Preprints
https://doi.org/10.5194/gmd-2022-214
https://doi.org/10.5194/gmd-2022-214
Submitted as: development and technical paper
15 Sep 2022
Submitted as: development and technical paper | 15 Sep 2022
Status: this preprint is currently under review for the journal GMD.

Improving scalability of Earth System Models through coarse-grained component concurrency – a case study with the ICON v2.6.5 modeling system

Leonidas Linardakis1, Irene Stemmler2,a, Moritz Hanke3, Lennart Ramme1, Fatemeh Chegini1, Tatiana Ilyina1, and Peter Korn1 Leonidas Linardakis et al.
  • 1Max Planck Institute for Meteorology, Hamburg, Germany
  • 2wobe-systems GmbH, Kiel, Germany
  • 3Deutsches Klimarechenzentrum, Hamburg, Germany
  • apreviously at the Max Planck Institute for Meteorology

Abstract. In the era of exascale computing, machines with unprecedented computing power are available. Making efficient use of these massively parallel machines, with millions of cores, presents a new challenge. Multi-level and multi-dimensional parallelism will be needed to meet this challenge.

Coarse-grained component concurrency provides an additional parallelism dimension, that complements typically used parallelization methods such as domain-decomposition and loop level shared memory approaches. The novel aspect is that component concurrency is a function parallel technique, while these parallelization methods are data parallel techniques. This additional dimension of parallelism allows us to extend scalability beyond the limits set by the established parallelization techniques. Furthermore, concurrency allows each component to run on different hardware, and thus leveraging the usage of heterogeneous hardware configurations.

We study the characteristics of component concurrency and analyse its behaviour in a general context. These generic considerations are complemented by an analysis of a specific case, namely the coarse-grained concurrency in the multi-level parallelism context of two components of the ICON modeling system: the ICON ocean model ICON-O and the marine biogeochemistry model HAMOCC. The additional computational cost incurred by the biogeochemistry module is about three times that of the ICON-O ocean stand alone model, and traditional parallelization techniques present a scaling limit that impedes the computational performance of the combined ICON-O-HAMOCC model. Scaling experiments, with and without concurrency, show that component concurrency extends the scaling, in cases doubling the parallel efficiency. The experiments’ scaling results are in agreement with the theoretical analysis.

Leonidas Linardakis et al.

Status: open (until 10 Nov 2022)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse

Leonidas Linardakis et al.

Data sets

Dataset for: Improving scalability of Earth System Models through coarse-grained component concurrency - a case study with the ICON v2.6.5 modelling system Leonidas Linardakis https://edmond.mpdl.mpg.de/dataset.xhtml?persistentId=doi:10.17617/3.FGFQZG

Leonidas Linardakis et al.

Viewed

Total article views: 221 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
162 53 6 221 1 2
  • HTML: 162
  • PDF: 53
  • XML: 6
  • Total: 221
  • BibTeX: 1
  • EndNote: 2
Views and downloads (calculated since 15 Sep 2022)
Cumulative views and downloads (calculated since 15 Sep 2022)

Viewed (geographical distribution)

Total article views: 160 (including HTML, PDF, and XML) Thereof 160 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 28 Sep 2022
Download
Short summary
In Earth system modeling we are facing the challenge of making efficient use of very large machines, with millions of cores. To meet this challenge we will need to employ multilevel and multidimensional parallelism. Component concurrency, being a function parallel technique, offers an additional dimension to the traditional data-parallel approaches. In this paper we examine the behavior of component concurrency and identify the conditions for its optimal application.