BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20241120T082409Z
LOCATION:HG E 1.2
DTSTART;TZID=Europe/Stockholm:20240605T113000
DTEND;TZID=Europe/Stockholm:20240605T133000
UID:submissions.pasc-conference.org_PASC24_sess127@linklings.com
SUMMARY:MS6D - What the FORTRAN? Lost in Formula Translation
DESCRIPTION:Minisymposium\n\nFortran, the primary programming language und
 erpinning many operational weather and climate codes, was built around the
  fundamental principle that performance optimisation is left to the compil
 er. However, with the emergence of GPU accelerators, significant refactori
 ng, often beyond simple addition of pragmas, is needed to achieve good GPU
  performance on established, operational, vectorised CPU code. This has le
 d to the rise of DSLs and source-to-source methods that often use elements
  from compiler theory to bridge the CPU-GPU gap, leaving one unspoken ques
 tion unanswered: Why the FORTRAN does my compiler not do this for me? In t
 his minisymposium we aim to explore this question by looking at ECMWF's CL
 OUDSC benchmark - an NWP mini-app designed to assess (and torture) compile
 rs. For this benchmark, many GPU-optimised flavours exist, including Fortr
 an and C-based offload flavours (OpenACC, OpenMP, CUDA, HIP), that provide
  an established performance baseline on different GPU architectures. Inste
 ad of further optimising these with more intrusive code changes, we ask th
 e question "How close to the original vector-style Fortran code can we get
  without sacrificing performance?" We aim to explore this question with te
 chnologists and compiler enthusiasts from across the HPC and academic spec
 trum.\n\nMaking Fortran Fly on AMD Instinct Accelerators\n\nFortran has pl
 ayed and is playing a vital tole in the HPC ecosystem, especially in the d
 omain of weather forecasting.  AMD is a key contributor to the LLVM/Flang 
 compiler infrastructure and plays a key role in establishing an LLVM-based
  compiler for Fortran for both CPU and GPU.  In this talk, we wi...\n\n\nM
 ichael Klemm (AMD, OpenMP) and Paul Mullowney (AMD)\n---------------------
 \nImpact of Code Refactoring for Heterogeneous Platforms to Performance Po
 rtability\n\nWe investigate how code refactoring of CLOUDSC benchmark impa
 cts the performance portability across different hardware platforms and co
 de revisions. For hardware platforms, we choose Intel® Xeon® Processor CPU
 s and Intel® Data Center GPU Max Series GPUs. For code revisions, we choos
 e Fortran with Op...\n\n\nMikko Byckling, Camilo Moreno, and Jacob Poulsen
  (Intel Corporation)\n---------------------\nSame Fortran, New Performance
  Portability\n\nFortran is the dominant language on many supercomputers. F
 ortran codes are highly tuned towards CPU performance but struggle to achi
 eve the same performance on GPUs despite their increasing adoption in supe
 rcomputers all over the world. To better support Fortran applications in H
 PC, we automatically...\n\n\nAlexandru Calotoiu and Torsten Hoefler (ETH Z
 urich)\n---------------------\nThe Role of Optimising Compilers in the GPU
  Programming Evolution\n\nGPU programming has traditionally centered aroun
 d established languages like C/C++, Fortran, and more recently, Python. De
 spite this, there exists a diverse range of programming models and APIs fo
 r specifying computation offloading to GPUs. Initially, GPU programming re
 lied on target-specific offloa...\n\n\nAnastasia Stulova (NVIDIA Inc.)\n\n
 Domain: Climate, Weather, and Earth Sciences, Computational Methods and Ap
 plied Mathematics\n\nSession Chairs: Michael Lange (ECMWF) and Balthasar R
 euter (ECMWF)
END:VEVENT
END:VCALENDAR