BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Heterogeneous Execution of Dataflow Programs
DTSTART:20190620T103000
DTEND:20190620T123000
DTSTAMP:20260407T055852Z
UID:decc3341c24d5796057ef06660e833bf47a59468839d6a925ca84cd6
CATEGORIES:Conferences - Seminars
DESCRIPTION:Seyedmahyar Emami\nEDIC candidacy exam\nExam president: Prof. 
 Babak Falsafi\nThesis advisor: Prof. James Larus\nCo-examiner: Prof. Paolo
  Ienne\n\nAbstract\nEconomies of scale have shifted computation to the War
 ehouse-Scale Computers (WSCs) in the cloud. The popularity of certain heav
 yweight workloads (e.g. machine learning) pushed the cloud providers to in
 tegrate accelerators\, such as GPUs and FPGAs\, to their infrastructure.\n
 The cloud is now comprised of heterogeneous WSCs\, combining processing el
 ements of different architectural characteristics. Exploiting the benefits
  of a heterogeneous WSC is challenging and requires new programming models
  to efficiently utilize the underlying resources at scale.\nDataflow is a 
 promising model which aids the development of large-scale programs\, makin
 g it easy for programmers to handle concurrency and to utilize the existin
 g optimized single-machine programs.\n\n\nBackground papers\nTensorFlow: A
  System for Large-Scale Machine Learning\, by Abadi\, M.\, et al.\nA cloud
 -scale acceleration architecture\, by Caulfield\, A. M.\, et al.\nXilinx A
 daptive Compute Acceleration Platform: Versal Architecture\, by  Gaide\, 
 B.\, et al.\n\n 
LOCATION:BC 229 https://plan.epfl.ch/?room==BC%20229
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR