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:20260422T000713Z
LOCATION:505
DTSTART;TZID=America/Denver:20231112T153000
DTEND;TZID=America/Denver:20231112T160000
UID:submissions.supercomputing.org_SC23_sess428_misc260@linklings.com
SUMMARY:HPC Software Scaling for ML Using CXL 3.0 GFAM
DESCRIPTION:Patrick Estep (Micron Technology Inc)\n\nTraditional HPC syste
 ms rely on balanced soft scaling, which adjusts the compute-to-memory rati
 o according to the workload. However, this approach is challenged by Machi
 ne Learning applications, especially Large Language Model (LLM) workloads,
  which demand much more memory than compute. This leads to wasted compute 
 resources and excessive data movement in the system. To address this issue
 , we propose to use CXL 3.0 Global Fabric Attached Memory (GFAM), which en
 ables independent scaling of compute and memory and reduces data movement.
  In this talk, we will explore how GFAM architectures require changes in m
 emory and compute placement, as well as software stacks, to optimize perfo
 rmance for LLM workloads.\n\nTag: Distributed Computing, Middleware and Sy
 stem Software, Runtime Systems\n\nRegistration Category: Workshop Reg Pass
 \n\nSession Chairs: Barbara Chapman (Hewlett Packard Enterprise (HPE), Sto
 ny Brook University); Joseph Manzano (Pacific Northwest National Laborator
 y (PNNL)); Shirley Moore (University of Texas at El Paso); EunJung (EJ) Pa
 rk (Qualcomm); and Joshua Suetterlein (Pacific Northwest National Laborato
 ry (PNNL))\n\n
END:VEVENT
END:VCALENDAR