Typical methodological approaches for designing parallel algorithms rely on the introduction of concurrency inside the most time consuming parts of the algorithm. Nevertheless, the key architectural changes of exascale machines will force the development of innovative approaches for introducing the parallelism throughout the overall layers of the math stack for simulation in ways that cannot be completely hidden from the associated numerical solvers and software libraries, situated at the bottom layer of the math stack. In line with such computational needs, we discuss the research activity which is being carried out within the H2020 Project NASDAC (iNnovative Approaches for Scalable Data Assimilation in oCeanography) in the scientific context of Data Assimilation (DA) for Ocean Circulation Models (OCMs). Numerical mathematicians from the University of Naples Federico II and Argonne National Laboratory in Chicago, physicists from the Imperial College London and University of California Santa Cruz, engineers from the Barcelona Supercomputing Center, are investigating on the design and development of scalable parallel algorithms for Variational DA models based on domain decomposition methods, designed for an efficient use in high resolution OCMs. Innovation starts from the introduction of multiple levels of the problem decomposition which conducts to a formulation of a scalable mathematical model. The multiple levels of decomposition follow a tree configuration by grafting each level of decomposition into the previous one. This approach allows us to pass the limitation on the scalability of the straightforward introduction of parallelism inside the most time consuming parts of the algorithm. These methodologies will be tested in the Incremental Strong Contraints 4DVAR module of the Regional Ocean Modelling System (ROMS). Validations will be performed on data collected in enclosed and semi-enclosed seas (e.g. Mediterranean and Caspian sea).

New strategies and algorithms to solve Data Assimilation problems in Oceanography efficiently on next generation computing systems / Arcucci, Rossella; D'Amore, Luisa; Murli, Almerico. - (2016).

New strategies and algorithms to solve Data Assimilation problems in Oceanography efficiently on next generation computing systems

ARCUCCI, ROSSELLA;D'AMORE, LUISA;MURLI, ALMERICO
2016

Abstract

Typical methodological approaches for designing parallel algorithms rely on the introduction of concurrency inside the most time consuming parts of the algorithm. Nevertheless, the key architectural changes of exascale machines will force the development of innovative approaches for introducing the parallelism throughout the overall layers of the math stack for simulation in ways that cannot be completely hidden from the associated numerical solvers and software libraries, situated at the bottom layer of the math stack. In line with such computational needs, we discuss the research activity which is being carried out within the H2020 Project NASDAC (iNnovative Approaches for Scalable Data Assimilation in oCeanography) in the scientific context of Data Assimilation (DA) for Ocean Circulation Models (OCMs). Numerical mathematicians from the University of Naples Federico II and Argonne National Laboratory in Chicago, physicists from the Imperial College London and University of California Santa Cruz, engineers from the Barcelona Supercomputing Center, are investigating on the design and development of scalable parallel algorithms for Variational DA models based on domain decomposition methods, designed for an efficient use in high resolution OCMs. Innovation starts from the introduction of multiple levels of the problem decomposition which conducts to a formulation of a scalable mathematical model. The multiple levels of decomposition follow a tree configuration by grafting each level of decomposition into the previous one. This approach allows us to pass the limitation on the scalability of the straightforward introduction of parallelism inside the most time consuming parts of the algorithm. These methodologies will be tested in the Incremental Strong Contraints 4DVAR module of the Regional Ocean Modelling System (ROMS). Validations will be performed on data collected in enclosed and semi-enclosed seas (e.g. Mediterranean and Caspian sea).
2016
New strategies and algorithms to solve Data Assimilation problems in Oceanography efficiently on next generation computing systems / Arcucci, Rossella; D'Amore, Luisa; Murli, Almerico. - (2016).
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11588/640967
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact