"The ECHO code for General-Relativistic Magneto-Hydrodynamics is used to model instabilities, turbulence, propagation of waves, stellar winds and magnetospheres, and astrophysical processes around black holes. DPEcho, our MPI-parallel rewriting of Fortran-based ECHO, is built around native SYCL* features (Unified Shared Memory, device queues, and mallocs) to take advantage of the capabilities of heterogeneous computing and of the latest GPU architectures. Thanks also to the synergies with Intel® MPI Library, the DPEcho can be executed on an LRZ SuperMUC-NG HPC Cluster based on an Intel® Xeon® processor (formerly code named Skylake). Comparing the DPEcho versus old ECHO on this Intel Xeon processor-based system, we experienced a 2-3 times performance speed up. Using the Intel® Data Center GPU Max Series, we achieved a further speedup by a factor of 9-10 compared to running DPEcho on an Intel Xeon CPU."
1