Implementation and validation of X-ray diffraction imaging systems in MC-GPU.

MC-GPU is a GPU-based Monte Carlo simulator that realizes enhanced computational speeds by exploiting the parallelism of the GPU architecture. Although MC-GPU is capable in theory of modeling XRD imaging systems, existing MC-GPU configurations are limited to simulating water or water-like materials. Here, we demonstrate MC-GPU's ability to rapidly simulate arbitrary materials and model real-world experimental X-ray diffraction (XRD) imaging systems. We found a mean-squared error (MSE) < 1.50E-02 and cross-correlation (CC) > 0.99 (with a speed-up of 1075x for MC-GPU compared to GEANT4) for all comparisons in the unit-testing/cross-validation phase. Additionally, we showcased MC-GPU's ability to model realistic experimental systems by modelling a fan beam coded aperture XRD imaging system. These results demonstrate MC-GPU's utility in rapidly simulating complex, real-world XRD imaging systems, which is critical in applications ranging from medical to security imaging. [ABSTRACT FROM AUTHOR]