BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Transparent Fault Tolerance for Scalable Functional Computation
DTSTART:20160726T120000
DTEND:20160726T130000
DTSTAMP:20260307T131211Z
UID:bca229883c901ab007c611c1d0b20fa1325e0f79ea184cee92b77fe7
CATEGORIES:Conferences - Seminars
DESCRIPTION:Rob Stewart (Heriot-Watt University\, Edinburgh)\nAbstract:\nR
 eliability is set to become a major concern on emergent large-scale archit
 ectures. While there are many parallel languages\, and indeed many paralle
 l functional languages\, very few address reliability. We investigate scal
 able transparent fault tolerance with automatic supervision and recovery o
 f tasks with HdpH-RS\, a DSL for fork/join parallelism on HPC architecture
 s. Stateless functions are key for proving a crucial property of the seman
 tics of HdpH-RS: fault recovery does not change the result of the program\
 , akin to deterministic parallelism. To eliminate elusive concurrency bugs
 \, HdpH-RS's work stealing protocol has been validated using the SPIN mode
 l checker.\nHdpH-RS has been benchmarked on conventional clusters and an H
 PC platform: all benchmarks survive Chaos Monkey random fault injection\; 
 the system scales well e.g. up to 1400 cores on the HPC\; reliability and 
 recovery overheads are consistently low even at scale.Bio:\nRob Stewart is
  a postdoc at Heriot-Watt University in Edinburgh. His research interests 
 cover parallel functional programming language design and implementation\,
  program transformations\, and embedded systems. He has previously develop
 ed Haskell libraries for fault tolerant distributed computing\, including 
 6 months using CloudHaskell in a start-up company. He is currently develop
 ing a parallel image processing DSL for FPGAs\, along with an IDE based tr
 ansformations toolkit for refactoring dataflow abstractions of FPGA circui
 ts to increase throughput performance.
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR