https://hgpu.org/?p=12591
Accelerated Matrix Element Method with Parallel Computing