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:20250626T233529Z
LOCATION:B302-B305
DTSTART;TZID=America/New_York:20241120T100000
DTEND;TZID=America/New_York:20241120T170000
UID:submissions.supercomputing.org_SC24_sess533_drs105@linklings.com
SUMMARY:Going Beyond the Chicken and Egg Situation with Modern MPI Feature
 s
DESCRIPTION:Tim Jammer (Technical University Darmstadt)\n\nMPI has become 
 the de facto standard for distributed memory computing since its inception
  in 1994. While the MPI standard has evolved to include new technologies l
 ike RDMA, many applications still rely on the original set of MPI operatio
 ns.\n\nThis thesis investigates the current usage of MPI. We note that dev
 elopers underutilize modern MPI features, as their implementations often a
 re not optimized. On the other hand, as many users rely on the "old" MPI f
 eatures, MPI implementation developers have no incentive to optimize imple
 mentations for the new features. As a consequence, there is no incentive f
 or MPI users to learn the new features, creating a vicious cycle.\n\nTo br
 eak this cycle, this thesis explores three main approaches:\n\n1) Facilita
 ting correctness checking tool support,\n\n2) Modernizing MPI codes with c
 ompiler based approaches, and\n\n3) Exploiting compiler knowledge to furth
 er optimize the implementation of modern MPI features.\n\nIn order to faci
 litate the development and improvement of tools aiding with MPI developmen
 t, this thesis introduces the correctness benchmark MPI-BugBench as a stan
 dardized benchmark to evaluate the real-world applicability of such tools.
  Further, we show that compiler-based automatic modernization methods can 
 encourage early adoption of new MPI features with minimal programmer effor
 t (for example, partitioned operations).\n\nLastly, compiler knowledge can
  be utilized in order to further optimize the performance of MPI implement
 ations (for example, in persistent). The use of compiler knowledge, in par
 ticular, enables modernization of existing MPI codes without the need for 
 application developers to rewrite existing MPI codes.\n\nRegistration Cate
 gory: Tech Program Reg Pass, Exhibits Reg Pass\n\nSession Chairs: Ayesha A
 fzal (Friedrich-Alexander University, Erlangen-Nuremberg; Erlangen Nationa
 l High Performance Computing Center); Sally Ellingson (University of Kentu
 cky); and Alan Sussman (University of Maryland)\n\n
END:VEVENT
END:VCALENDAR