BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:SABER: Hybrid Data Processing with Heterogeneous Servers
DTSTART:20170523T151500
DTSTAMP:20260407T051042Z
UID:5fc4928c0c6bf953ff5b6f19b1f5bd74f6e1b0b88f4c4a4d95cc5abb
CATEGORIES:Conferences - Seminars
DESCRIPTION:Peter Pietzuch is an Associate Professor (Reader) at Imperial 
 College London\, where he leads the Large-scale Distributed Systems (LSDS)
  group (http://lsds.doc.ic.ac.uk) in the Department of Computing. His rese
 arch focuses on the design and engineering of scalable\, reliable and secu
 re large-scale software systems\, with a particular interest in performanc
 e\, data management and networking issues. He has published papers in prem
 ier international venues\, including SIGMOD\, VLDB\, ICDE\, OSDI\, USENIX 
 ATC\, SoCC\, ICDCS\, CCS\, CoNEXT\, NSDI\, and Middleware. Before joining 
 Imperial College London\, he was a post-doctoral fellow at Harvard Univers
 ity. He holds PhD and MA degrees from the University of Cambridge.\nModern
  servers in data centres have become increasingly heterogeneous\, e.g. com
 bining multi-core CPUs with many-core GPUs. This has implications on the d
 esign of future data-intensive systems for stream processing or machine le
 arning: first\, systems must exploit all the available parallelism of the 
 hardware\, independently of the processing semantics\; and\, second\, inst
 ead of offloading computation entirely to an accelerator\, systems must fu
 lly utilise all heterogeneous processors in a server\, thus making acceler
 ators first-class compute elements.\n\n \n\nIn this talk\, I will describ
 e SABER\, a new hybrid stream processing engine for CPUs and GPUs. Under a
  hybrid execution model\, SABER executes streaming SQL queries in a data-p
 arallel fashion on all available CPUs and GPUs simultaneously. Instead of 
 statically assigning query tasks to heterogeneous processors\, SABER adapt
 ively schedules computation on the best available processor. It parallelis
 es stream queries in a way that suits the properties of the hardware\, ind
 ependently of the window-based query semantics. Our experiments show how S
 ABER's hybrid execution model can aggregate the performance of multiple he
 terogeneous processors in a server.\n\n \n\nThis talk is based on a SIGMO
 D'16 paper.
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR