BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T000609Z
LOCATION:704-706
DTSTART;TZID=America/Denver:20231112T090000
DTEND;TZID=America/Denver:20231112T123000
UID:submissions.supercomputing.org_SC23_sess424@linklings.com
SUMMARY:13th International Workshop on Runtime and Operating Systems for S
 upercomputers (ROSS)
DESCRIPTION:RDARuntime:  An OS for AI Accelerators\n\nToday's supercompute
 rs are more heterogeneous than ever before. As the share of AI workloads i
 n data centers continues to grow, the share of GPUs and AI-specific hardwa
 re grows with it. AI accelerators are different from traditional hardware,
  affecting all aspects of system design, from data-center ...\n\n\nBenjami
 n Glick, Arjun Sabnis, Renate Kempf, Arnav Goel, Aarti Lalwani, Guoyao Fen
 g, and Kiran Ranganath (SambaNova Systems Inc)\n---------------------\nCAR
 AT KOP:  Toward Protecting the Core HPC Kernel from Linux Kernel Modules\n
 \nExtending Linux through kernel modules offers immense potential benefits
  and capabilities for HPC.  Deployment is also more likely since Linux is 
 typically the only supported vendor OS. However, because Linux is monolith
 ic, kernel modules are free to access any address with maximum permissions
 . A po...\n\n\nThomas Filipiuk, Nick Wanninger, Nadharm Dhiantravan, and C
 arson Surmeier (Northwestern University); Alex Bernat (Harvard University)
 ; and Peter Dinda (Northwestern University)\n---------------------\nFine-G
 rained Accelerator Partitioning for Machine Learning and Scientific Comput
 ing in Function as a Service Platform\n\nFunction-as-a-service (FaaS) is a
  promising execution environment for high-performance computing (HPC) and 
 machine learning (ML) applications, as it offers developers a simple way t
 o write and deploy programs. Nowadays, GPUs and other accelerators are ind
 ispensable for HPC and ML workloads. However,...\n\n\nAditya Dhakal and Ph
 ilipp Raith (Hewlett Packard Labs), Logan Ward (Argonne National Laborator
 y (ANL)), Rolando P. Hong Enriquez and Gourav Rattihalli (Hewlett Packard 
 Labs), Kyle Chard (University of Chicago), Ian Foster (Argonne National La
 boratory (ANL)), and Dejan Milojicic (Hewlett Packard Labs)\n-------------
 --------\nROSS – Opening Panel:  Is Accelerator Firmware the New HPC OS?  
 Opportunities and Challenges for the OS/R Research Community\n\nThe propos
 ition for the panel is that specialized (lightweight) OS kernel architectu
 res have become the dominant OS architecture for large-scale HPC systems, 
 with the caveat that these specialized OSes are located in the opaque firm
 ware blobs running on hardware accelerators (primarily GPUs). It is ...\n\
 n\nJohn Lange (Oak Ridge National Laboratory (ORNL), University of Pittsbu
 rgh); Giri Chukkapalli (NVIDIA Corporation); Larry Kaplan (Hewlett Packard
  Enterprise (HPE)); Stefan Lankes (RWTH Aachen University); Kevin Pedretti
  (Sandia National Laboratories); and Robert W. Wisniewski (Samsung)\n-----
 ----------------\nAnalysis and Characterization of Performance Variability
  for OpenMP Runtime\n\nIn the high performance computing (HPC) domain, per
 formance variability is a major scalability issue for parallel computing a
 pplications with heavy synchronization and communication.  We present an e
 xperimental performance analysis of OpenMP benchmarks regarding the variat
 ion of execution time, and ...\n\n\nMinyu Cui, Nikela Papadopoulou, and Mi
 quel Pericàs (Chalmers University of Technology, Sweden)\n----------------
 -----\nGPU Acceleration in Unikernels Using Cricket GPU Virtualization\n\n
 To achieve maximum performance on current heterogeneous architectures, app
 lications have to be tailored to the available hardware by using special A
 PIs to interact with the hardware resources, such as the CUDA APIs for NVI
 DIA GPUs. Simultaneously, unikernels emerge as a solution for the increasi
 ng o...\n\n\nNiklas Eiling, Martin Kröning, Jonathan Klimt, Philipp Fensch
 , Stefan Lankes, and Antonello Monti (RWTH Aachen University)\n-----------
 ----------\nROSS – Morning Break\n---------------------\n13th Internationa
 l Workshop on Runtime and Operating Systems for Supercomputers (ROSS)\n\nT
 he complexity of node architectures in supercomputers increases as we cros
 s milestones on the way toward exascale and beyond. Increasing levels of p
 arallelism in multi- and many-core chips and emerging heterogeneity of com
 putational resources coupled with energy and memory constraints force a re
 eva...\n\n\nBalazs Gerofi (Intel Corporation, RIKEN); Torsten Hoefler (ETH
  Zurich - Swiss Federal Institute of Technology, Microsoft Corporation); a
 nd Kamil Iskra (Argonne National Laboratory (ANL))\n---------------------\
 nROSS – Welcome and Introduction\n\nKamil Iskra (Argonne National Laborato
 ry (ANL))\n\nTag: Middleware and System Software, Programming Frameworks a
 nd System Software, Runtime Systems\n\nRegistration Category: Workshop Reg
  Pass\n\nSession Chairs: Balazs Gerofi (Intel Corporation, RIKEN Center fo
 r Computational Science (R-CCS)); Torsten Hoefler (ETH Zürich, Microsoft C
 orporation); and Kamil Iskra (Argonne National Laboratory (ANL))
END:VEVENT
END:VCALENDAR