BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20240626T180033Z
LOCATION:3001\, 3rd Floor
DTSTART;TZID=America/Los_Angeles:20240626T170000
DTEND;TZID=America/Los_Angeles:20240626T171500
UID:dac_DAC 2024_sess161_RESEARCH909@linklings.com
SUMMARY:Cache-aware Task Decomposition for Efficient Intermittent Computin
 g Systems
DESCRIPTION:Research Manuscript\n\nShuo Xu, Wei Zhang, Mengying Zhao, Zime
 ng Zhou, and Lei Ju (Shandong University)\n\nEnergy harvesting offers a sc
 alable and cost-effective power solution for IoT devices, but it introduce
 s the challenge of frequent and unpredictable power failures due to the un
 stable environment. \nTo address this, intermittent computing has been pro
 posed, which periodically backs up the system state to non-volatile memory
  (NVM), enabling robust and sustainable computing even in the face of unre
 liable power supplies.\nIn modern processors, write back cache is extensiv
 ely utilized to enhance system performance. \nHowever, it poses a challeng
 e during backup operations as it buffers updates to memory, potentially le
 ading to inconsistent system states.\nOne solution is to adopt a write-thr
 ough cache, which avoids the inconsistency issue but incurs increased memo
 ry access latency for each write reference. \nSome existing work enforces 
 a cache flushing before backups to maintain a consistent system state, res
 ulting in significant backup overhead.\nIn this paper, we point out that a
 lthough cache delays updates to the main memory, it may preserve a recover
 able system state in the main memory. \nLeveraging this characteristic, we
  propose a cache-aware task decomposition method that divides an applicati
 on into multiple tasks, ensuring that no dirty cache lines are evicted dur
 ing their execution.\nFurthermore, the cache-aware task decomposition main
 tains a unchanged memory state during the execution of each task, enabling
  us to parallelize the backup process with task execution and effectively 
 hide the backup latency.\nExperimental results with different power traces
  demonstrate the effectiveness of the proposed system.\n\nTopic: Embedded 
 Systems\n\nKeyword: Embedded Software
END:VEVENT
END:VCALENDAR