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:20260422T000712Z
LOCATION:403-404
DTSTART;TZID=America/Denver:20231116T110000
DTEND;TZID=America/Denver:20231116T113000
UID:submissions.supercomputing.org_SC23_sess178_pap552@linklings.com
SUMMARY:Calculon: a Methodology and Tool for High-Level Codesign of System
 s and Large Language Models
DESCRIPTION:Mikhail Isaev (Georgia Institute of Technology), Nic McDonald 
 and Larry Dennison (NVIDIA Corporation), and Richard Vuduc (Georgia Instit
 ute of Technology)\n\nThis paper presents a parameterized analytical perfo
 rmance model of transformer-based Large Language Models (LLMs) for guiding
  high-level algorithm-architecture codesign studies. This model derives fr
 om an extensive survey of performance optimizations that have been propose
 d for the training and inference of LLMs; the model's parameters capture a
 pplication characteristics, the hardware system, and the space of implemen
 tation strategies. With such a model, we can systematically explore a join
 t space of hardware and software configurations to identify optimal system
  designs under given constraints, like the total amount of system memory. 
 We implemented this model and methodology in a Python-based open-source to
 ol called Calculon. Using it, we identified novel system designs that look
  significantly different from current inference and training systems, show
 ing quantitatively the estimated potential to achieve higher efficiency, l
 ower cost, and better scalability.\n\nTag: Artificial Intelligence/Machine
  Learning, Codesign, Performance Optimization, Programming Frameworks and 
 System Software\n\nRegistration Category: Tech Program Reg Pass\n\nReprodu
 cibility Badges: Artifact Available, Artifact Functional, Results Reproduc
 ed\n\nSession Chair: Aparna Chandramowlishwaran (University of California,
  Irvine)\n\n
END:VEVENT
END:VCALENDAR