With Work Package 2, things get interesting: we will launch an attempt to implement the AcceleratedLattice, a version of the MultiBlockLattice which is GPU ready, with a structure-of-array and AA-pattern implementation. On the side of community involvement, we encourage you
- To run the benchmark cases we have implemented so far on your own hardware, for now as a CPU version with MPI. Please do report your measured performance to us, by posting it on the forum or sending me a message, and we will add it to the shared spreadsheet. We hope to be able to add GPU benchmarks to the same sheet soon.
- To keep learning about the possibility to run programs on GPU with help of the parallel STL. I will also give a presentation on this topic later today at the AMS seminar.
Finally, everything related to this project is summarized on the project page.