BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260421T090512Z
LOCATION:Bldg. 6 - Room 102
DTSTART;TZID=Europe/Stockholm:20260629T133000
DTEND;TZID=Europe/Stockholm:20260629T153000
UID:submissions.pasc-conference.org_PASC26_sess106@linklings.com
SUMMARY:MS1D - 2nd Sustainable Computing for AI and Data-Intensive Infrast
 ructures
DESCRIPTION:Artificial Intelligence (AI) has become a primary driver of co
 mpute demand and energy consumption, with workloads such as large language
  models and AI-powered computational science applications requiring sustai
 ned, infrastructure-scale resources. At the same time, modern data centers
  operate under increasingly tight power, carbon, and power grid constraint
 s, particularly as they integrate variable renewable energy sources. Toget
 her, these trends highlight the growing tension between performance scalab
 ility, energy availability, and environmental impact. Addressing these cha
 llenges requires integrated efforts from component-level optimization to c
 o-designing computing infrastructures, workloads, and energy systems. This
  minisymposium examines how sustainability can be embedded by design acros
 s AI and data-intensive computing infrastructures. Topics include (1) data
 center-power grid co-design to enable load flexibility under renewable int
 egration, (2) infrastructure-level optimization of AI and AI-powered workl
 oads for improved energy efficiency, (3) hardware-software co-design appro
 aches that explore accelerator design trade-offs for AI and scientific com
 puting, and (4) methods, metrics, and lifecycle assessment tools. The sess
 ion also highlights the economic and environmental tensions between cost-o
 ptimal and emissions-optimal system configurations. By bringing together r
 esearchers and practitioners working on architectures, systems, and infras
 tructure design, this session aims to foster a shared understanding of the
  design principles needed to support sustainable growth in AI and data-int
 ensive computing.\n\nHow Design Choices Shape Sustainable Computing: A Har
 dware Accelerator Perspective\n\nData centers are experiencing rapidly ris
 ing energy consumption and associated greenhouse-gas emissions. Today, GPU
 s are the dominant accelerator in many data centers, yet they are not nece
 ssarily the most efficient choice for every application domain because the
 y remain relatively general-purpose. ...\n\n\nRuben Rodriguez Alvarez and 
 Denisa-Andreea Constantinescu (EPFL)\n---------------------\nModelling the
  Environmental Footprint of Future AI and HPC Datacenters: Enabling Hardwa
 re-Software Codesign Including Life Cycle Assessment\n\nArtificial intelli
 gence (AI) has emerged as a pivotal force shaping the future of scientific
  discovery and technological innovation in day-to-day life. It has become 
 a key enabler for future scientific research and development of new applic
 ations. As data centers dedicated to these workloads experie...\n\n\nNitis
 h Satya Murthy, Vincent Schellekens, and Timon Evenblij (IMEC)\n----------
 -----------\nTowards Green AI: Reducing the Environmental Footprint of LLM
  Workloads\n\nContemporary large-scale computing systems are becoming incr
 easingly heterogeneous, complex, and decentralized, driven by growing digi
 tization demands. This transformation is most evident in the rapid rise of
  Generative AI applications, which have fueled the expansion of hyperscale
  data centers, now...\n\n\nShashikant Ilager (University of Amsterdam) and
  Ivona Brandic (TU Wien)\n---------------------\nMiddlebox: Resolving the 
 Datacenter-Grid tension and Decarbonizing Both!\n\nDatacenter’s (DCs) cons
 tant power requirement is difficult to balance with variable solar and win
 d generation.  DC load flexibility is the key to solving the problem, but 
 it conflicts stable capacity requirements.\n\nWe propose ``power Middlebox
 '', a new system architecture, as a bridge. Middleb...\n\n\nAndrew Chien (
 Univ of Chicago, Argonne National Laboratory)\n\nDomain: Engineering, Comp
 utational Methods and Applied Mathematics\n\nSession Chairs: Denisa-Andree
 a Constantinescu (EPFL), Florina Ciorba (University of Basel), and Can Han
 kendi (Boston University)
END:VEVENT
END:VCALENDAR