Аннотация:The emergence of an increasing number of tools for processing data on the GPU makes us think about optimizing the processing of a large array of coefficients. This article examines the performance of the implementation of the calculation of the absorption coefficient according to the data of the Hitran base on the GPU. An algorithm for calculating the absorption coefficient for water vapor is presented. The results show a significant increase in productivity due to parallelization of sequential calculation steps, which makes it possible to use this approach to speed up the calculation of the global atmospheric model.