Entry Li:1992:ANL from sigplan1990.bib

Last update: Thu Apr 12 03:37:15 MDT 2012                Valid HTML 3.2!

Index sections

Top | Symbols | Numbers | Math | A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z

BibTeX entry

@Article{Li:1992:ANL,
  author =       "Wei Li and Keshav Pingali",
  title =        "Access normalization: loop restructuring for {NUMA}
                 compilers",
  journal =      j-SIGPLAN,
  volume =       "27",
  number =       "9",
  pages =        "285--295",
  month =        sep,
  year =         "1992",
  CODEN =        "SINODQ",
  ISSN =         "0362-1340 (print), 1523-2867 (print), 1558-1160 (electronic)",
  ISSN-L =       "0362-1340",
  bibdate =      "Sun Dec 14 09:16:26 MST 2003",
  bibsource =    "http://portal.acm.org/; http://www.acm.org/pubs/toc/",
  URL =          "http://www.acm.org:80/pubs/citations/proceedings/asplos/143365/p285-li/",
  abstract =     "In scalable parallel machines, processors can make
                 local memory accesses much faster than they can make
                 remote memory accesses. In addition, when a number of
                 remote accesses must be made, it is usually more
                 efficient to use block transfers of data rather than to
                 use many small messages. To run well on such machines,
                 software must exploit these features. We believe it is
                 too onerous for a programmer to do this by hand, so we
                 have been exploring the use of restructuring compiler
                 technology for this purpose. In this paper, we start
                 with a language like FORTRAN-D with user-specified data
                 distribution and develop a systematic loop
                 transformation strategy called {\em access
                 normalization\/} that restructures loop nests to
                 exploit locality and block transfers. We demonstrate
                 the power of our techniques using routines from the
                 BLAS (Basic Linear Algebra Subprograms) library. An
                 important feature of our approach is that we model loop
                 transformations using {\em invertible\/} matrices and
                 integer lattice theory, thereby generalizing Banerjee's
                 framework of unimodular matrices [5].",
  acknowledgement = ack-nhfb,
  keywords =     "design; languages; measurement; performance",
  subject =      "{\bf D.3.4} Software, PROGRAMMING LANGUAGES,
                 Processors, Compilers. {\bf C.1.2} Computer Systems
                 Organization, PROCESSOR ARCHITECTURES, Multiple Data
                 Stream Architectures (Multiprocessors), Parallel
                 processors**. {\bf D.4.2} Software, OPERATING SYSTEMS,
                 Storage Management.",
}

Related entries