BEGIN:VCALENDAR
PRODID:-//Microsoft Corporation//Outlook MIMEDIR//EN
VERSION:1.0
BEGIN:VEVENT
DTSTART:20131120T001500Z
DTEND:20131120T020000Z
LOCATION:Mile High Pre-Function
DESCRIPTION;ENCODING=QUOTED-PRINTABLE:ABSTRACT: NVRAM devices can help processing of extremely large-scale graphs with=0Aover DRAM capacity on a single node; however, performance studies and=0Aaccess pattern analyses of graph kernels using both DRAM and NVRAM=0Adevices are limited. In order to address the issue, we propose a graph=0Adata offloading technique by using NVRAM for the hybrid BFS algorithm=0Aand conduct performance analysis of the Hybrid BFS implementation with=0Aour data offloading technique. Experimental results show that our=0Aapproach can achieve 2.8 GTEPS at the maximum and reduce half the size=0Aof DRAM with 47.1% performance degradation. The poster also includes=0Aperformance analyses of our Hybrid BFS approach, which suggests that=0Awe can process large-scale graphs with minimum performance degradation=0Ausing NVRAM by carefully considering the data structures and the=0Aaccess patterns.
SUMMARY:Performance Analysis of Hybrid BFS Approach Using Semi-External Memory
PRIORITY:3
END:VEVENT
END:VCALENDAR