Optimization of computation methods calculating regional geologic models of ore-bearing isometric structures in Python

Abstract

Numeric regional models are necessary to reconstruct the nature and conditions of formation of i geological structures, to obtain ideas about the evolution of different zones of the crust and upper mantle, the evolution of fold and fault dislocations and predicting expression of manifestation of deep structures on earth’s surface. Numeric reconstructions are based on computational dynamics methods, in which rock is represented as highly viscous liquid. «Markers-in-cell» method is one of the implementations of that approach. We propose to accelerate model ported into Python, as compared to Matlab, without rejecting sequential enumeration of markers. It is shown that the solutions can be either code vectorization (refactoring with binary indices instead of enumeration loops and numpy functions) or acceleration of loops by their pre-compilation or parallel computations. The results obtained can be used in the implementation of numeric geological modeling in Python.

Keywords

• minerageny
• python
• markers-in-cell
• numeric modeling
• petrology