You are here: Home » Archive » Phase 1 » Project » Description of work » JRA2/WP8 - European ESM: Performance Enhancement

JRA2/WP8 - European ESM: Performance Enhancement


This WP will undertake research into the performance aspects of configuring, deploying and running Earth System Models (ESMs). This WP covers a number of key areas relating to model performance including: researching portability and performance of key models on a range of platforms to provide improved models; to improve the links between model support initiatives and to ensure the European climate science community is better prepared to work together on the petascale machines that are imminent; to develop tools to ease the composition of new ESMs from existing model components and coupler technologies which will lower the technical hurdle for small climate research organisations.

The WP is structured into 3 complementary tasks, each of which is strongly application-led:

  • Evaluation Suite, will gather key existing and developing ESMs, as examples of ESM infrastructure, and provide an application-focussed lead to the other tasks, ensuring cross-fertilisation of effort and knowledge.
  • Portability, performance analysis and improvement, seeks to understand and improve the performance of the ESM models for current state-of-the-art computing systems and prepare for future architectures such as PRACE and DEISA. The applications work in this project will complement PRACE.
  • Flexible construction of ESMs, will undertake research into future coupling systems, seeking to provide flexibility in the construction and deployment of future, community-based, ESMs based on a clear understanding of the strengths and weaknesses of the best practice coupling technologies in Europe (OASIS) and the US (ESMF).

The results of the User Survey in JRA1 will be used to drive the application lead in this work package and help determine the specific interactions between the application holders and other partners.

See the full description in the DoW




Effort in this work package has focused on gathering information on the status of a number of current ESMs available to the project, and on making that available to all partners (Task 1). Following that, a number of porting and performance analysis experiments have been undertaken by teams including application owners and performance specialists (Task 2). Finally, work has started on looking at future coupling technologies (Task 3).


Task 1: Evaluation Suite:

In this task, the initial collection of ESMs for the project were identified and documented in deliverable D8.1, Definition of the Evaluation Suite.

Task 2: Portability, Performance Analysis and Improvement:

D8.2 describes the results of specific porting and performance-related activities undertaken to-date by partners in this WP. These activities have been undertaken, either individually or, more commonly, jointly with other partners in small teams consisting of application owners and technical specialists at HPC centres. The general aim of these activities has been to perform initial porting and performance analyses of ESMs in order to collect data and experiences which are being used to inform current and future developments of ESM models, and related software infrastructure, for existing and future HPC systems. Preparations for interaction with PRACE and DEISA have also been undertaken (in the IS-ENES HPC Task Force).


Task 3: Flexible Construction of ESMs:

Task 3 is concerned with future coupling techniques and technologies for Earth System Modelling. Activities have included work on the JULES land surface model to enable it to be more easily configured either within the Met Office’s UM (one of the ESMs in the IS-ENES test suite) or as a stand-alone model, and, potentially, to be configured more easily with other models in the future. Tests show that there is no loss of performance when configuring the UM with the new, more modular, JULES when compared with the original, ‘hard-wired’ equivalent configuration. This work is part of the effort on flexible construction of ESMs.

Some work has also been done on supporting the extraction of Metadata from code in pursuit of ensuring consistency between the code and Metadata. (This work will build on our experiences in METAFOR). As a result of this work, UNIMAN has been invited to become a developer partner of the Open FORTRAN Parser (OFP, see:




D8.1 Report on the Definition of the Evaluation Suite


D8.2 Report on the Description of the Evaluation Suite and Base-case Results
© Copyright ENES Portal 2011



IS-ENES3 has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 824084