Seismic data stacking cloud acceleration with CUDA, OpenCL and SPITS

• Gustavo Ciotto Pinton UNICAMP
• Edson Borin UNICAMP https://orcid.org/0000-0003-1783-4231

DOI: https://doi.org/10.5753/eradsp.2020.16892

Abstract

Seismic processing algorithms have been important in many industry applications, notably oil and gas exploration. Such methods tend to be computationally expensive mainly due to the big amount of data. In this article, we evaluate an application to search for the environment parameters that maximizes the coherence measure for three different traveltime estimation models (including a brand new implementation for the OCT model), which automatically distributes idempotent and independent tasks to CUDA/OpenCL supporting nodes of a computing cloud. Besides, to avoid performance degradation due to phenomena such as data transfer and cache misses, we introduce a heuristic to select which fraction of the data should be indeed considered.

Keywords: CUDA, OpenCL, Cloud computing, SPITS, Distributed computing

References

Barros, T., Ferrari, R., Krummenauer, R., and Lopes, R. (2015). Differential evolution-based optimization procedure for automatic estimation of the common-reflection surface traveltime parameters. Geophysics, 80(6):WD189–WD200.

Borin, E., Rodrigues, I. L., Novo, A. T., Sacramento, J. D., Breternitz, M., and Tygel, M. (2015). Efficient and fault tolerant computation of partially idempotent tasks. In 14th International Congress of the Brazilian Geophysical Society & EXPOGEF, Rio de Janeiro, Brazil, 3-6 August 2015, pages 367–372. Brazilian Geophysical Society.

Coimbra, T. A., Novais, A., and Schleicher, J. (2012). Offset continuation (oco) ray tracing using oco trajectories. Studia Geophysica et Geodaetica, 56(1):65–82.

da Silva, H. C. (2017). Aceleração de métodos de processamento sísmico com opencl. Master’s thesis, Universidade de Campinas, Campinas.

Faccipieri, J. H. (2012). Separação e processamento de difrações em dados geofísicos de reflexão. Master’s thesis, Universidade de Campinas, Campinas.

Faccipieri, J. H., Coimbra, T. A., and Bloot, R. (2019). Stretch-free generalized normal moveout correction. Geophysical Prospecting, 67(1):52–68.

Gimenes, T. L., Pisani, F., and Borin, E. (2018). Evaluating the performance and cost of acce- lerating seismic processing with cuda, opencl, openacc, and openmp. In 2018 IEEE Interna- tional Parallel and Distributed Processing Symposium (IPDPS), pages 399–408. IEEE.

Mendonça, A., Silveira, D., Alvarenga, D., and Barreira, G. (2019). Cessão onerosa: governo arrecada r$ 69,96 bilhões com megaleilão do pré-sal. G1 Globo.