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:20260605T154541Z
LOCATION:Bldg. 8 - Room B 101
DTSTART;TZID=Europe/Stockholm:20260701T140000
DTEND;TZID=Europe/Stockholm:20260701T160000
UID:submissions.pasc-conference.org_PASC26_sess164@linklings.com
SUMMARY:MS5G - Computational Storage for Scientific Computing
DESCRIPTION:Organizer(s): Jakob Luettgau (INRIA), Kira Duwe (CERN), and Mi
 chael Kuhn (Otto-von-Guericke-Universität Magdeburg)\n\nThe exponential gr
 owth of data handled by HPC systems is driven by growing simulation fideli
 ty and an explosion of affordable, high-resolution sensing devices. The ri
 se of machine learning applications and other large-scale data analysis ta
 sks imposes workload patterns on existing storage solutions that often lea
 d to large, avoidable data copies and transfers, as well as contention nea
 r storage devices and in the network. This creates challenges for the repr
 oducibility of science and thus limits trust in science. Computational sto
 rage is a promising solution to lower barriers to public data access and r
 educe the time and cost for many workloads that require search across or a
 ggregation of large volumes of data. The session will feature four speaker
 s from academia, government labs, and industry to give their perspectives 
 on the challenges and opportunities of computational storage for scientifi
 c computing and how the technology may help to increase trust in science d
 espite the data deluge.\n\nStorage Systems and Programming Models for Comp
 utational Storage Landscapes\n\nStorage and I/O are increasingly important
  due to exponentially growing data volumes and more dynamic and diverse I/
 O loads. Traditionally, there has been a clear separation of computation a
 nd storage with the bulk of storage and I/O requirements being handled by 
 centralized parallel file systems. A...\n\n\nMichael Kuhn (Otto von Gueric
 ke University Magdeburg)\n---------------------\nOpportunities for Computa
 tional Storage in Scientific Computing\n\nThis talk will give an overview 
 of computational storage as a promising technology both to lower barriers 
 to public data access while also reducing the time and cost for many exist
 ing workloads that require search across or aggregation of large volumes o
 f data. The talk will give an overview of diff...\n\n\nJakob Luettgau (INR
 IA)\n---------------------\nPower Efficiency as seen from the Storage Pers
 pective\n\nWith computational resources increasingly commoditized, perform
 ance optimization now prioritizes power efficiency. Effective energy optim
 ization, however, necessitates preliminary empirical observations. If powe
 r monitoring is well established for compute nodes, storage servers tend t
 o be overlooked...\n\n\nJean-Thomas Acquaviva (DDN)\n---------------------
 \nIn-Situ Data Analysis Meets Computational Storage Devices\n\nSimulations
  from domains such as climate science produce increasing amounts of data a
 s we approach Exascale. The current workflow for generating knowledge from
  the data relies on post-mortem analysis, which requires storing the raw d
 ata. The amount of storage space needed to facilitate this analysis...\n\n
 \nNiclas Schroeter (DKRZ, Otto-von-Guericke-Universitat Magdeburg)\n\nDoma
 in: Computational Methods and Applied Mathematics\n\nSession Chairs: Jakob
  Luettgau (INRIA) and Michael Kuhn (Otto von Guericke University Magdeburg
 )
END:VEVENT
END:VCALENDAR