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:20260422T000711Z
LOCATION:601
DTSTART;TZID=America/Denver:20231113T103000
DTEND;TZID=America/Denver:20231113T105000
UID:submissions.supercomputing.org_SC23_sess443_ws_ai4dev102@linklings.com
SUMMARY:MPI-RICAL: Data-Driven MPI Distributed Parallelism Assistance with
  Transformers
DESCRIPTION:Nadav Schneider and Tal Kadosh (Ben-Gurion University of the N
 egev, Israel; IAEC); Niranjan Hasabnis and Timothy Mattson (Intel Labs); Y
 uval Pinter (Ben-Gurion University of the Negev, Israel); and Gal Oren (Te
 chnion - Israel Institute of Technology)\n\nMessage Passing Interface (MPI
 ) plays a crucial role in distributed memory parallelization across multip
 le nodes. However, parallelizing MPI code manually, and specifically, perf
 orming domain decomposition, is a challenging, error-prone task. In this p
 aper, we address this problem by developing MPI-RICAL, a novel data-driven
 , programming-assistance tool that assists programmers in writing domain d
 ecomposition based distributed memory parallelization code. Specifically, 
 we train a supervised language model to suggest MPI functions and their pr
 oper locations in the code on the fly. We also introduce MPICodeCorpus, th
 e first publicly available corpus of MPI-based parallel programs that is c
 reated by mining more than 15,000 open-source repositories on GitHub. Expe
 rimental results have been done on MPICodeCorpus and more importantly, on 
 a compiled benchmark of MPI-based parallel programs for numerical computat
 ions that represent real-world scientific applications. MPI-RICAL achieves
  F1 scores between 0.87-0.91 on these programs, demonstrating its accuracy
  in suggesting correct MPI functions at appropriate code locations.\n\nTag
 : Artificial Intelligence/Machine Learning, Software Engineering\n\nRegist
 ration Category: Workshop Reg Pass\n\nSession Chairs: Giorgis Georgakoudis
  (Lawrence Livermore National Laboratory (LLNL)), Ignacio Laguna (Lawrence
  Livermore National Laboratory (LLNL)), and Konstantinos Parasyris (Lawren
 ce Livermore National Laboratory (LLNL))\n\n
END:VEVENT
END:VCALENDAR