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:20240116T190002Z
LOCATION:710
DTSTART;TZID=America/Denver:20231112T090000
DTEND;TZID=America/Denver:20231112T123000
UID:submissions.supercomputing.org_SC23_sess419@linklings.com
SUMMARY:5th Workshop on Programming and Performance Visualization Tools (P
 roTools 2023)
DESCRIPTION:Workshop\n\nFROOM:  A Framework of Operators for OTF2 Modifica
 tion\n\nIn recent years, High Performance Computing (HPC) has become incre
 asingly important for many industries and research areas besides ‘classic’
  applications.  As new domains emerge, applications, implementations and f
 rameworks become more diverse. Generic performance analysis tools often ca
 ...\n\n\nJan Frenzel, Apurv Kulkarni, and Sebastian Döbel (Technical Unive
 rsity Dresden); Bert Wesarg and Maximilian Knespel (GWT-TUD GmbH); and Hol
 ger Brunst (Technical University Dresden)\n---------------------\nGPUscout
 : Locating Data Movement-Related Bottlenecks on GPUs\n\nGPUs pose an attra
 ctive opportunity for delivering high-performance applications. However, G
 PU codes are often limited due to memory contention, resulting in overall 
 performance degradation. Since GPU scheduling is transparent to the user, 
 and GPU memory architectures are very complex compared to on...\n\n\nSoumy
 a Sen, Stepan Vanecek, and Martin Schulz (Technical University of Munich)\
 n---------------------\nProTools 2023 – Morning Break\n-------------------
 --\nAn Event Model for Trace-Based Performance Analysis of MPI Partitioned
  Point-to-Point Communication\n\nThe MPI 4.0 standard introduced the conce
 pt of partitioned point-to-point communication.  One facet that may help i
 n encouraging application developers to use this new concept in their prog
 rams is the availability of proper tool support in a timely manner.  We th
 erefore propose nine new events exten...\n\n\nIsabel Thärigen and Marc-And
 ré Hermanns (RWTH Aachen University) and Markus Geimer (Forschungszentrum 
 Jülich)\n---------------------\nExtra-Deep:  Automated Empirical Performan
 ce Modeling for Distributed Deep Learning\n\nWith the rapidly increasing s
 ize and complexity of DNNs, equally sophisticated methods are needed to tr
 ain them efficiently, including distributed training and various model/hyb
 rid parallelism approaches. Even though developers heavily rely on state-o
 f-the-art frameworks such as PyTorch and TensorFl...\n\n\nMarcus Ritter an
 d Felix Wolf (Technical University of Darmstadt)\n---------------------\nI
 nvited Talk: Using XDMoD for HPC Performance and Quality-of-Service Analys
 is\n\nNikolay Simakov (SUNY University at Buffalo)\n---------------------\
 nEnabling Agile Analysis of I/O Performance Data with PyDarshan\n\nModern 
 scientific applications utilize numerous software and hardware layers to e
 fficiently access data. This approach poses a challenge for I/O optimizati
 on because of the need to instrument and correlate information across thos
 e layers. The Darshan characterization tool seeks to address this chall...
 \n\n\nJakob Luettgau (French Institute for Research in Computer Science an
 d Automation (INRIA)); Shane Snyder (Argonne National Laboratory (ANL)); T
 yler Reddy and Nikolaus Awtrey (Los Alamos National Laboratory (LANL)); Ke
 vin Harms (Argonne National Laboratory (ANL)); Jean Bez (Lawrence Berkeley
  National Laboratory (LBNL)); and Rui Wang, Rob Latham, and Philip Carns (
 Argonne National Laboratory (ANL))\n---------------------\nFiltering and R
 anking of Code Regions for Parallelization via Hotspot Detection and OpenM
 P Overhead Analysis\n\nMany high-performance computing applications reach 
 millions of code lines and hundreds of code regions. Analyzing all code re
 gions for parallelization with OpenMP is neither efficient nor necessary. 
 To facilitate this task and minimize the effort by the user, the code regi
 ons of the application need...\n\n\nSeyed Ali Mohammadi, Lukas Rothenberge
 r, Gustavo de Morais, Bertin Nico Görlich, Erik Lille, Hendrik Rüthers, an
 d Felix Wolf (Technical University of Darmstadt)\n\nTag: Performance Measu
 rement, Modeling, and Tools, Programming Frameworks and System Software\n\
 nRegistration Category: Workshop Reg Pass\n\nSession Chairs: David Boehme 
 (Lawrence Livermore National Laboratory (LLNL)); Anthony Danalis (Universi
 ty of Tennessee); and Josef Weidendorfer (Leibniz Supercomputing Centre, T
 echnical University of Munich)
END:VEVENT
END:VCALENDAR