Scientific computing's energy footprint has become a critical challenge. What once was primarily about computational cost is now equally about electricity consumption and environmental impact. We need to rethink how we approach modeling, simulation, programming, and hardware—not in isolation, but as interconnected parts of a sustainable computing ecosystem.

This workshop brings together diverse experts to tackle sustainability from multiple angles: algorithm designers, programming language architects, software developers, hardware engineers, cloud computing specialists, and machine learning practitioners. By connecting these perspectives, we aim to identify not just local optimizations, but systemic solutions that reduce energy waste across the entire computational pipeline.

Workshop Format

• Participation: Both onsite and online attendance available
• Capacity: 45 onsite seats, unlimited online participation

Special Feature: Flash Talk Bazaar

We're creating a unique opportunity for early-career researchers (PhD students, postdocs) through our Flash Talk Bazaar session. Young developers can present their sustainable computing innovations in 5-minute pitches to an audience of distinguished experts. Think of it as a marketplace where emerging talent meets established leaders—perfect for sparking collaborations, finding mentors, or catching the attention of potential employers.

Scientific Context

Sustainability challenges appear at every level:

• Models & Algorithms: Balancing accuracy with efficiency
• Programming Languages: How do Fortran, C++, and Python compare for energy consumption?
• Software Development: Optimizing memory, data movement, and runtime
• Hardware: Leveraging GPUs, CPUs, and HPC systems efficiently
• Cloud Computing: Managing energy and costs on Azure, AWS, Google Cloud
• Machine Learning: Using ML to enhance efficiency while managing its own energy demands

Each domain faces unique challenges, but solutions often require cross-domain thinking. A hardware innovation might enable new algorithmic approaches; an efficient algorithm might change how we use programming languages.

Workshop Objectives

We're focused on practical outcomes:

1. Exchange optimization strategies across programming languages and software tools
2. Align software with hardware capabilities for maximum efficiency
3. Develop energy-efficient algorithms for computational chemistry and molecular simulations
4. Optimize GPU utilization and hybrid computing approaches
5. Establish cloud computing best practices for scientific workloads
6. Integrate machine learning thoughtfully, managing both benefits and costs
7. Create actionable best practices that researchers can implement immediately
8. Build lasting collaborations between industry and academia
9. Support early-career researchers through networking and mentorship
10. Form working groups to continue progress beyond the workshop

Expected Outcomes

• Practical guidelines for energy-efficient software and algorithm design
• Framework for assessing and comparing power efficiency
• Best practices for GPU-centric computing and cloud platforms
• Strategies for integrating ML while managing energy footprint
• Policy recommendations for HPC centers and cloud providers
• Stronger industry-academia partnerships
• Career opportunities for early-career researchers
• Ongoing working groups for continued collaboration

Who Should Attend

• Programming language developers (C++, Fortran, Python)
• Scientific software creators (open-source and commercial)
• Algorithm designers and computational scientists
• Hardware engineers and HPC specialists
• Cloud computing experts
• Machine learning practitioners
• Early-career researchers developing sustainable computing tools
• Anyone passionate about reducing the environmental impact of scientific computing

Join us in Lugano to help build a more sustainable future for computational science—whether in person or online.