User: Guest  Login
Document type:
Buchbeitrag
Author(s):
Mario Wille; Tobias Weinzierl; Gonzalo Brito Gadeschi; Michael Bader
Title:
Efficient GPU Offloading with OpenMP for a Hyperbolic Finite Volume Solver on Dynamically Adaptive Meshes
Pages contribution:
65-85
Chapter contribution:
HPC Algorithms and Applications
Abstract:
We identify and show how to overcome an OpenMP bottleneck in the administration of GPU memory. It arises for a wave equation solver on dynamically adaptive block-structured Cartesian meshes, which keeps all CPU threads busy and allows all of them to offload sets of patches to the GPU. Our studies show that multithreaded, concurrent, non-deterministic access to the GPU leads to performance breakdowns, since the GPU memory bookkeeping as offered through OpenMP’s map clause, i.e., the allocation an...     »
Book title:
High Performance Computing
Book subtitle:
38th International Conference, ISC High Performance 2023, Hamburg, Germany, May 21–25, 2023, Proceedings
Volume:
Lecture Notes in Computer Science
Edition:
13948
Publisher:
Springer
Date of publication:
10.05.2023
Year:
2023
Year / month:
2023-05
Month:
May
Pages:
65-85
Print-ISBN:
978-3-031-32040-8
E-ISBN:
978-3-031-32041-5
Reviewed:
ja
Language:
en
DOI:
doi:https://doi.org/10.1007/978-3-031-32041-5_4
Notes:
https://doi.org/10.5281/zenodo.7741217
TUM Institution:
Department of Informatics
 BibTeX