http:^^www.netlib.org^scalapack^
来自「This data set contains WWW-pages collect」· ORG^SCALAPACK^ 代码 · 共 370 行 · 第 1/2 页
ORG^SCALAPACK^
370 行
Date: Wed, 20 Nov 1996 19:18:57 GMT
Server: Apache/1.1.1
Content-type: text/html
Content-length: 15247
Last-modified: Wed, 20 Nov 1996 10:04:38 GMT
<html><head><TITLE>ScaLAPACK</TITLE><meta name="BIDM.asset.name" content="ScaLAPACK"><meta name="BIDM.asset.abstract" content="ScaLAPACK is a library of high performance linear algebraroutines for distributed memory MIMD computers. It is acontinuation of the LAPACK project, which designed andproduced analogous software for workstations, vectorsupercomputers, and shared memory parallel computers. Bothlibraries contain routines for solving systems of linearequations, least squares problems, and eigenvalue problems.Most machine dependencies are limited to two standardlibraries called the BLAS, or Parallel Basic Linear Algebra Subroutines,and the BLACS, or Basic Linear Algebra Communication Subroutines.LAPACK and ScaLAPACK will run on any machine where the PBLASand the BLACS are available."><meta name="BIDM.asset.cost" content="free"><meta name="BIDM.asset.targetenvironment" content="PVM 3.3 or later; Intel distributedmemory computers; IBM SP series; Thinking Machines CM-5."><meta name="BIDM.asset.domain" content="numerical/linear algebra"><meta name="NHSE.asset.contactis.organization.email" content="scalapack@cs.utk.edu"><meta name="BIDM.asset.keyword" content="GAMS D. Linear algebra"><meta name="BIDM.asset.keyword" content="parallel numerical library;distributed memory multiprocessor; MIMD machine"><link rel="NHSE.asset.reference" href="http://www.netlib.org/lapack/lawns/lawn100.ps"><link rel="BIDM.asset.islocatedin.library" title="Netlib" href="http://www.netlib.org/" urn="urn:inet:netlib.org:netlib"><link rel="BIDM.asset.webpage" href="http://www.netlib.org/scalapack/"></head><body><h1>ScaLAPACK Project</h1>There have been<!-- Don't edit this line--> <!WA0><A HREF="http://www.netlib.org/netlib/master_counts.html">21,579</A> accesses to this library. (Count updated 11/13/96 at 02:12:44)<p>The <!WA1><a href=http://www.netlib.org/scalapack/darpa-progress-report.html><B>DARPA/ITO Progress Report</B></a> for the project ``A Scalable Library For Numerical Linear Algebra'' for fiscal year 1995/1996 isavailable for viewing.</p><hr><!WA2><A HREF="http://www.netlib.org/scalapack/poster.html"><!WA3><IMG border=1 SRC="http://www.netlib.org/scalapack/slap_thumbnail.gif" align="left">Click here to see <br>theSuperComputing'96 <br>ScaLAPACK poster</A><br><br><br><br><br><br><hr><p>The ScaLAPACK project is a collaborative effort between:</p><ul><li>Oak Ridge National Laboratory<li>Rice University<li>University of California, Berkeley<li>University of California, Los Angeles<li>University of Illinois<li>University of Tennessee, Knoxville</ul><p>The ScaLAPACK project is made up of 4 components:</p><ul><li>dense matrix software (<!WA4><a href="http://www.netlib.org/scalapack/scalapack_home.html"><B>ScaLAPACK</B></a>)<li>large sparse eigenvalue software (<!WA5><a href="http://www.caam.rice.edu/~kristyn/parpack_home.html"><B>PARPACK</B></a>)<li>sparse direct systems software (<B>CAPSS</B>)<li>preconditioners for large sparse iterative solvers (<!WA6><a href="http://www.math.ucla.edu/~eijkhout/parpre.html"><B>PARPRE</B></a>)</ul><p><!WA7><a href="http://www.netlib.org/scalapack/scalapack_home.html"><B>ScaLAPACK</B></a>,version 1.4, includes routines for the solution of dense,band, and tridiagonal linear systems of equations, condition estimation anditerative refinement, for LU and Cholesky factorization, matrix inversion,full-rank linear least squares problems, orthogonal and generalized orthogonalfactorizations, orthogonal transformation routines, reductions to upperHessenberg, bidiagonal and tridiagonal form, reduction of a symmetric-definite/Hermitian-definite generalized eigenproblem to standard form, thesymmetric/Hermitian, generalized symmetric/Hermitian, and the nonsymmetriceigenproblem.</p><p>Software is available in single precision real, double precision real,single precision complex, and double precision complex. The software hasbeen written to be portable across a wide range of distributed-memoryenvironments such as the Cray T3, IBM SP, Intel series, TM CM-5,clusters of workstations, and any system for which <!WA8><a href="http://www.netlib.org/pvm3/index.html">PVM</a>or <!WA9><a href="http://www.mcs.anl.gov/mpi/">MPI</a> is available.