LATIN AMERICAN HIGH PERFORMANCE COMPUTING CONFERENCE

Energy Efficiency

2nd Energy Efficiency and Sustainability  in AI, HPC, and Quantum Computing

Chairs Workshop

  • Arthur Francisco Lorenzon – Federal University of Rio Grande do Sul, Brazil
  • João Carlos Garcia da Cunha Barbosa – IT4Innovation @ VSB-TUO
  • Maximilian Höb – Leibniz Supercomputing Centre, Munich, Germany

Description of the workshop

High-performance computing (HPC), artificial intelligence (AI), and quantum computing (QC) are essential tools for accelerating scientific discovery, engineering innovation, and decision-making across disciplines. From climate modeling and genomics over large-scale simulations and AI-driven analytics to quantum-enhanced simulations, these technologies continue to push the boundaries of what is computationally possible. However, the rapid growth of compute-intensive workloads has led to a dramatic rise in energy consumption, raising serious concerns around operational costs, infrastructure demands, and environmental sustainability. The rapid expansion of large-scale AI training and inference infrastructures, including GPU-intensive AI factories and sovereign AI initiatives, has further amplified the urgency of sustainable and energy-aware computing strategies.

These challenges are particularly pressing in Latin America, where the HPC and AI ecosystem is advancing rapidly, yet must operate within unique regional constraints—including higher electricity costs, limited cooling infrastructure, and sustainability priorities driven by environmental, social, and economic factors. As such, energy efficiency is not just a technical goal, but a strategic imperative for ensuring the long-term viability and accessibility of high-performance digital technologies in the region.

The “2nd Energy Efficiency and Sustainability in AI, HPC, and Quantum Computing” workshop at CARLA 2026 seeks to address this critical need by providing a dedicated forum for sharing knowledge, strategies, and best practices related to sustainable computing. It aims to catalyze innovation through dialogue among system architects, application developers, hardware and software researchers, datacenter operators, and policy makers.

The workshop will explore both foundational and applied aspects of energy-aware computing, including novel algorithms and architectures, software-hardware co-optimization, real-world monitoring and tuning approaches, and institutional or policy-level solutions. Special emphasis will be placed on case studies, experiences, and collaborative efforts originating in or applicable to the Latin American context.

By fostering a regional and international community focused on sustainability in HPC and AI, the workshop aspires to build bridges between emerging research and practical implementation—ensuring that performance, energy efficiency, and social responsibility evolve hand in hand.

Topics of Interest

We invite submissions on (but not limited to) the following themes:

  • Energy-efficient algorithms, AI models, and foundation model training/inference
  • Scheduling, orchestration, and workload management for energy and power efficiency
  • Hardware-software co-design and optimization for sustainable computing
  • Low-power, accelerator-based, and heterogeneous computing architectures
  • Monitoring, telemetry, modeling, and benchmarking of energy consumption in HPC and AI systems
  • Carbon-aware computing and energy-aware resource management
  • Sustainable GPU computing and large-scale AI infrastructures
  • Renewable energy integration and sustainable operation of data centers
  • Energy-aware data center design, cooling technologies, and waste heat reuse
  • AI-driven optimization of infrastructure energy consumption and operations
  • Case studies and operational experiences from HPC, AI, and quantum computing infrastructures
  • Sustainability challenges and operational experiences from AI factories and exascale systems
  • Policy, governance, and educational frameworks for sustainable digital infrastructures
  • Energy-efficient quantum computing systems and hybrid HPC-QC workflows

Submission Guidelines

  • Guidelines: https://revistas.usfq.edu.ec/index.php/avances/about/submissions  (Change language at end of webpage)
  • Submission Types: Full research papers, experience reports, or position papers
  • Page Limit: Minimum 6 / Maximum 15 pages (including references)
  • Language: English
  • Format: Microsoft Word or LaTeX using the ACI template
  • ACI Submission Info: https://revistas.usfq.edu.ec/index.php/avances/about/submissions
  • Review Process: Double-blind peer review
  • Submission Portal: https://meteor.springer.com/carla2026

 * Select Track: “Workshop: Energy Efficiency”

All accepted papers will be presented during the workshop and published in the CARLA 2026 Workshop Proceedings (Avances en Ciencias e Ingeniería (ACI)).

Policy on Originality and Generative AI Use

Submissions must be original, not under review elsewhere, and must comply with CARLA and Springer policies. The use of generative AI tools (e.g., ChatGPT, GitHub Copilot) is permitted only for improving language clarity—not for producing scientific content or experimental results. Any AI assistance must be disclosed in the final manuscript.
Plagiarism and self-plagiarism will result in rejection.

Workshop Format – Half-Day

  • Keynote Talk (Latin America and international perspectives)
  • Paper presentations (15 minutes each, with Q&A)
  • Concluding panel discussion on future directions in sustainable HPC, AI and QC in the region

Workshop Keynote Speaker

Antigoni Georgiadou – ORNL
Title: Treating Energy and Uncertainty as First-Class Metrics in AI and HPC

Short Bio:
Antigoni Georgiadou is an applied mathematician in the Science Engagement Section of the Algorithms & Performance Analysis (APA) Group at the Oak Ridge Leadership Computing Facility (OLCF). She supports major computational science campaigns by partnering with teams to improve performance, scalability, and scientific throughput on OLCF supercomputers and emerging GPU-accelerated systems. In this role, she serves as liaison for two INCITE 2027 proposal efforts: “APEX: AI- and Physics-Driven EXperimental Workflow for Healthcare Applications” (PI: Peter Coveney) and “Exascale Cardiovascular Digital Twins Across Temporal and Cellular Scales” (PI: Amanda Randles), helping align application goals with facility capabilities and readiness. She also leads OLCF’s Uncertainty Quantification Working Group and contributes to uncertainty visualization for large-scale simulations, including cardiovascular flow studies of aortic stenosis. Dr. Georgiadou earned her Ph.D. in Mathematics from Florida State University and previously held roles at the European Space Agency and Fermilab applying AI, Gaussian processes, and machine learning to scientific workflows.

Abstract: Energy efficiency has become a critical constraint in high-performance computing (HPC) as systems scale toward larger node counts. In modern HPC platforms, energy consumption is influenced by complex interactions among hardware and software parameters, including operating frequencies, concurrency levels, memory behavior, and runtime policies. These interactions are further affected by execution-time variability arising from hardware heterogeneity, resource contention, and system noise. As a result, energy measurements often exhibit significant fluctuations that are not captured by deterministic models or single-run experiments, limiting the reliability of traditional energy-aware optimization approaches. This presentation will cover the usage of uncertainty quantification (UQ) as a systematic method for analyzing energy efficiency in scientific HPC applications. Instead of relying on average-case behavior, we explicitly model the variability in execution time and energy consumption as functions of system configuration parameters. We present a tool-based methodology that combines controlled experimental measurements with variance-based sensitivity analysis, surrogate modeling, and uncertainty propagation. This approach enables the identification of configuration parameters that most strongly affect energy consumption and performance variability, as well as the assessment of energy-performance trade-offs under uncertainty. We will dive in application-agnostic tools and applicable across a wide range of scientific HPC workloads and architectures.