We describe MGARD, a software providing MultiGrid Adaptive Reduction for floating-point scientific data on structured and unstructured grids. With exceptional data compression capability and precise error control, MGARD addresses a wide range of requirements, including storage reduction, high-performance I/O, and in-situ data analysis. It features a unified application programming interface (API) that seamlessly operates across diverse computing architectures. MGARD has been optimized with highly-tuned GPU kernels and efficient memory and device management mechanisms, ensuring scalable and rapid operations.
@article{arxiv.2401.05994,
title = {MGARD: A multigrid framework for high-performance, error-controlled data compression and refactoring},
author = {Qian Gong and Jieyang Chen and Ben Whitney and Xin Liang and Viktor Reshniak and Tania Banerjee and Jaemoon Lee and Anand Rangarajan and Lipeng Wan and Nicolas Vidal and Qing Liu and Ana Gainaru and Norbert Podhorszki and Richard Archibald and Sanjay Ranka and Scott Klasky},
journal= {arXiv preprint arXiv:2401.05994},
year = {2024}
}