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:20260422T143006Z
LOCATION:B302-B305
DTSTART;TZID=America/New_York:20241119T120000
DTEND;TZID=America/New_York:20241119T170000
UID:submissions.supercomputing.org_SC24_sess487_drs102@linklings.com
SUMMARY:Toward Performance & Portability & Productivity in Parallel Progra
 mming
DESCRIPTION:Ari Rasch (University of Münster, Germany)\n\nAchieving *perfo
 rmance*, *portability*, and *productivity* for data-parallel computations 
 (e.g., MatMul and convolutions) has emerged as a major research challenge.
  The complex hardware design of contemporary parallel architectures, inclu
 ding GPUs and CPUs, requires advanced program optimizations to fully explo
 it the performance potential of architectures. Furthermore, due to the div
 erse hardware landscape, it has proven challenging to achieve (performance
 ) portability: different architectures require different kinds of optimiza
 tions, thereby often posing challenging, often even contradicting requirem
 ents on code optimization. Also, the complexity of achieving performance a
 nd portability must be hidden behind a user-productive programming interfa
 ce to make programming modern architectures amenable.\n\nThis thesis intro
 duces a novel approach to code *generation* and *optimization* for data-pa
 rallel computations targeting modern parallel architectures. The ultimate 
 goal of our approach is to simultaneously achieve *performance*, *portabil
 ity*, and *productivity*, in one combined approach, which is identified as
  a major research challenge.\n\nThe first part of this thesis introduces t
 he algebraic formalism of Multi-Dimensional Homomorphisms (MDH) — a novel 
 approach to generating code that can be fully automatically optimized (aut
 o-tuned) for a particular target architecture and characteristics of the i
 nput and output data (such as size and memory layout); our code generation
  approach is hidden behind a productive user interface that expresses a wi
 de range of data-parallel computations.\n\nThe second part of this thesis 
 introduces the Auto-Tuning Framework (ATF) for automatically optimizing pa
 rameterized program code (as generated by our MDH approach). In contrast t
 o existing auto-tuners, ATF supports so-called constrained tuning paramete
 rs which are ubiquitous in modern parallel programming.\n\nRegistration Ca
 tegory: Tech Program Reg Pass, Exhibits Reg Pass\n\nSession Chairs: Ayesha
  Afzal (Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen Nation
 al High Performance Computing Center); Sally Ellingson (University of Kent
 ucky); and Alan Sussman (University of Maryland)\n\n
END:VEVENT
END:VCALENDAR