High Performance Multivariate Geospatial Statistics on Manycore Systems
Type
ArticleAuthors
Salvaña, Mary Lai O.
Abdulah, Sameh
Huang, Huang

Ltaief, Hatem

Sun, Ying

Genton, Marc G.

Keyes, David E.

KAUST Department
Statistics ProgramComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Extreme Computing Research Center
Applied Mathematics and Computational Science Program
Office of the President
Date
2021-04-06Permanent link to this record
http://hdl.handle.net/10754/666235
Metadata
Show full item recordAbstract
Modeling and inferring spatial relationships and predicting missing values of environmental data are some of the main tasks of geospatial statisticians. These routine tasks are accomplished using multivariate geospatial models and the cokriging technique, which requires the evaluation of the expensive Gaussian log-likelihood function. This large-scale cokriging challenge provides a fertile ground for supercomputing implementations for the geospatial statistics community as it is paramount to scale computational capability to match the growth in environmental data. In this paper, we develop large-scale multivariate spatial modeling and inference on parallel hardware architectures. To tackle the increasing complexity in matrix operations and the massive concurrency in parallel systems, we leverage low-rank matrix approximation techniques with task-based programming models and schedule the asynchronous computational tasks using a dynamic runtime system. The proposed framework provides both the dense and approximated computations of the Gaussian log-likelihood function. It demonstrates accuracy robustness and performance scalability on a variety of computer systems. Using both synthetic and real datasets, the low-rank matrix approximation shows better performance compared to exact computation, while preserving the application requirements in both parameter estimation and prediction accuracy. We also propose a novel algorithm to assess the prediction accuracy after the online parameter estimation.Citation
Salvana, M., Abdulah, S., Huang, H., Ltaief, H., Sun, Y., Genton, M. M., & Keyes, D. (2021). High Performance Multivariate Geospatial Statistics on Manycore Systems. IEEE Transactions on Parallel and Distributed Systems, 1–1. doi:10.1109/tpds.2021.3071423Sponsors
The authors would like to thank NVIDIA Inc., Cray Inc., and Intel Corp., the Cray Center of Excellence and Intel Parallel Computing Center awarded to the Extreme Computing Research Center (ECRC) at KAUST. For computer time, this research used GPU-based systems as well as Shaheen supercomputer, both hosted at the Supercomputing Laboratory at King Abdullah University of Science and Technology (KAUST).arXiv
2008.07437Additional Links
https://ieeexplore.ieee.org/document/9397281/https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9397281
ae974a485f413a2113503eed53cd6c53
10.1109/TPDS.2021.3071423