</p><p>A draft of the<!WA10><a href="http://www.netlib.org/scalapack/scalapack_ug.ps">ScaLAPACK Users' Guide</a> and a comprehensive<!WA11><a href="http://www.netlib.org/scalapack/scalapack_install.ps">Installation Guide</a> isprovided, as well as test suites for all <B>ScaLAPACK</B>,<!WA12><a href="http://www.netlib.org/scalapack/html/pblas_qref.html"><B>PBLAS</B></a>, and<!WA13><a href="http://www.netlib.org/blacs/index.html"><B>BLACS</B></a> routines.</p><p>Future releases of<!WA14><a href="http://www.netlib.org/scalapack/scalapack_home.html"><B> ScaLAPACK</B></a> will include routines for the solution ofgeneral banded linear systems, general and symmetric positive definitetridiagonal systems, rank-deficient linear least squares problems,generalized linear least squares problems, and the singular valuedecomposition. Also available will be the full-index <B>PBLAS</B>, which willhave the majority of alignment restrictions removed, as well as theability to operate on partial first blocks. The next release of<B>ScaLAPACK</B> is slated for Fall, 1996.</p><p><!WA15><a href="http://www.caam.rice.edu/~kristyn/parpack_home.html"><B>PARPACK</B></a> (Parallel ARPACK) is an extension of the <B>ARPACK</B> softwarepackage used for solving large scale eigenvalue problems on distributed memoryparallel architectures. The message passing layers currently supportedare <!WA16><a href="http://www.netlib.org/blacs/index.html">BLACS</a>and <!WA17><a href="http://www.mcs.anl.gov/mpi/">MPI</a>. <I>Serial ARPACK must beretrieved and installed prior to installing PARPACK</I>. All core<B>ARPACK</B> routinesare available in single precision real, double precision real, singleprecision complex, and double precision complex. An extensive set of driverroutines are available for <B>ARPACK</B> and a subset of these are availablefor parallel computation with <B>PARPACK</B>. These may be used as templatesthat are easily modified to construct a problem specific parallel interface to<B>PARPACK</B>.</p><p><B>CAPSS</B> is a fully parallel package to solve a sparse linear system ofthe form Ax=b on a message passing multiprocessor; the matrix A is assumed tobe symmetric positive definite and associated with a mesh in two or threedimensions. This version has been tested on the Intel Paragon and makespossible efficient parallel solution for several right hand side vectors.</p><p><!WA18><a href="http://www.math.ucla.edu/~eijkhout/parpre.html"><B>PARPRE</B></a> isa package of parallel preconditioners for general sparsematrices. It includes classical point/block relaxation methods,generalised block SSOR preconditioners (this includes ILU), and domaindecomposition methods (additive and multiplicative Schwarz, Schurcomplement). The communication protocol is<!WA19><a href="http://www.mcs.anl.gov/mpi/">MPI</a>, and low levelroutines from the <B>Petsc</B> library are used, but installing the complete<B>Petsc</B> library is not necessary.</p><p>Funding for this effort comes in part from DARPA, DOE, NSF, and CRPC.</p><p>Comments/suggestions may be sent to <address>scalapack@cs.utk.edu</address>.</p><p><hr><pre># SCALAPACK directory# =========# The ScaLAPACK project is a collaborative effort between:# Oak Ridge National Laboratory Rice University# Univ. of Tenn., Knoxville Univ. of Calif., Berkeley# Univ. of Calif., Los Angeles Univ. of Illinois, Urbana-Champaign# The ScaLAPACK project is made up of 4 components:# dense matrix software (ScaLAPACK)# large sparse eigenvalue software (PARPACK)# sparse direct systems software (CAPSS)# preconditioners for large sparse iterative solvers (PARPRE)# This directory contains a number of different software packages# involved in the ScaLAPACK project.# *********** # PLEASE NOTE: As of November, 1996, this directory was updated.# *********** PROTOTYPE codes are in scalapack/prototype directory.# Pre-built ScaLAPACK libraries are now available!* ## The URL for the BLAS is http://www.netlib.org/blas/index.html.# The URL for the BLACS is http://www.netlib.org/blacs/index.html.# The URL for PVM is http://www.netlib.org/pvm3/index.html.# A URL for MPI is http://www.mcs.anl.gov/mpi/.# The ARPACK/PARPACK files may also be obtained by anonymous ftp# from ftp.caam.rice.edu.## Questions/comments should be directed to scalapack@cs.utk.edu.⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?