelerator Programming and Directives (WACCPD 2023)

Download
129.05 MB

References

[1]

Carlo Bertolli, Samuel F. Antao, Gheorghe-Teodor Bercea, Arpith C. Jacob, Alexandre E. Eichenberger, Tong Chen, Zehra Sura, Hyojin Sung, Georgios Rokos, David Appelhans, and Kevin O’Brien. 2015. Integrating GPU Support for OpenMP Offloading Directives into Clang. In Proceedings of the Second Workshop on the LLVM Compiler Infrastructure in HPC (Austin, Texas) (LLVM ’15). 1–11. https://doi.org/10.1145/2833157.2833161

Digital Library

[2]

Carlo Bertolli, Samuel F. Antao, Alexandre E. Eichenberger, Kevin O’Brien, Zehra Sura, Arpith C. Jacob, Tong Chen, and Olivier Sallenave. 2014. Coordinating GPU Threads for OpenMP 4.0 in LLVM. In 2014 LLVM Compiler Infrastructure in HPC. 12–21. https://doi.org/10.1109/LLVM-HPC.2014.10

Digital Library

[3]

OpenMP Architecture Review Board. 2023. The OpenMP API specification for parallel programming.https://www.openmp.org/specifications/

[4]

Holger Brunst, Sunita Chandrasekaran, Florina M. Ciorba, Nick Hagerty, Robert Henschel, Guido Juckeland, Junjie Li, Verónica G. Melesse Vergara, Sandra Wienke, and Miguel Zavala. 2022. First Experiences in Performance Benchmarking with the New SPEChpc 2021 Suites. In 2022 22nd IEEE International Symposium on Cluster, Cloud and Internet Computing (CCGrid). 675–684. https://doi.org/10.1109/CCGrid54584.2022.00077

[5]

Joshua Hoke Davis, Christopher Daley, Swaroop Pophale, Thomas Huber, Sunita Chandrasekaran, and Nicholas J. Wright. 2021. Performance Assessment of OpenMP Compilers Targeting NVIDIA V100 GPUs. In Accelerator Programming Using Directives, Sridutt Bhalachandra, Sandra Wienke, Sunita Chandrasekaran, and Guido Juckeland (Eds.). Springer International Publishing, 25–44.

[6]

Tom Deakin, James Price, Matthew Martineau, and Simon McIntosh-Smith. 2018. Evaluating Attainable Memory Bandwidth of Parallel Programming Models via BabelStream. International Journal of Computational Science and Engineering 17, 3 (Jan 2018), 247–262. https://doi.org/10.1504/IJCSE.2018.095847

[7]

Johannes Doerfert, Jose Manuel Monsalve Diaz, and Hal Finkel. 2019. The TRegion Interface and Compiler Optimizations for OpenMP Target Regions. In OpenMP: Conquering the Full Hardware Spectrum: 15th International Workshop on OpenMP, IWOMP 2019, Auckland, New Zealand, September 1113, 2019, Proceedings (Auckland, New Zealand). 153–167. https://doi.org/10.1007/978-3-030-28596-8_11

Digital Library

[8]

Johannes Doerfert, Atemn Patel, Joseph Huber, Shilei Tian, Jose M Monsalve Diaz, Barbara Chapman, and Giorgis Georgakoudis. 2022. Co-Designing an OpenMP GPU Runtime and Optimizations for Near-Zero Overhead Execution. In 2022 IEEE International Parallel and Distributed Processing Symposium (IPDPS). 504–514. https://doi.org/10.1109/IPDPS53621.2022.00055

[9]

Simon Garcia De Gonzalo, Sitao Huang, Juan Gómez-Luna, Simon Hammond, Onur Mutlu, and Wen-mei Hwu. 2019. Automatic Generation of Warp-Level Primitives and Atomic Instructions for Fast and Portable Parallel Reduction on GPUs. In 2019 IEEE/ACM International Symposium on Code Generation and Optimization (CGO). 73–84. https://doi.org/10.1109/CGO.2019.8661187

[10]

Joseph Huber, Melanie Cornelius, Giorgis Georgakoudis, Shilei Tian, Jose M Monsalve Diaz, Kuter Dinel, Barbara Chapman, and Johannes Doerfert. 2022. Efficient Execution of OpenMP on GPUs. In 2022 IEEE/ACM International Symposium on Code Generation and Optimization (CGO). 41–52. https://doi.org/10.1109/CGO53902.2022.9741290

Digital Library

[11]

LLVM 2023. The LLVM Compiler Infrastructure.https://llvm.org

[12]

Justin Luitjens. 2014. Faster Parallel Reductions on Kepler. https://developer.nvidia.com/blog/faster-parallel-reductions-kepler

