The Computing Community Consortium (CCC) spoke with council member Michela Taufer who will be moderating a panel discussion at the 2024 Supercomputing Conference on the future sustainability of computing.
Here’s what she had to say.
How do you see the current trends in HPC influencing the future of this topic?
High-Performance Computing (HPC) and artificial Intelligence (AI) are increasingly pivotal for tackling urgent global issues, like predicting climate and health crises. However, AI’s energy demands are escalating, potentially conflicting with sustainability goals. The rapid pace of AI and HPC expansion challenges our infrastructure, especially the power grid, to keep up. These trends underscore a critical question: Can HPC and AI truly scale while being sustainable, or is this inherently an unsustainable pursuit? In the panel, I will challenge both panelists and attendees to address this balance carefully to ensure AI and HPC do not inadvertently contribute to the crises it is aiming to solve.
Are there any innovative solutions or technologies that you’re particularly excited about discussing during the panel?
We are excited to explore decentralized energy management for AI-driven systems, which could alleviate energy burdens on centralized power grids and enhance resilience. Another promising area is the integration of renewable energy sources directly with edge systems, a shift that could significantly cut down on emissions. Standardizing transparency in energy usage and data accountability is also an area with transformative potential, helping ensure that AI applications are used for meaningful, high-impact purposes rather than trivial or resource-intensive tasks.
What are the three main takeaways you hope attendees will gain from this session?
Urgency in Sustainability: HPC and AI expansion cannot continue unchecked without sustainable practices. I will challenge the panelists and the attendees on whether the community must adopt renewable energy and responsible growth strategies.
Academia and Industry Collaboration: Effective solutions require joint efforts between academia and industry to address sustainability challenges. I will challenge the panelists and the attendees on whether each strength must be leveraged to create scalable, sustainable solutions.
Balance in Purpose and Transparency: A commitment to transparency in AI’s energy and data use will ensure responsible applications, driving meaningful impacts rather than resource-intensive low-value tasks. I will challenge the panelists and the attendees on whether and how we should recognize the value of a balanced approach, ensuring AI advancements align with sustainability and ethical considerations.
Where do you see HPC’s impact on the environment in 10 years?