CaDS Seminar 2025 - June 3

Dr. Daniel Rohe (SDL Quantum Materials)

Migrating from CPU to CPU+GPU - A somewhat unusual example

Abstract:

A short report is given how a CPU-oriented parallel code which relies on high-level parallelism via C++-objects and OpenMP was ported to GPU by means of OpenACC and undocumented features of the Nvidia HPC toolkit, to make use of CPU and GPU concurrently.

Last Modified: 02.06.2025