BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Incremental Parallel and Distributed Systems - Pramod Bhatotia\, M
 PI-SWS
DTSTART:20140328T100000Z
DTEND:20140328T110000Z
UID:TALK51689@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Many applications exhibit the property that small\, localized 
 changes to the input result in only minor changes to the output. Increment
 al computations seek to exploit this observation to speed-up successive ru
 ns of an application by tracking control and data dependencies\, and re-ex
 ecuting only those parts of the computation that are affected by a given i
 nput change. To realize the benefits of incremental computation\, research
 ers and practitioners are developing new systems where the programmer can 
 provide efficient update mechanisms for changing application data. Unfortu
 nately\, most of the existing solutions are limiting because they either d
 epart from existing programming models\, or require programmers to devise 
 an incremental update mechanism (or a dynamic algorithm) on a per-applicat
 ion basis.\n\nThe high-level goal of my thesis research is to enable pract
 ical\, automatic\, and efficient incremental computations in parallel and 
 distributed systems. My approach neither requires a radical departure from
  current models of programming nor complex\, application-specific dynamic 
 algorithms.\n\nIn this talk\, I will first present a high-level descriptio
 n of my thesis research. Thereafter\, as a concrete example I will describ
 e my recent project “iThreads”. iThreads is a POSIX-compliant threadin
 g library to support parallel incremental computation targeting unmodified
  C/C++ multithreaded programs. The iThreads library can be used as a drop-
 in replacement for the pthreads library\, making it trivial to obtain the 
 benefits of incremental computation by a simple exchange of libraries link
 ed\, without even recompiling the application code. To achieve this result
 \, we design our algorithms and implementation to operate at the compiled 
 binary code level by leveraging operating system-specific mechanisms. Our 
 design choice of iThreads tries to strike a balance between efficiency of 
 incremental computation and transparency for unmodified C/C++ pthreads-bas
 ed programs. Our evaluation on a multicore platform using benchmarks from 
 the PARSEC and Phoenix suites shows significant performance improvements f
 or the majority of applications.
LOCATION:Small Lecture Theatre\, Microsoft Research Ltd\, 21 Station Road\
 , Cambridge\, CB1 2FB
END:VEVENT
END:VCALENDAR