Preprints
https://doi.org/10.5194/gmd-2024-40
https://doi.org/10.5194/gmd-2024-40
Submitted as: development and technical paper
 | 
09 Apr 2024
Submitted as: development and technical paper |  | 09 Apr 2024
Status: this preprint is currently under review for the journal GMD.

Refactoring the EVP solver for improved performance – a case study based on CICE v6.5

Till Andreas Soya Rasmussen, Jacob Poulsen, Mads Hvid Ribergaard, Ruchira Sasanka, Anthony P. Craig, Elizabeth Clare Hunke, and Stefan Rethmeier

Abstract. This study focuses on the performance of CICE and its Elastic-Viscous-Plastic (EVP) dynamical solver. The study has been conducted in two steps. First, the standard EVP solver has been extracted from CICE for experiments with refactored versions of it. Secondly, one refactored version was integrated and tested as part of the full model. Two dominant bottlenecks were revealed. The first is the number of MPI and OpenMP synchronization points required for halo exchanges during each time-step combined with the irregular domain of active sea ice points. The second is the lack of Single Instruction Multiple Data (SIMD) code generation.

The study refactors the standard EVP solver based on two generic patterns. The first pattern exposes how general finite-differences on masked multi-dimensional arrays can be expressed in order to produce significantly better code generation. The primary change is that the memory access pattern is changed from random access to direct access. The second pattern exposes an alternative approach to handle static grid properties.

The measured single core improvement is increased by more than a factor of five compared to the standard implementation. The refactored implementation strong scales on the Intel® Xeon® Scalable Processor Series node until the available bandwidth of the node is used. For the Intel® Xeon® CPU Max Series Series there is sufficient bandwidth to allow the strong scaling to continue for all the cores on the node resulting in a single node improvement factor of 35 over the standard implementation.

This study also show improved performance on GPU processors.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Till Andreas Soya Rasmussen, Jacob Poulsen, Mads Hvid Ribergaard, Ruchira Sasanka, Anthony P. Craig, Elizabeth Clare Hunke, and Stefan Rethmeier

Status: open (until 04 Jun 2024)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CEC1: 'Comment on gmd-2024-40 - No compliance with GMD's policy', Juan Antonio Añel, 11 May 2024 reply
    • AC1: 'Reply on CEC1', Till Rasmussen, 11 May 2024 reply
      • CEC2: 'Reply on AC1', Juan Antonio Añel, 12 May 2024 reply
        • AC2: 'Reply on CEC2', Till Rasmussen, 12 May 2024 reply
          • CEC3: 'Reply on AC2', Juan Antonio Añel, 13 May 2024 reply
            • AC3: 'Reply on CEC3', Till Rasmussen, 23 May 2024 reply
              • CEC4: 'Reply on AC3', Juan Antonio Añel, 23 May 2024 reply
  • RC1: 'Comment on gmd-2024-40', Anonymous Referee #1, 24 May 2024 reply
Till Andreas Soya Rasmussen, Jacob Poulsen, Mads Hvid Ribergaard, Ruchira Sasanka, Anthony P. Craig, Elizabeth Clare Hunke, and Stefan Rethmeier
Till Andreas Soya Rasmussen, Jacob Poulsen, Mads Hvid Ribergaard, Ruchira Sasanka, Anthony P. Craig, Elizabeth Clare Hunke, and Stefan Rethmeier

Viewed

Total article views: 290 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
220 44 26 290 16 18
  • HTML: 220
  • PDF: 44
  • XML: 26
  • Total: 290
  • BibTeX: 16
  • EndNote: 18
Views and downloads (calculated since 09 Apr 2024)
Cumulative views and downloads (calculated since 09 Apr 2024)

Viewed (geographical distribution)

Total article views: 283 (including HTML, PDF, and XML) Thereof 283 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 25 May 2024
Download
Short summary
Earth system models (ESM) today strive towards better quality based on improved resolution and improved physics. One of the limiting factors is the super computers at hand and how to utilize these. This study focus on a refactorization of one part of a sea ice model (CICE), namely the dynamics. The study shows that the performance can be significantly reduced, which means that one can either run the same simulations much cheaper or advance the system according to what is needed.