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:20181221T160742Z
LOCATION:D221
DTSTART;TZID=America/Chicago:20181113T113000
DTEND;TZID=America/Chicago:20181113T114500
UID:submissions.supercomputing.org_SC18_sess277_drs105@linklings.com
SUMMARY:In-Memory Accelerator Architectures for Machine Learning and Bioin
 formatics
DESCRIPTION:Doctoral Showcase\nComputational Biology, Exascale, GPUs, Grap
 h Algorithms, Linear Algebra, Machine Learning, Sparse Computation, Worksh
 op Reg Pass, Tutorial Reg Pass, Tech Program Reg Pass, Exhibits Reg Pass, 
 Exhibits - Exhibit Hall Only Reg Pass, Doctoral Showcase\n\nIn-Memory Acce
 lerator Architectures for Machine Learning and Bioinformatics\n\nKaplan, G
 inosar\n\nMost contemporary accelerators are von Neumann machines. With th
 e increasing sizes of gathered and then processed data,  memory bandw
 idth is the main limiting of performance. One approach to mitigate the ban
 dwidth constraint is to bring the processing units closer to the data. Thi
 s approach is known as <em>near-data processing</em> (NDP). However, NDP a
 rchitecture (e.g., Hybrid Memory Cube) are still inherently limited becaus
 e they are based on replicating the von Neumann architecture in memory.&nb
 sp;<br />My research proposes two new processing-in-storage architectures,
  where each bitcell can both store information and perform computation. Th
 e main building block of the architectures are memristors, an emerging mem
 ory technolgy. <br />The first architecture I propose, PRinS, was applied 
 to accelerate machine learning and large-scale DNA sequence alignment. Usi
 ng Associative Processing, PRinS achieves massive parallelism. The se
 cond, RASSA, accelerates DNA long read mapping, with approximate Hamming d
 istance computation and quantifying mismatch of a pattern to voltage 
 level.
URL:https://sc18.supercomputing.org/presentation/?id=drs105&sess=sess277
END:VEVENT
END:VCALENDAR