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:20260422T000611Z
LOCATION:702
DTSTART;TZID=America/Denver:20231113T090000
DTEND;TZID=America/Denver:20231113T173000
UID:submissions.supercomputing.org_SC23_sess449@linklings.com
SUMMARY:The 6th Annual Parallel Applications Workshop, Alternatives to MPI
 +X (PAW-ATM)
DESCRIPTION:As supercomputers become more and more powerful, the number an
 d diversity of applications that can be tackled with these machines grows.
  Unfortunately, the architectural complexity of these supercomputers grows
  as well, with heterogeneous processors, multiple levels of memory hierarc
 hy, and many ways to move data and synchronize between processors. The MPI
 +X programming model, use of which is considered by many to be standard pr
 actice, demands that a programmer be expert in both the application domain
  and the low-level details of the architecture(s) on which that applicatio
 n will be deployed, and the availability of such superhuman programmers is
  a critical bottleneck. Things become more complicated when evolution and 
 change in the underlying architecture translates into significant re-engin
 eering of the MPI+X code to maintain performance.\n\nNumerous alternatives
  to the MPI+X model exist, and by raising the level of abstraction on the 
 application domain and/or the target architecture, they offer the ability 
 for “mere mortal” programmers to take advantage of the supercomputing reso
 urces that are available to advance science and tackle urgent real-world p
 roblems. However, compared to the MPI+X approach, these alternatives gener
 ally lack two things. First, they aren’t as well known as MPI+X and a doma
 in scientist may simply not be aware of models that are a good fit to thei
 r domain. Second, they are less mature than MPI+X and likely have more fun
 ctionality or performance “potholes” that need only be identified to be ad
 dressed.\n\nPAW-ATM is a forum for discussing HPC applications written in 
 alternatives to MPI+X. Its goal is to bring together application experts a
 nd proponents of high-level languages to present concrete example uses of 
 such alternatives, describing their benefits and challenges.\n\nSurvey of 
 Technologies for Developers of Parallel Applications:  Legate and cuNumeri
 c\n\nThis talk is about Legate, a framework for scalable and composable di
 stributed software. Legate enables cuNumeric, a GPU accelerated distribute
 d NumPy library, to grow rapidly by providing high-productivity abstractio
 ns on top of a scalable runtime system. In this talk, we will explain how 
 Legate en...\n\n\nRohan Yadav (Stanford University)\n---------------------
 \nHigh-Performance Programming and Execution of a Coral Biodiversity Mappi
 ng Algorithm Using Chapel\n\nWe will demonstrate how the parallelism and e
 xpressiveness of the Chapel programming language are used to achieve an en
 ormous improvement in computational speed for a problem related to coral r
 eef conservation. Chapel’s concise syntax and versatile data structures en
 able this problem to be sol...\n\n\nScott Bachman (National Center for Atm
 ospheric Research (NCAR)), Rebecca Green (The Coral Reef Alliance), Anna B
 akker (University of Miami), Helen Fox (The Coral Reef Alliance), Sam Purk
 is (University of Miami), and Ben Harshbarger (Hewlett Packard Enterprise 
 (HPE))\n---------------------\nPAW-ATM – Morning Break\n------------------
 ---\nPAW-ATM – Afternoon Break\n---------------------\nDesign and Analysis
  of the Network Software Stack of an Asynchronous Many-Task System – The L
 CI Parcelport of HPX\n\nThe HPX asynchronous many-task runtime system has 
 been using TCP and MPI as its communication backends (parcelports). We dev
 eloped a new HPX parcelport using a new communication library, the Lightwe
 ight Communication Interface (LCI) that was designed to better match the n
 eeds of systems such as HPX....\n\n\nJiakun Yan (University of Illinois), 
 Hartmut Kaiser (Louisiana State University), and Marc Snir (University of 
