CES Units

• **Xavier Yepes** (Unit Leader)
• Iria Ayan
• Marcel Cases
• Roc Salvador
• Sergi Palomas
• Simone Vacondio
• Eric Ferrer (attending meetings)

Contribute to the development and performance improvement of the EC-Earth model and its associated tools.

• Fixing known issues that are hindering prouctions runs
• Improving coupled load-balance
• Profiling and identifying areas for optimization
• Improving I/O
• Refactoring code
• Efficient deployment
• Code development of new features/components
• Develop tools such as Auto-LB and APP

• EC-Earth
• Auto-LB
• APP

Efficient executions of the EC-Earth model in terms of energy and computatinal resources.

• Antoni Portero (Unit Leader)
• Aitor Sanchez
• Carlos Peña

CHIP focuses on bridging application workloads and emerging architectures by driving hardware/software co-design and cross-stack analysis to optimize performance, energy efficiency, and the evolution of next-generation RISC-V and heterogeneous systems.

• Architecture-driven HPC workload profiling.
• Future-generation RISC-V architectures.
• Earth science codes (EC) co-design & optimisation.
• Cross-stack performance analysis (application ↔ architecture).
• Hardware/software co-design.
• Performance & energy efficiency (performance/W).
• Architecture modelling & evaluation.
• Workload-driven architecture design.

• DARE
• LOCA

• Development of domain-specific mini-apps and computational kernels for architecture evaluation and hardware/software co-design.
• Improved performance and energy efficiency (performance/W) through architecture-aware optimisation strategies.
• Workload-informed insights to guide the evolution of future RISC-V and heterogeneous computing architectures.
• Strengthened collaboration across application domains, computer architecture, and system software communities.
• Validation of methodologies for hardware/software co-design, including profiling, modelling, and architectural feedback loops.
• Scientific dissemination and broader impact:
  1. Publications in peer-reviewed journals and conferences.
  2. Workshops and collaborative research activities.
  3. Contributions to open, reproducible, and reusable research practices.

• Carlos Peña (Unit Leader)

Improve CPU efficiency, scalability, and profiling practices of Earth system models and components, across projects through shared expertise.

• Standardise profiling methods, metrics, and reporting across projects
• Support teams in identifying bottlenecks and improving CPU performance
• Optimise MPI/OpenMP configurations, load balance, and communication efficiency
• Monitor and investigate performance regressions
• Share best practices, training, and optimisation knowledge
• Evaluate new CPU technologies and future performance opportunities

• Destination Earth (DestinE) – Climate Digital Twins
• TerraDT – Digital Twin for Cryosphere and Land Surface Systems

• Future Profiling trainings, and resources.
• Future profilings (TerraDT)
• Publications derived from the profiling and optimisations resultant of the work

• ** Gilbert Montané** (Unit Leader)
• Alejandro García
• Breixo Soliño
• Eric Ferrer
• Iria Ayan
• Matias F. Rebolledo
• Marcel Cases
• Roc Salvador
• Sergi Palomas
• Simone Vacondio
• Supriyo Ghosh

Contribute to the development, maintenance, and improvement of the EC-Earth model and its associated tools.

• Collaborate in the development of new model features in close collaboration with scientific groups and other EC-Earth consortium partners
• Study and improve the performance of the model
• Develop and maintain the Auto-EC-Earth tool
• Support users on the execution of the model and postprocessing tools

• EC-Earth
• Auto-EC-Earth
• ece2cmor
• Auto-LB
• APP

Auto-EC-Earth workflow including the execution of the model and the postprocessing of the outputs, new features and improvements in the model.

• ** Alexey Medvedev** (Unit Leader)
• Aitor Sanchez
• Andrew Forembski
• Christian Guzman
• Okke L. Van Eck
• Rommel A. Quintanilla

Create and profile modeling and compute intensive GPU code in C/C++ and Fortran.

• Port existing model code to GPUs
• Automate GPU code profiling
• Research on best automation methods for GPU code porting

• Destination Earth – Climate Digital Twins
• European Digital Twin Ocean
• MONARCH
• NEMO v.4
• NEMO v.5
• CAMP

The modeling software components that are able to effectively use the GPU resources of EuroHPC ecosystem.

• ** Hannah Ross** (Unit Leader)
• Alba Martinez
• Alejandro García
• Andrew Forembski
• Christian Guzman
• Joan Bernat

Develop and improve the performance of Monarch.

• Fixing known issues that are hindering prouctions runs
• Porting to GPUs
• Profiling and identifying areas for optimization
• Improving IO
• Refactoring code
  

• Monarch
• auto-monarch
• CAMP

• Improved performance of Monarch
• Future publications
  

• ** Kai Keller** (Unit Leader)
• Carlos Romero
• Marta Alerany
• Nicholas T. Abel

Develop and implement features and performance improvements for earth system model components, and develop and apply model diagnostics.

• Develop model integration testing tools
• Implement scientific features
• Implement technical features
• Implement performance improvements
• Develop and apply model diagnostics

• Destination Earth
• TerraDT
• HANAMI
• PREDDYCT
• pyhanami diagnostics suite
• HANAMI workflow

• Novel model diagnostics
• Scientific publications
• Improved model physics and realism
• Better model performance
• International collaborations

• ** Victoria Agudetse** (Unit Leader)
• Ariadna Batalla
• Theertha Kariyathan

Develop R packages and software tools for the analysis and visualization of climate data.

See more information here.

• Maintain current in-house R packages
• Coordinate with other groups to integrate and organize common R code according to research and project needs.
• Contribute to the operationalization of climate services for department projects
• Provide user support to researchers working with the R Tools
  

• RstartR
• multiApply
• easyNCDF
• s2dv
• ClimProjDiags
• CSTools
• CSIndicators
• esviz
• SUNSET

• startR, multiApply, easyNCDF, s2dv, ClimProjDiags, CStools and CSIndicators are published on the Comprehensive R Archive Network (CRAN).
• SUNSET poster presented at conferences (CERISE General Assembly, EMS2024, XIII Asociación Española de Climatología)
• Technical papers published for some of the R Packages (CSTools, CSIndicators)
• Co-authorship in papers by other groups

• ** Elliott Rose** (Unit Leader)
• Alba Vilanova
• Luiggi G. Tenorio
• Rebeca Nunes
• Santiago Ramirez

Develop websites and web applications for the various projects within the department.

• Build and maintain web applications, websites, and manage relevant infrastructure

• INES-CORE
• INES-BDRC
• INES-CALIOPE
• INES-EPIOUTLOOK
• INES-RESPIRE
• CLIMVAL
• Contacts-app
• ESS Volto Components
• Autosubmit-gui
• Autosubmit-api
• Storage-monitoring dashboard
• Respire Website
• Caliope-Dataviz
• TEEDY DB

• Daily Dust Products
• CALIOPE
• CLIMVAL
• Autosubmit GUI
• Autosubmit API
• Storage Monitor Dashboard
• CALIOPE
• User Video Guide