Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Appears in First USENIX Workshop on Hot Topics in Parallelism (HotPar-09), Berkeley, CA, March 2009 Lithe: Enabling Efficient Composition of Parallel Libraries
 

Summary: Appears in First USENIX Workshop on Hot Topics in Parallelism (HotPar-09), Berkeley, CA, March 2009
Lithe: Enabling Efficient Composition of Parallel Libraries
Heidi Pan
Massachusetts Institute of Technology
xoxo@csail.mit.edu
Benjamin Hindman Krste Asanovi┤c
University of California, Berkeley
{benh,krste}@eecs.berkeley.edu
Abstract
For the software industry to take advantage of multi-
core processors, we must allow programmers to arbi-
trarily compose parallel libraries without sacrificing per-
formance. We argue that high-level task or thread ab-
stractions and a common global scheduler cannot pro-
vide effective library composition. Instead, the operat-
ing system should expose unvirtualized processing re-
sources that can be shared cooperatively between parallel
libraries within an application. In this paper, we describe
a system that standardizes and facilitates the exchange
of these unvirtualized processing resources between li-

  

Source: AsanoviŠ, Krste - Computer Science and Artificial Intelligence Laboratory & Department of Electrical Engineering and Computer Science, Massachusetts Institute of Technology (MIT)

 

Collections: Computer Technologies and Information Sciences