 Illinois)\n---------------------\nSurvey of Technologies for Developers of
  Parallel Applications:  SHMEM\n\nOpenSHMEM was introduced more than a dec
 ade ago to standardize SHMEM, a library-based communications interface tha
 t was originally developed as a proprietary application interface by Cray 
 for their T3D systems.  An alternative to MPI that implements a Partitione
 d Global Address Space (PGAS) programm...\n\n\nBarbara Chapman (Hewlett Pa
 ckard Enterprise (HPE))\n---------------------\nsymPACK:  A GPU-Capable Fa
 n-Out Sparse Cholesky Solver\n\nSparse symmetric positive definite systems
  of equations are ubiquitous in scientific workloads and applications. Par
 allel sparse Cholesky factorization is the method of choice for solving su
 ch linear systems. Therefore, the development of parallel sparse Cholesky 
 codes that can efficiently run on to...\n\n\nJulian Bellavita (Lawrence Be
 rkeley National Laboratory (LBNL)); Mathias Jacquelin (Cerebras Systems); 
 and Esmond Ng, Dan Bonachea, Johnny Corbino, and Paul H. Hargrove (Lawrenc
 e Berkeley National Laboratory (LBNL))\n---------------------\nPure:  Evol
 ving Message Passing To Better Leverage Shared Memory within Nodes\n\nPure
  is a new programming model and runtime system explicitly designed to take
  advantage of shared memory within nodes in the context of a mostly messag
 e passing interface enhanced with the ability to use tasks to make use of 
 idle cores. We use microbenchmarks to evaluate Pure’s key messaging ...\n\
 n\nJames Psota and Armando Solar-Lezama (Massachusetts Institute of Techno
 logy (MIT))\n---------------------\nIntroduction to The 6th Annual Paralle
 l Applications Workshop, Alternatives to MPI+X\n\nAs supercomputers become
  more and more powerful, the number and diversity of applications that can
  be tackled with these machines grows. Unfortunately, the architectural co
 mplexity of these supercomputers grows as well, with heterogeneous process
 ors, multiple levels of memory hierarchy, and many way...\n\n\nKarla V. Mo
 rris Wright (Sandia National Laboratories)\n---------------------\nSurvey 
 of Technologies for Developers of Parallel Applications:  Julia\n\nThis ta
 lk will highlight distributed and gpu computing using the Julia ecosystem.
  The Julia language proposes an integrated development end-to-end co-desig
 n model as a LLVM front-end for science to close the gap between high-prod
 uctivity languages and the desired performance of traditional compiled ...
 \n\n\nJohannes Blaschke (Lawrence Berkeley National Laboratory (LBNL))\n--
 -------------------\nPAW-ATM Distinguished Speaker:  Ethan Gutmann – Natio
 nal Center for Atmospheric Research: Trials and Tribulations and Joys of D
 eveloping with Alternative Parallel Frameworks\n\nDeveloping large scienti
 fic applications is challenging for many reasons, and alternative programm
 ing can help with better support for the implementation.  These applicatio
 ns need to incorporate the latest domain specific scientific information, 
 be applicable to real world problems, and be robust ac...\n\n\nEthan Gutma
 nn (National Center for Atmospheric Research (NCAR))\n--------------------
 -\nSurvey of Technologies for Developers of Parallel Applications:  Q&A\n\
 nThe speakers will respond to Q&A for the technologies they presented in t
 he session.\n\n\nBarbara Chapman (Hewlett Packard Enterprise (HPE)), Justi
 n Wozniak (Argonne National Laboratory (ANL)), Johannes Blaschke (Lawrence
  Berkeley National Laboratory (LBNL)), and Rohan Yadav (Stanford Universit
 y)\n---------------------\nImplementing Scalable Matrix-Vector Products fo
 r the Exact Diagonalization Methods in Quantum Many-Body Physics\n\nExact 
 diagonalization is a well-established method for simulating small quantum 
 systems. Its applicability is limited by the exponential growth of the Ham
 iltonian matrix that needs to be diagonalized. Physical symmetries are usu
 ally utilized to reduce the matrix dimension, and distributed-memory par..
