The ever-increasing heterogeneity in supercomputing applications has given rise to complex compute node architectures offering multiple, heterogeneous levels of massive parallelism. As a result, the ‘X’ in MPI+X demands more focus. Exploiting the maximum available parallelism out of such systems necessitates sophisticated programming approaches such as OpenMP offloading, OpenACC, OneAPI, Kokkos and Raja among several others that can provide scalable as well as portable solutions without compromising performance. A programmer’s expectation from the software community is to deliver solutions without the need to duplicate effort.
Toward this end, the workshop will highlight the improvements to the state-of-the-art through accepted papers and prompt keynote/panel discussion, drawing the community’s attention to key areas that will facilitate the transition to accelerator-based high-performance computing (HPC). The workshop aims to showcase all aspects of innovative high-level language features, lessons learned while using directives and abstractions to migrate scientific legacy code (among others).