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:20211207T054806Z
LOCATION:227-228
DTSTART;TZID=America/Chicago:20211117T133000
DTEND;TZID=America/Chicago:20211117T140000
UID:submissions.supercomputing.org_SC21_sess146_pap199@linklings.com
SUMMARY:Single-Node Partitioned-Memory for Huge Graph Analytics: Cost and 
 Performance Trade-Offs
DESCRIPTION:Paper\n\nSingle-Node Partitioned-Memory for Huge Graph Analyti
 cs: Cost and Performance Trade-Offs\n\nGhosh, Tallent, Minutoli, Halappana
 var, Peri...\n\nBecause of cost, non-volatile memory NVDIMMs such as Intel
  Optane are attractive in single-node big-memory systems. We evaluate perf
 ormance and cost trade-offs when using Optane as volatile memory for huge-
 graph analytics. We study two scalable graph applications with different w
 ork locality, access patterns and parallelism. We evaluate single and part
 itioned address spaces; Memory and AppDirect modes; and compare with distr
 ibuted executions on GPU-accelerated and CPU-based supercomputers.\n\nWe s
 how that AppDirect can perform and scale better than Memory for the larges
 t working sets (12%), even when dominated by irregular access patterns, if
  most accesses are NUMA-local and Optane accesses are frequently reads. Su
 rprisingly, between Memory and AppDirect, processor-cache performance can 
 change due to line invalidations; updates to the caching policy (via non-t
 emporal hints) can make a 25% improvement. We observe that single-node gra
 ph analytics frequently has >4–10x cost/performance advantages over distri
 buted-memory executions on supercomputers.\n\nTag: Algorithms, Extreme Sca
 le Computing, Heterogeneous Systems\n\nRegistration Category: Tech Program
  Reg Pass
END:VEVENT
END:VCALENDAR