Export Publication

Benedicto, C., Rodrigues, I. L., Tygel, M., Breternitz, M. & Borin, E. (2017). Harvesting the computational power of heterogeneous clusters to accelerate seismic processing   . In Global Meeting Expanded Abstracts. Rio de Janeiro: Sociedade Brasileira de Geofísica.

C. benedito et al.,  "Harvesting the computational power of heterogeneous clusters to accelerate seismic processing   ", in Global Meeting Expanded Abstracts, Rio de Janeiro, Sociedade Brasileira de Geofísica, 2017

@inproceedings{benedito2017_1764949377122,
	author = "Benedicto, C. and Rodrigues, I. L. and Tygel, M. and Breternitz, M. and Borin, E.",
	title = "Harvesting the computational power of heterogeneous clusters to accelerate seismic processing   ",
	booktitle = "Global Meeting Expanded Abstracts",
	year = "2017",
	editor = "",
	volume = "",
	number = "",
	series = "",
	doi = "10.1190/sbgf2017-070",
	publisher = "Sociedade Brasileira de Geofísica",
	address = "Rio de Janeiro",
	organization = "",
	url = "https://library.seg.org/doi/10.1190/sbgf2017-070"
}

TY  - CPAPER
TI  - Harvesting the computational power of heterogeneous clusters to accelerate seismic processing   
T2  - Global Meeting Expanded Abstracts
AU  - Benedicto, C.
AU  - Rodrigues, I. L.
AU  - Tygel, M.
AU  - Breternitz, M.
AU  - Borin, E.
PY  - 2017
SN  - 2159-6832
DO  - 10.1190/sbgf2017-070
CY  - Rio de Janeiro
UR  - https://library.seg.org/doi/10.1190/sbgf2017-070
AB  - Cluster environments are crucial to modern geophysics. Major processing companies make use of one or more computational environments, whether they be in-house clusters or third-party public clouds, to guarantee the efficient execution of their processing flows. But the diversification of such environments created a demand for software tools that are able to scale with efficiency in these ever-increasing ecosystems. Aside from efficiency requirements, these tools must also be able to handle and recover automatically from the faults that arise from these new and complex ecosystems. In this paper, we discuss how we leverage the Scalable Partially Idempotent Tasks System (SPITS) programming model and the PY-PITS runtime system to efficiently harvest the computing power of heterogeneous systems in order to solve geophysics problems. We also present an experiment in which we combine the computational resources from several clusters and workstations simultaneously to perform the regularization of seismic data and demonstrate the scalability and robustness of the system.


Read More: https://library.seg.org/doi/abs/10.1190/sbgf2017-070

ER  -