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:20210402T160558Z
LOCATION:Track 6
DTSTART;TZID=America/New_York:20201113T154500
DTEND;TZID=America/New_York:20201113T161500
UID:submissions.supercomputing.org_SC20_sess225_ws_h2rc103@linklings.com
SUMMARY:FPGAs-as-a-Service Toolkit (FaaST)
DESCRIPTION:Workshop\n\nFPGAs-as-a-Service Toolkit (FaaST)\n\nRankin, Krup
 a, Harris, Acosta, Holzman...\n\nComputing needs for high energy physics a
 re already intensive and are expected to increase drastically in the comin
 g years. In this context, heterogeneous computing, specifically as-a-servi
 ce computing, has the potential for significant gains over traditional com
 puting models. Although previous studies and packages in the field of hete
 rogeneous computing have focused on GPUs as accelerators, FPGAs are an ext
 remely promising option as well. A series of workflows are developed to es
 tablish the performance capabilities of FPGAs as a service. Multiple diffe
 rent devices and a range of algorithms for use in high energy physics are 
 studied. For a small, dense network, the throughput can be improved by an 
 order of magnitude with respect to GPUs as a service. For large convolutio
 nal networks, the throughput is found to be comparable to GPUs as a servic
 e. This work represents the first open-source FPGAs-as-a-service toolkit.\
 n\nTag: Accelerators, FPGA, and GPUs, Architectures, Emerging Technologies
 , Heterogeneous Systems, Reconfigurable Computing\n\nRegistration Category
 : Workshop Reg Pass
END:VEVENT
END:VCALENDAR