Synchrotron-based x-ray micro-tomography systems often suffer from severe ring artifacts in reconstructed images. This article presents an implementation of superior techniques for eliminating these artifacts using cuda and high-performance computing (hpc) on a multi-gpu system. The proposed methods are designed to handle various types of stripe artifacts, providing efficient and high-quality results. Performance evaluations demonstrate significant speedup and enhanced artifact removal quality compared to traditional cpu based methods. Current implementations of this method in open packages do not offer acceleration in multi-gpu systems. In this sense, the present work makes a significant contribution to the imaging community by providing an implementation strategy in python/c/cuda, allowing acceleration.