GPU offloading of a large-scale gyrokinetic particle-in-cell Fortran code: From OpenACC to OpenMP
Event Type
Workshop
Online Only
Accelerator-based Architectures
Parallel Programming Languages and Models
Performance
State of the Practice
W
TimeSunday, 14 November 202112pm - 12:30pm CST
LocationOnline
DescriptionGPU offloading of a large-scale gyrokinetic particle-in-cell Fortran code is converted from using OpenACC to using OpenMP. Particle pushing and deposition are completely offloaded to GPU. Performance is compared between CPU and GPU, and between OpenACC and OpenMP. Good weak scaling (increasing particle number with fixed grid number) is obtained. Issues encountered when porting OpenMP GPU offloading are discussed.
Author/Presenters
