BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230124T171524Z
LOCATION:C155
DTSTART;TZID=America/Chicago:20221114T153000
DTEND;TZID=America/Chicago:20221114T160000
UID:submissions.supercomputing.org_SC22_sess453_ws_pmbsf108@linklings.com
SUMMARY:Benchmarking Fortran DO CONCURRENT on CPUs and GPUs Using BabelStr
 eam
DESCRIPTION:Workshop\n\nBenchmarking Fortran DO CONCURRENT on CPUs and GPU
 s Using BabelStream\n\nHammond, Deakin, Cownie, McIntosh-Smith\n\nFortran 
 DO CONCURRENT has emerged as a new way to achieve parallel execution of lo
 ops on CPUs and GPUs. This paper studies the performance portability of th
 is construct on a range of processors and compares it with the incumbent m
 odels: OpenMP, OpenACC and CUDA. To do this study fairly, we implemented t
 he BabelStream memory bandwidth benchmark from scratch, entirely in modern
  Fortran, for all of the models considered, which include Fortran DO CONCU
 RRENT, as well as two variants of OpenACC, four variants of OpenMP (2 CPU 
 and 2 GPU), CUDA Fortran, and both loop- and array-based references. Babel
 Stream Fortran matches the C++ implementation as closely as possible, and 
 can be used to make language-based comparisons. This paper represents one 
 of the first detailed studies of the performance of Fortran support on het
 erogeneous architectures; we include results for AArch64 and x86_64 CPUs a
 s well as AMD, Intel and NVIDIA GPU platforms.\n\nSession Format: Recorded
 \n\nTag: Applications, Architectures, Benchmarking, Exascale Computing, Mo
 deling and Simulation, Performance, Performance Portability\n\nRegistratio
 n Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR