BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20250626T234223Z
LOCATION:B304
DTSTART;TZID=America/New_York:20241118T140000
DTEND;TZID=America/New_York:20241118T173000
UID:submissions.supercomputing.org_SC24_sess761@linklings.com
SUMMARY:High Performance Python for Science at Scale
DESCRIPTION:This symposium-style workshop aims to connect researchers, dev
 elopers, and Python practitioners to share their experiences scaling Pytho
 n-based applications and workflows on supercomputers. The goal is to provi
 de a platform for topical discussion of best practices, hands-on demonstra
 tions, and community engagement via open-source contributions to new libra
 ries, runtimes, and frameworks. Based on talks and demos that survey and s
 ummarize the best practices and recent success stories and developments – 
 the workshop will serve as a requirements gathering exercise for the futur
 e of Python in HPC and science.\n\nHigh Performance Python for Science at 
 Scale — Afternoon Break\n---------------------\nAccelerating Python Applic
 ations with Dask and ProxyStore\n\nApplications are increasingly written a
 s dynamic workflows underpinned by an execution framework that manages asy
 nchronous computations across distributed hardware. However, execution fra
 meworks typically offer one-size-fits-all solutions for data flow manageme
 nt, which can restrict performance and ...\n\n\nJ. Gregory Pauloski (Unive
 rsity of Chicago, Argonne National Laboratory (ANL)); Klaudiusz Rydzy (Loy
 ola University, Chicago); and Valerie Hayot-Sasson, Ian Foster, and Kyle C
 hard (University of Chicago, Argonne National Laboratory (ANL))\n---------
 ------------\nInvited Speaker Presentation\n\nMatthew Rocklin (Dask/Colied
 )\n---------------------\nVisualizing Workflows with the Dragon Telemetry 
 Service\n\nThe Dragon telemetry service is an easy-to-use, scalable means 
 for users to visualize both hardware and custom metrics for complex workfl
 ows. We discuss in-depth the Dragon runtime, the architecture and capabili
 ties of the telemetry service, and how the telemetry service compares to e
 xisting tools. ...\n\n\nIndira Pimpalkhare, Colin Wahl, and Maria Kalantzi
  (Hewlett Packard Enterprise (HPE))\n---------------------\nExploring Data
  at Scale with Arkouda: A Practical Introduction to Scalable Data Science\
 n\nData scientists can be thought of as modern-day explorers, venturing in
 to the vast unknown of information. However, this exciting journey is not 
 without its hurdles. One of the biggest challenges they face is the sheer 
 immensity of data they encounter. Modern datasets cannot fit in laptop mem
 ory, co...\n\n\nBen McDonald (Hewlett Packard Enterprise (HPE))\n---------
 ------------\nWorkshop Introduction\n\nPete Mendygral (HPE)\n-------------
 --------\nWork-in-progress: CUDA Python object models and parallelism mode
 ls\n\nToday Python developers typically access GPUs from either deep learn
 ing frameworks or tools that haven't kept up with modern CUDA practices. I
 n this work in progress update, the CUDA Python team will demonstrate new 
 interfaces using the CUDA Core Compute Libraries and an updated Pythonic o
 bject mode...\n\n\nAndy Terrel (NVIDIA Corporation)\n---------------------
 \nLightning Talks\n\nPete Mendygral (HPE)\n---------------------\nSeamless
 ly scale your python program from single CPU core to multi-GPU multi-node 
 HPC cluster with cuNumeric\n\nPython is a powerful and user-friendly progr
 amming language widely adopted by researchers and scientists. As data size
 s and computational complexities grow, CPU-based Python struggles to meet 
 the speed and scale demanded by cutting-edge research. Distributed acceler
 ated computing offers an infrastru...\n\n\nWonchan Lee, Manolis Papadakis,
  Mike Bauer, and Bo Dong (NVIDIA Corporation)\n---------------------\nPyOM
 P: Parallel programming for CPUs and GPUs with OpenMP and Python\n\nPython
  is the most popular programming language.  OpenMP is the most\npopular pa
 rallel programming API. Projecting OpenMP into Python\nwill help expand th
 e HPC community.  We \ncall our Python-based OpenMP system PyOMP.\n\nIn th
 is short paper we describe PyOMP and \nits use for parallel programming\nf
 or CP...\n\n\nGiorgis Georgakoudis (Lawrence Livermore National Laboratory
  (LLNL)), Todd Anderson (Intel Corporation), Stuart Archibald (The Numba p
 roject), Bronis de Supinski (Lawrence Livermore National Laboratory (LLNL)
 ), and Timothy Mattson (Human Learning Group)\n\nTag: Applications and App
 lication Frameworks, Artificial Intelligence/Machine Learning, Parallel Pr
 ogramming Methods, Models, Languages and Environments\n\nRegistration Cate
 gory: Workshop Reg Pass\n\nSession Chairs: Sunita Chandrasekaran (Universi
 ty of Delaware), Sam Foreman (Argonne National Laboratory (ANL)), Daniel M
 argala (Lawrence Berkeley National Laboratory (LBNL)), and Pete Mendygral 
 (Cray Inc.)
END:VEVENT
END:VCALENDAR