Challenges Porting a C++ Template-Metaprogramming Abstraction Layer to Directive-based Offloading

TL;DR

This paper discusses porting a complex C++ scientific code to GPU accelerators using directive-based models, addressing challenges in standards and compiler support to maintain portability and performance.

Contribution

It introduces a method for porting a C++ template metaprogramming abstraction layer to directive-based offloading models like OpenACC and OpenMP, with minimal code changes.

Findings

Successful porting of PIConGPU to OpenACC and OpenMP

Identification of conflicts between standards and compiler support

Practical insights into porting complex C++ code to accelerators

Abstract

HPC systems employ a growing variety of compute accelerators with different architectures and from different vendors. Large scientific applications are required to run efficiently across these systems but need to retain a single code-base in order to not stifle development. Directive-based offloading programming models set out to provide the required portability, but, to existing codes, they themselves represent yet another API to port to. Here, we present our approach of porting the GPU-accelerated particle-in-cell code PIConGPU to OpenACC and OpenMP target by adding two new backends to its existing C++-template metaprogramming-based offloading abstraction layer alpaka and avoiding other modifications to the application code. We introduce our approach in the face of conflicts between requirements and available features in the standards as well as practical hurdles posed by immature compiler support.