ESiWACE3: The Centre of Excellence in Simulation of Weather and Climate in Europe
Wednesday, June 24, 2026 3:45 PM to 5:15 PM · 1 hr. 30 min. (Europe/Berlin)
Foyer D-G - 2nd Floor
Project Poster
Community EngagementEarth, Climate and Weather ModelingEducation and TrainingEnergy Efficiency and SustainabilityPerformance Tools and Simulators
Information
Poster is on display.
The ESiWACE3 Centre of Excellence (CoE) represents the third phase of Europe’s initiative to advance high-resolution Earth system modeling. Building on previous phases that demonstrated the ability to run weather and climate models at ~1 km resolution on EuroHPC supercomputers, ESiWACE3 focuses on enabling the community to achieve exascale readiness while fostering knowledge transfer across Europe’s Earth system modeling centers. Its objectives are threefold: (i) transfer and establish scalable simulation techniques for weather and climate, (ii) address common technology gaps and provide collaborative toolboxes for high-resolution modeling, and (iii) serve as a sustainable hub for training, communication, and dissemination of HPC best practices. By combining expertise from multiple modeling groups, providers and compute centers, ESiWACE3 provides targeted HPC support, fosters synergies across local efforts, and educates the next generation of researchers.
A major contribution of ESiWACE3 is the High Performance Climate and Weather (HPCW) benchmark suite, a domain-specific benchmark (DSB) designed to evaluate HPC system performance for weather and climate applications. Unlike generic synthetic benchmarks, which mainly measure peak floating-point performance, HPCW captures the complex computational and memory access patterns of Earth system models, providing meaningful insights into real-world performance. Developed initially under the ESCAPE before being handed over to ESiWACE projects, HPCW continues to evolve through european collaborations with EPI-SGA2, and even further with GANANA (India), HANAMI (Japan) and EPI-SGA2 (European Processor Initiative - Specific Grant Agreement, phase 2). The suite comprises widely used European weather and climate models and simplified “dwarfs” that retain key computational characteristics. HPCW is portable, reproducible, self-contained, and integrates with profiling tools, offering modelers and technology providers a common framework for co-design, performance optimization, and benchmarking on emerging HPC architectures. Its focus on weather and climate ensures that scientific requirements directly inform HPC innovation.
ESiWACE3 also demonstrates its impact through tailored HPC services: collaborative projects where the experts from the CoE spend 0.5FTE to advance community codes towards the exascale. Two success stories illustrate the CoE’s contributions: DALES and OGSTM. In DALES, physics modules and atmospheric dynamics were ported to GPUs using OpenACC, accelerating simulations with minimal disruption to the Fortran90 codebase. In OGSTM, GPU support was implemented using the Nvidia Fortran compiler and OpenACC directives, while a new Intel ifx toolchain provided CPU reference runs. These deployments exemplify ESiWACE3’s ability to deliver practical HPC solutions, enabling users to exploit modern architectures efficiently and reproducibly.
In summary, ESiWACE3 provides a robust framework for high-resolution weather and climate modeling on exascale systems. Through the HPCW benchmark suite and dedicated HPC services, the CoE bridges the gap between scientific ambition and computational capability, fostering collaboration across Europe and internationally while preparing the community for next-generation supercomputing challenges. By combining expertise, benchmarking, and hands-on services, ESiWACE3 ensures that the weather and climate modeling community is ready to fully leverage future HPC systems for scientific discovery.
Contributors:
The ESiWACE3 Centre of Excellence (CoE) represents the third phase of Europe’s initiative to advance high-resolution Earth system modeling. Building on previous phases that demonstrated the ability to run weather and climate models at ~1 km resolution on EuroHPC supercomputers, ESiWACE3 focuses on enabling the community to achieve exascale readiness while fostering knowledge transfer across Europe’s Earth system modeling centers. Its objectives are threefold: (i) transfer and establish scalable simulation techniques for weather and climate, (ii) address common technology gaps and provide collaborative toolboxes for high-resolution modeling, and (iii) serve as a sustainable hub for training, communication, and dissemination of HPC best practices. By combining expertise from multiple modeling groups, providers and compute centers, ESiWACE3 provides targeted HPC support, fosters synergies across local efforts, and educates the next generation of researchers.
A major contribution of ESiWACE3 is the High Performance Climate and Weather (HPCW) benchmark suite, a domain-specific benchmark (DSB) designed to evaluate HPC system performance for weather and climate applications. Unlike generic synthetic benchmarks, which mainly measure peak floating-point performance, HPCW captures the complex computational and memory access patterns of Earth system models, providing meaningful insights into real-world performance. Developed initially under the ESCAPE before being handed over to ESiWACE projects, HPCW continues to evolve through european collaborations with EPI-SGA2, and even further with GANANA (India), HANAMI (Japan) and EPI-SGA2 (European Processor Initiative - Specific Grant Agreement, phase 2). The suite comprises widely used European weather and climate models and simplified “dwarfs” that retain key computational characteristics. HPCW is portable, reproducible, self-contained, and integrates with profiling tools, offering modelers and technology providers a common framework for co-design, performance optimization, and benchmarking on emerging HPC architectures. Its focus on weather and climate ensures that scientific requirements directly inform HPC innovation.
ESiWACE3 also demonstrates its impact through tailored HPC services: collaborative projects where the experts from the CoE spend 0.5FTE to advance community codes towards the exascale. Two success stories illustrate the CoE’s contributions: DALES and OGSTM. In DALES, physics modules and atmospheric dynamics were ported to GPUs using OpenACC, accelerating simulations with minimal disruption to the Fortran90 codebase. In OGSTM, GPU support was implemented using the Nvidia Fortran compiler and OpenACC directives, while a new Intel ifx toolchain provided CPU reference runs. These deployments exemplify ESiWACE3’s ability to deliver practical HPC solutions, enabling users to exploit modern architectures efficiently and reproducibly.
In summary, ESiWACE3 provides a robust framework for high-resolution weather and climate modeling on exascale systems. Through the HPCW benchmark suite and dedicated HPC services, the CoE bridges the gap between scientific ambition and computational capability, fostering collaboration across Europe and internationally while preparing the community for next-generation supercomputing challenges. By combining expertise, benchmarking, and hands-on services, ESiWACE3 ensures that the weather and climate modeling community is ready to fully leverage future HPC systems for scientific discovery.
Contributors:
Format
on-demandon-site
