Sustainable Computing

Prof. Tevfik Kosar

Sustainable Computing podcasts by Prof. Tevfik Kosar and his team at the University at Buffalo cover the latest trends in green computing and sustainability with implications on AI workloads, cloud computing, and beyond. Credits: This podcast series is prepared by Prof. Tevfik Kosar and his team at the University at Buffalo using Google NotebookLM. Prof. Kosar's published papers and blog posts are used as the feed for the podcast generation.

Folgen

  1. Carbon-aware Networking

    17.12.2024

    Carbon-aware Networking

    The growing adoption of cloud, edge, and distributed computing, as well as the rise in the use of AI/ML workloads, have created a significant need to measure, monitor, and reduce the carbon emissions associated with these resource-intensive tasks. One significant but often overlooked source of emissions is data transfers over wide-area networks (WANs), primarily due to the challenges in accurately measuring the carbon footprint of end-to-end network paths. This podcast discusses the need and available mechanisms to measure network carbon footprint. It explains how users can achieve significant carbon savings by shifting data transfer tasks across time and geographic regions based on local carbon intensity. Link to the related paper: https://arxiv.org/pdf/2406.09650 Credits: This podcast series is prepared by Prof. Tevfik Kosar and his team at the University at Buffalo using Google NotebookLM. The team's published papers and blog posts are used as the feed for the podcast generation.

    13 Min.
  2. Sustainable AI - Best Practices

    06.12.2024

    Sustainable AI - Best Practices

    This episode explores the significant environmental impact of AI, particularly the substantial energy consumption of large-scale AI models and data centers. Several studies are cited, quantifying the carbon footprint of AI training and highlighting the projected increase in energy usage by 2030. The podcast examines strategies for creating more sustainable AI, including choosing efficient AI frameworks (comparing PyTorch and TensorFlow), simplifying model structure, utilizing low-precision data formats, and employing energy-aware code refactoring techniques. Finally, it discusses GPU power capping as a method to reduce energy consumption during AI model training and inference. Link to the related blog post: http://magnus-notitia.blogspot.com/2024/09/the-path-to-sustainable-ai-core.html Credits: This podcast series is prepared by Prof. Tevfik Kosar and his team at the University at Buffalo using Google NotebookLM. The team's published papers and blog posts are used as the feed for the podcast generation.

    24 Min.
  • 2 Folgen

Info

Sustainable Computing podcasts by Prof. Tevfik Kosar and his team at the University at Buffalo cover the latest trends in green computing and sustainability with implications on AI workloads, cloud computing, and beyond. Credits: This podcast series is prepared by Prof. Tevfik Kosar and his team at the University at Buffalo using Google NotebookLM. Prof. Kosar's published papers and blog posts are used as the feed for the podcast generation.

Informationen

  • Erstellt von
    Prof. Tevfik Kosar
  • Jahre aktiv
    2024
  • Folgen
    2