Optimization of Finite-Differencing Kernels for Numerical Relativity Applications
A simple optimization strategy for the computation of 3D finite-differencing kernels on many-cores architectures is Phone Holder proposed.The 3D finite-differencing computation is split direction-by-direction and exploits two level of parallelism: in-core vectorization and multi-threads shared-memory parallelization.The main Insulated Delivery Bags