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:20260422T000616Z
LOCATION:708
DTSTART;TZID=America/Denver:20231113T090000
DTEND;TZID=America/Denver:20231113T173000
UID:submissions.supercomputing.org_SC23_sess451@linklings.com
SUMMARY:ESPM2 2023: Eighth International Workshop on Extreme Scale Program
 ming Models and Middleware
DESCRIPTION:Challenge on Extreme-Hetero Application Programming\n\nGPU-cen
 tric accelerated supercomputing is still on the main stream for HPC and AI
  applications. However, in the next generation's systems, we need to consi
 der wider variety of accelerators in different style of systems from the a
 rchitecture level. On such complicated systems, what is the best way of...
 \n\n\nTaisuke Boku (University of Tsukuba, Japan)\n---------------------\n
 Cross-Stack System Techniques for Trillion-Parameter Scale Model Inference
 \n\nLeon Song (Microsoft Corporation)\n---------------------\nProgramming 
 Model for Habana/Gaudi2 Accelerators and Its Impact on Deep Learning Infer
 ence/Training Performance at Scale\n\nI will discuss the multi-stream base
 d execution environment of Habana/Gaudi systems that is exposed to deep le
 arning frameworks and I will show how one can combine compute, networking 
 and DMA at high performance and with low run-time overheads. I will highli
 ght the performance of Habana Collective C...\n\n\nKarthikeyan Vaidyanatha
 n (Intel Corporation)\n---------------------\nThe MI300 APU:  Programming 
 for CPUs and GPUs on a Single Package\n\nModern extreme scale computing sy
 stems rely on heterogeneous CPU and GPU architectures. While this design h
 as enabled several remarkable achievements in high-performance computing, 
 applications running at exascale have already identified multiple opportun
 ities where this paradigm can be improved; no...\n\n\nNicholas Malaya (Adv
 anced Micro Devices (AMD) Inc)\n---------------------\nTop 5 Challenges  i
 n Programming Models and Runtimes for Large Language Models Training/Infer
 ence\n\nIn this panel, we focus on the challenges in programming models an
 d runtime system for large language model training/inference. We invite re
 searchers across academia, national labs, and industry to share their expe
 rience and vision on programming tools, runtime performance, architecture,
  optimizatio...\n\n\nZhao Zhang (Rutgers University), Rick Stevens (Univer
 sity of Chicago), Rio Yokota (Tokyo Institute of Technology), Leon Song (M
 icrosoft Corporation), and Torsten Hoefler (ETH Zurich - Swiss Federal Ins
 titute of Technology)\n---------------------\nWho's Winning the Performanc
 e Portability Race on GPU Platforms?\n\nEnsuring high productivity in scie
 ntific software development necessitates developing and maintaining a sing
 le codebase that can run efficiently on a range of accelerator-based super
 computing platforms. This requires the use of performance portability laye
 rs such as OpenMP, RAJA, Kokkos and SYCL for...\n\n\nAbhinav Bhatele (Univ
 ersity of Maryland)\n---------------------\nESPM2 2023 – Morning Break\n--
 -------------------\nPerformance Portability in the Age of Extreme Heterog
 eneity\n\nMoore’s Law is a techno-economic model that has enabled the IT i
 ndustry to double the performance and functionality of digital electronics
  roughly every 2 years within a fixed cost, power and area. This expectati
 on has led to a relatively stable ecosystem (e.g. electronic design automa
 tion too...\n\n\nJohn Shalf (Lawrence Berkeley National Laboratory (LBNL))
 \n---------------------\nDomain-Specific Programming Methodologies for Dom
 ain-Specific and Emerging Computing Systems\n\nProgramming heterogeneous c
 omputing systems is a daunting task which is becoming even more challengin
 g with the advent of emerging, non Von-Neumann computer architectures. Inn
 ovation in programming abstractions and compilers are thus badly needed to
  cope with the current golden age of computer archi...\n\n\nJeronimo Castr
 illon (TU Dresden, Germany)\n---------------------\nFeatured Talk:  Aurora
  Exascale Architecture\n\nAurora is an exascale supercomputer in the final
  stages of assembly at the Argonne Leadership Computing Facility (ALCF) in
  the U.S. This talk will focus on the Aurora hardware and software archite
 ctures with emphasis on the interconnect and programming models, and their
  impact on application perform...\n\n\nKalyan Kumaran (Argonne National La
 boratory (ANL))\n---------------------\nESPM2 – Afternoon Break\n---------
 ------------\nAn Autonomous Execution Model for GPUs:  When CPUs Take a Ba
 ck Seat\n\nIn conventional multi-GPU configurations, the host manages exec
 ution, kernel launches, communication, and synchronization, incurring unne
 cessary overhead. To mitigate this, we present a CPU-free model that deleg
 ates control to the devices themselves, especially benefiting communicatio
 n-intensive app...\n\n\nDidem Unat (Koç University, Turkey)\n-------------
 --------\nESPM2 2023 – Lunch Break\n\nTag: Large Scale Systems, Middleware
  and System Software, Programming Frameworks and System Software\n\nRegist
 ration Category: Workshop Reg Pass\n\nSession Chairs: Dhabaleswar K. (DK) 
 Panda (The Ohio State University), Karl Schulz (Advanced Micro Devices (AM
 D) Inc), Aamir Shafi (The Ohio State University), and Hari Subramoni (The 
 Ohio State University)
END:VEVENT
END:VCALENDAR