Optimization of Parallel Particle-to-Grid Interpolation on Leading Multicore Platforms