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:20240116T191658Z
LOCATION:601
DTSTART;TZID=America/Denver:20231113T115000
DTEND;TZID=America/Denver:20231113T123000
UID:submissions.supercomputing.org_SC23_sess443_misc153@linklings.com
SUMMARY:Unlocking the Potential of Large Language Models for High-Performa
 nce Computing Code
DESCRIPTION:Workshop\n\nGal Oren (Technion - Israel Institute of Technolog
 y)\n\nHigh-Performance Computing (HPC) has long been the driving force beh
 ind advancements in science, engineering, and beyond. Yet, realizing the f
 ull potential of HPC applications has often been hampered by the intricate
  nature of programming for the underlying parallel systems. In this keynot
 e, we explore a transformative approach that bridges the gap between human
  ingenuity and computational power using the capabilities of large languag
 e models (LLMs).\n\nOur research is an exploration of how cutting-edge LLM
 s can be tailored to the demanding domain of HPC, where computational spee
 d and efficiency reign supreme. While LLMs have showcased remarkable profi
 ciency in understanding and generating code, their training data primarily
  comes from general-purpose codebases. In stark contrast, HPC code involve
 s intricate mathematical modeling, parallelism, and optimization, demandin
 g customized adaptations.\n\nThat is why our journey for ‘HPC LLMs’ began 
 with the collection of an extensive dataset, HPCorpus, that represents the
  culmination of HPC code in C, C++, and Fortran from diverse domains. Arme
 d with this invaluable resource, we embarked on an ambitious mission to en
 hance the capabilities of language models in the realm of HPC. The creatio
 n of Tokompiler, a pioneering HPC-specific code tokenizer, marked a pivota
 l turning point. Tokompiler, designed to preprocess code for language mode
 ls, introduced a revolutionary approach that harnessed abstract syntax tre
 es (ASTs) into the source code itself and reshaped the way language models
  comprehend and generate code, resembling how compilers perceive our codes
 , not humans. Building upon this innovation, we undertook comprehensive pr
 e-training efforts with CompCoder, adapting transformer-based language mod
 els to the intricacies of HPC. This journey has culminated in novel downst
 ream tasks, including the generation of OpenMP and MPI code, where our mod
 els shine by transforming serial code into efficient parallel one. Togethe
 r, these milestones represent a great leap forward in the convergence of A
 I and HPC from a different perspective, promising to redefine the landscap
 e of computational science.\n\nAs we stand at the crossroads of AI and HPC
 , the possibilities are boundless. Our journey is merely the prologue, unv
 eiling a multitude of untapped opportunities in HPC code comprehension, ge
 neration, and optimization. From refining domain-specific code to tackling
  complex simulations and accelerating scientific breakthroughs, the horizo
 ns are vast. The symbiotic partnership between LLMs and HPC promises to re
 volutionize how HPC practitioners write code. Looking ahead, we envision a
  future where LLMs for HPC become indispensable tools for researchers and 
 developers in their quest for unprecedented speed, accuracy, and efficienc
 y.\n\nTag: Artificial Intelligence/Machine Learning, Software Engineering\
 n\nRegistration Category: Workshop Reg Pass\n\nSession Chairs: Giorgis Geo
 rgakoudis (Lawrence Livermore National Laboratory (LLNL)), Ignacio Laguna 
 (Lawrence Livermore National Laboratory (LLNL)), and Konstantinos Parasyri
 s (Lawrence Livermore National Laboratory (LLNL))
END:VEVENT
END:VCALENDAR