“…An early attempt to use parallel computation for Monte Carlo simulation is Chong and Hendry (1986), while Swann (2002) develops a parallel implementation of maximum likelihood estimation. Creel and Goffe (2008) low diffusion of this technology in economics and econometrics, according to Creel (2005), is mainly due to two issues, which are the high cost of the hardware, e.g., a computing cluster, and the steep learning curve of dedicated programming languages such as CUDA (compute unified device architecture, see NVIDIA Corporation 2010), OpenCL (Khronos OpenCL Working Group 2009), Thrust (Hoberock and Bell 2011) and C++ AMP (C++ accelerated massive parallelism, see Gregory and Miller 2012). Table 1 compares different currently available GPGPU approaches.…”