BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Hardware Acceleration of Recursive Application
DTSTART:20190826T140000
DTEND:20190826T160000
DTSTAMP:20260406T211541Z
UID:e12c3b607764e4a3415847e90865876cbab3ed6fd8b02f1f8daea299
CATEGORIES:Conferences - Seminars
DESCRIPTION:Jovan Blanusa\nEDIC candidacy exam\nExam president: Prof. Baba
 k Falsafi\nThesis advisor: Prof. Paolo Ienne\nThesis co-advisor: Dr. Kubil
 ay Atasu\nCo-examiner: Prof. Wenzel Jakob\n\nAbstract\nMany data- and grap
 h-mining problems rely on recursive search and exploration. Examples of su
 ch problems include finding communities in social networks (e.g. cliques 
 and clusters) and detecting patterns in graphs in the form of a subgraph w
 ith a specific structure. Due to the irregularity of the underlying data 
 structures and the parallelism that arises dynamically\, it is challenging
  to parallelize these workloads. The search tree constructed during the re
 cursion can be very unbalanced\, which can lead to uneven distribution of 
 the workload across different threads. Scheduling\, synchronization\, and 
 random memory accesses are additional challenges. The goal of this project
  is to discover the limits of scaling recursive algorithms on existing mul
 ti-core platforms and to propose more efficient and scalable architecture
 s that can address the existing limitations. In this proposal\, we discuss
  three papers that serve as the foundation for this project. First\, we in
 troduce a state-of-theart algorithm that performs a recursive search and c
 omputes all the maximal cliques of a given graph. Next\, we discuss a pape
 r which introduces a new approach for parallelizing ”irregular” algori
 thms\, by exposing data parallelism that arises dynamically\, and presents
  a graph processing framework which uses this approach. Finally\, we exami
 ne a specific FPGA implementation of a graph processing accelerator\, whi
 ch uses hardware transactional memory\, and discuss the way the authors ta
 ckled some problems specific to the parallelization of graph algorithms.\
 n\nBackground papers\nListing All Maximal Cliques in Sparse Graphs in Near
 -optimal Time\, by Eppstein et al.\nThe Tao of Parallelism in Algorithms 
  by Pingali et al.\nFPGA-Accelerated Transactional Execution of Graph Work
 loads\, by Ma et al. 
LOCATION:INF 113 https://plan.epfl.ch/?room==INF%20113
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR