Newsgroups: comp.parallel From: "Robert D. Blumofe" Subject: Re: object oriented parallel processing Organization: The University of Texas at Austin Date: 24 Jun 1998 18:04:22 GMT Message-ID: <6mrf36$qcr$1@encore.ece.cmu.edu> Michael W. Bigrigg wrote: > So a few years back I asked the group if anyone was > using C++ for parallel processing. So I'll ask it > again to see how things have changed. I have been working on a C++ user-level threads library for multiprogrammed multiprocessors. It is called "Hood"; we should be doing our first release within the next few weeks; and this first release will run on SPARC/Solaris SMPs. Hood's distinguishing feature is that it uses a scheduling algorithm --- the non-blocking work stealer --- that is provably efficient, even in multiprogrammed environments (without any need for coscheduling or process control). A description and algorithmic analysis of the non-blocking work stealer is in a paper to appear at SPAA later this month. An empirical evaluation is in a tech report. Both papers are on my web page: www.cs.utexas.edu/users/rdb. Among the applications that we have written in C++ using Hood are dense matrix codes (matrix multiply and LU decomposition), sorting (a parallelization of merge sort), barnes-hut, and ray tracing. -- Bobby -- Articles to bigrigg+parallel@cs.cmu.edu (Admin: bigrigg@cs.cmu.edu) Archive: http://www.hensa.ac.uk/parallel/internet/usenet/comp.parallel