BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Concurrent Algorithms for Emerging Hardware Platforms - Irina Calc
 iu\, Brown University 
DTSTART:20150212T090000Z
DTEND:20150212T100000Z
UID:TALK57938@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Computer architecture has recently seen an explosion of innova
 tion that has enabled more parallel execution. The number of computing cor
 es used in every area of the software ecosystem continues to increase\, an
 d parallelism within programs is now ubiquitous. Ideally\, performance wou
 ld scale linearly with the number of cores\, but in practice that is rarel
 y the case. Communication and synchronization between cores running the sa
 me application are often necessary\, but usually come at a high cost. This
  results in negative scalability and a significant drop in performance. In
  this context\, parallel software needs to provide more simplified program
 ming patterns and tools that enable a higher potential for parallelism wit
 hout increasing the burden on the programmer. \nIn this talk\, I will disc
 uss new techniques to simplify writing efficient parallel code that levera
 ge novel architectural features from many current systems. First\, I will 
 describe programming abstractions\, such as delegation\, elimination\, com
 bining and transactional memory\, which improve scalability and performanc
 e of concurrent programs. I will then present a new concurrent priority qu
 eue design that is based on these abstractions and is significantly faster
  than prior implementations. Finally\, I will describe Invyswell\, a new h
 ybrid transactional memory that combines Intel’s Restricted Transactiona
 l Memory capability with a software transactional memory to guarantee effi
 cient execution and forward progress of transactional programs. Developers
  can use all of these abstractions as building blocks to create efficient 
 code that is able to scale on very diverse platforms\, without requiring s
 pecialized knowledge of parallel programming.
LOCATION:Indigo Meeting room\, Microsoft Research Ltd\, 21 Station Road\, 
 Cambridge\, CB1 2FB
END:VEVENT
END:VCALENDAR