BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210402T160210Z
LOCATION:Track 7
DTSTART;TZID=America/New_York:20201119T113000
DTEND;TZID=America/New_York:20201119T120000
UID:submissions.supercomputing.org_SC20_sess292_drs104@linklings.com
SUMMARY:Machine Intelligent and Timely Data Management for Hybrid Memory S
 ystems
DESCRIPTION:Doctoral Showcase\n\nMachine Intelligent and Timely Data Manag
 ement for Hybrid Memory Systems\n\nDoudali, Gavrilovska\n\nBig data analyt
 ics in datacenter platforms and data intensive simulations in exascale com
 puting environments create the need for massive main memory capacities, on
  the order of terabytes, to boost application performance. To satisfy thes
 e requirements, memory hierarchies become more complex, incorporating emer
 ging types of technologies or disaggregation techniques to offset the skyr
 ocketing cost that DRAM-only systems would impose. As we shift away from t
 raditional memory hierarchies, the effectiveness of existing data manageme
 nt solutions decreases, as these have not provisioned against the even big
 ger disparity in the access speeds of the heterogeneous components that ar
 e now part of the memory subsystem. Additionally, system-level configurati
 on knobs need to be re-tuned to adjust to the speeds of the newly introduc
 ed memory hardware. In the face of this complexity, conventional approache
 s to designing data management solutions with empirically-derived configur
 ation parameters become impractical. This makes the case for leveraging ma
 chine intelligence in building a new generation of data management solutio
 ns for hybrid memory systems. This thesis identifies the machine intellige
 nt methods that can be effective for and practically integrated with syste
 m-level memory management, and demonstrates their importance through the d
 esign of new components of the memory management stack; from system-level 
 support for configuring stack parameters to memory scheduling.\n\nRegistra
 tion Category: Tech Program Reg Pass
END:VEVENT
END:VCALENDAR