 .\n\n\nTom Westerhout (Radboud University) and Bradford L. Chamberlain (He
 wlett Packard Enterprise (HPE))\n---------------------\nThe 6th Annual Par
 allel Applications Workshop, Alternatives to MPI+X (PAW-ATM)\n\nSupercompu
 ters get faster and more complex every year. MPI, long the dominant model 
 for distributed computation, has adapted by combining with models for intr
 a-node parallelism (e.g. OpenMP, CUDA). These MPI+X hybrids offer performa
 nce but demand significant programmer effort to write, debug and tun...\n\
 n\nKarla Vanessa Morris Wright (Sandia National Laboratories); Elliott Sla
 ughter (SLAC National Accelerator Laboratory); Engin Kayraklioglu (Hewlett
  Packard Enterprise (HPE)); Irene Moulitsas (Cranfield University, England
 ); Daniele Lezzi (Barcelona Supercomputing Center (BSC)); and William Long
  (Hewlett Packard Enterprise (HPE))\n---------------------\nPAW-ATM Panel 
 Discussion:  Charting Paths to Success with Alternatives to MPI+X\n\nDiffe
 rent aspects of the workshop and other questions from the moderator and au
 dience will be discussed in the panel.\n\n\nMichelle Strout (Hewlett Packa
 rd Enterprise (HPE)), Ewa Deelman (University of Southern California), Jon
 athan Graham (Los Alamos National Laboratory (LANL)), Ethan Gutmann (Natio
 nal Center for Atmospheric Research (NCAR)), Laxmikant Kale (University of
  Illinois), and George Karypis (University of Minnesota)\n----------------
 -----\nPAW-ATM – Lunch Break\n---------------------\nshmem4py: High-Perfor
 mance One-Sided Communication for Python Applications\n\nWe describe shmem
 4py, a Python wrapper for the OpenSHMEM application programming interface 
 (API) which follows a design similar to that of the well-known mpi4py pack
 age. OpenSHMEM is a descendant of the one-sided communication library for 
 the Cray T3D and it is known for its uncompromising performan...\n\n\nMarc
 in Rogowski (King Abdullah University of Science and Technology (KAUST), N
 VIDIA Corporation); Jeff R. Hammond (NVIDIA Helsinki Oy); and David E. Key
 es and Lisandro Dalcin (King Abdullah University of Science and Technology
  (KAUST))\n---------------------\nSurvey of Technologies for Developers of
  Parallel Applications:  Swift/T\n\nI will describe how Swift/T, an automa
 tically parallel scripting language, serves as an alternative to MPI by cr
 eating a higher-level programming model for workflow-like applications.  S
 wift/T essentially translates a functional description of a workflow into 
 an MPI program runnable at the largest s...\n\n\nJustin Wozniak (Argonne N
 ational Laboratory (ANL))\n\nTag: Accelerators, Artificial Intelligence/Ma
 chine Learning, Algorithms, Applications, Architecture and Networks, Distr
 ibuted Computing, Compilers, Data Analysis, Visualization, and Storage, Ex
 ascale, Heterogeneous Computing, Linear Algebra, Message Passing, Performa
 nce Optimization, Programming Frameworks and System Software, Quantum Comp
 uting, Runtime Systems, Software Engineering, Sustainability, Task Paralle
 lism, Tensors\n\nRegistration Category: Workshop Reg Pass\n\nSession Chair
 s: Engin Kayraklioglu (Hewlett Packard Enterprise (HPE)), Daniele Lezzi (B
 arcelona Supercomputing Center (BSC)), Bill Long (Retired), Karla Vanessa 
 Morris Wright (Sandia National Laboratories), Irene Moulitsas (Cranfield U
 niversity), and Elliott Slaughter (SLAC National Accelerator Laboratory)
END:VEVENT
END:VCALENDAR