Shift with MPI_Sendrecv

• Performs reasonably well; simpler than many other approaches

• T3D performance is ok but other approaches are better

Previous slide

Next slide

Back to first slide

View graphic version