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:20231112T103000
DTEND;TZID=America/Denver:20231112T110000
UID:submissions.supercomputing.org_SC23_sess426_ws_llvmf109@linklings.com
SUMMARY:Fortran Performance Optimisation and Auto-Parallelization by Lever
 aging MLIR-Based Domain Specific Abstractions in Flang
DESCRIPTION:Nick Brown (Edinburgh Parallel Computing Centre (EPCC), Univer
 sity of Edinburgh); Maurice Jamieson (Edinburgh Parallel Computing Centre 
 (EPCC)); and Anton Lydike, Emilien Bauer, and Tobias Grosser (University o
 f Edinburgh)\n\nMLIR has become popular since it was open sourced in 2019.
  A sub-project of LLVM, the flexibility provided by MLIR to represent Inte
 rmediate Representations (IR) as dialects at different abstraction levels,
  to mix these, and to leverage transformations between dialects provides o
 pportunities for automated program optimisation and parallelisation. In ad
 dition to general purpose compilers built upon MLIR, domain specific abstr
 actions have also been developed. \n\nIn this paper, we explore compliment
 ing the Flang MLIR general purpose compiler by combining with the domain s
 pecific Open Earth Compiler’s MLIR stencil dialect. Developing transformat
 ions to discover and extracts stencils from Fortran, this specialisation d
 elivers between a 2- and 10-times performance improvement for our benchmar
 ks on a Cray supercomputer compared to using Flang alone. Furthermore, by 
 leveraging existing MLIR transformations we develop an auto-parallelisatio
 n approach targeting multi-threaded and distributed memory parallelism, an
 d optimised execution on GPUs, without any modifications to the serial For
 tran source code.\n\nTag: Compilers, Heterogeneous Computing, Performance 
 Optimization\n\nRegistration Category: Workshop Reg Pass\n\nSession Chairs
 : James Brodman (AMD); Ryan Kabrick (Tactical Computing Laboratories LLC, 
 University of Delaware); Patrick S. McCormick (Los Alamos National Laborat
 ory (LANL)); and Alexis Perry-Holby (Los Alamos National Laboratory (LANL)
 )\n\n
END:VEVENT
END:VCALENDAR