AnACor2.0: a GPU-accelerated open-source software package for analytical absorption corrections in X-ray...

Area of Science:

• Crystallography and Materials Science
• Computational Science and Engineering

Background:

• Analytical absorption corrections are essential for processing diffraction data from highly absorbing crystalline samples, particularly in long-wavelength crystallography.
• Traditional empirical corrections are often inadequate, and existing analytical methods can be computationally intensive due to ray-tracing complexities.

Purpose of the Study:

• To develop and evaluate AnACor2.0, an accelerated software package for calculating analytical absorption corrections.
• To significantly reduce the computational time required for absorption correction calculations without compromising accuracy.

Main Methods:

• AnACor2.0 employs ray-tracing of X-ray paths through a voxelized 3D sample model.
• Acceleration is achieved through systematic sampling of crystal voxels and modifications to standard ray-tracing algorithms.
• The bisection method (reducing complexity to O(log2 n)) and gridding with interpolation are utilized, alongside optimized CUDA implementations for NVIDIA GPUs.

Main Results:

• Execution time for analytical absorption corrections was reduced by up to 175x compared to previous methods.
• Absorption factor calculations for datasets like insulin were completed in under 10 seconds.
• Systematic sampling yielded accurate results with minimal variance (mean difference ≤ 2% for absorption factors, ≤ 1% for anomalous peak heights).

Conclusions:

• AnACor2.0 effectively refines and accelerates the analytical absorption correction process.
• Innovative sampling and computational techniques ensure high efficiency and accuracy for crystallographic data analysis.
• The software provides a significant advancement for handling highly absorbing samples in X-ray diffraction studies.