[13]

ROCm 2023. AMD ROCmTM Platform.https://rocm.docs.amd.com/en/latest/

[14]

Ben Sander. 2016. AMD GCN Assembly: Cross-Lane Operations. https://gpuopen.com/learn/amd-gcn-assembly-cross-lane-operations/

[15]

SPEC 2023. Standard Performance Evaluation Corporation. https://www.spec.org/

[16]

SPEC ACCEL 2019. Standard Performance Evaluation Corporation ACCEL benchmark suite. https://www.spec.org/accel/

[17]

SPEChpc 2021. Standard Performance Evaluation Corporation SPEChpc 2021 benchmark suite. https://www.spec.org/hpc2021/

[18]

Shilei Tian, Jon Chesterfield, Johannes Doerfert, and Barbara Chapman. 2021. Experience Report: Writing a Portable GPU Runtime with OpenMP 5.1. In OpenMP: Enabling Massive Node-Level Parallelism: 17th International Workshop on OpenMP, IWOMP 2021, Bristol, UK, September 1416, 2021, Proceedings (Bristol, United Kingdom). 159–169. https://doi.org/10.1007/978-3-030-85262-7_11

Digital Library

Cited By

Lehr JHalkenhäuser MChakrabarti DIslam SPalermo DLieberman R(2024)ompTest – Unit Testing with OMPTSC24-W: Workshops of the International Conference for High Performance Computing, Networking, Storage and Analysis10.1109/SCW63240.2024.00031(196-203)Online publication date: 17-Nov-2024
https://doi.org/10.1109/SCW63240.2024.00031

Index Terms

Specialized Kernels for Optimizing GPU Offload in OpenMP
1. Software and its engineering
  1. Software notations and tools
    1. Compilers
      1. Runtime environments

Recommendations

Offload Compiler Runtime for the Intel® Xeon Phi Coprocessor
IPDPSW '13: Proceedings of the 2013 IEEE 27th International Symposium on Parallel and Distributed Processing Workshops and PhD Forum

The Intel® Xeon Phi™ coprocessor platform has a new software stack that enables new programming models. One such model is offload of computation from a host processor to a coprocessor that is a fully-capable Intel® Architecture CPU, namely, the Intel® ...
Performance of CPU/GPU compiler directives on ISO/TTI kernels

GPUs are slowly becoming ubiquitous devices in High Performance Computing, as their capabilities to enhance the performance per watt of compute intensive algorithms as compared to multicore CPUs have been identified. The primary shortcoming of a GPU is ...
A Virtual GPU as Developer-Friendly OpenMP Offload Target
ICPP Workshops '21: 50th International Conference on Parallel Processing Workshop

While parallel programming is hard, programming accelerators has always been even more complicated. One fundamental reason is the lack of mature tooling that can be used to inspect a program that executes on two different architectures. As GPU software ...

Comments

Information & Contributors

Information

Published In

cover image ACM Other conferences

SC-W '23: Proceedings of the SC '23 Workshops of The International Conference on High Performance Computing, Network, Storage, and Analysis

November 2023

2180 pages

ISBN:9798400707858

DOI:10.1145/3624062

Copyright © 2023 ACM.

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected].

Publisher

Association for Computing Machinery

New York, NY, United States

Publication History

Published: 12 November 2023

Permissions

Request permissions for this article.

Request Permissions

Check for updates

Author Tags

Qualifiers

Research-article
Research
Refereed limited

Conference

SC-W 2023

SC-W 2023: Workshops of The International Conference on High Performance Computing, Network, Storage, and Analysis

November 12 - 17, 2023

CO, Denver, USA

Contributors

Other Metrics

View Article Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

1
Total Citations
View Citations
81
Total Downloads

Downloads (Last 12 months)54
Downloads (Last 6 weeks)5

Reflects downloads up to 27 Jan 2025

Other Metrics

View Author Metrics

Citations

Cited By

Lehr JHalkenhäuser MChakrabarti DIslam SPalermo DLieberman R(2024)ompTest – Unit Testing with OMPTSC24-W: Workshops of the International Conference for High Performance Computing, Networking, Storage and Analysis10.1109/SCW63240.2024.00031(196-203)Online publication date: 17-Nov-2024
https://doi.org/10.1109/SCW63240.2024.00031

View Options

Login options

Check if you have access through your login credentials or your institution to get full access on this article.

Full Access

Get this Publication

View options

PDF

View or Download as a PDF file.

eReader

View online with eReader.

HTML Format

View this article in HTML Format.

Figures

Tables

Media

View Table of Conten