1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61
|
<!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN">
<!--Converted with LaTeX2HTML 96.1-h (September 30, 1996) by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds -->
<HTML>
<HEAD>
<TITLE>Solving Linear Systems of Equations</TITLE>
<META NAME="description" CONTENT="Solving Linear Systems of Equations">
<META NAME="keywords" CONTENT="slug">
<META NAME="resource-type" CONTENT="document">
<META NAME="distribution" CONTENT="global">
<LINK REL=STYLESHEET HREF="slug.css">
</HEAD>
<BODY LANG="EN" >
<A NAME="tex2html3671" HREF="node118.html"><IMG WIDTH=37 HEIGHT=24 ALIGN=BOTTOM ALT="next" SRC="http://www.netlib.org/utk/icons/next_motif.gif"></A> <A NAME="tex2html3669" HREF="node108.html"><IMG WIDTH=26 HEIGHT=24 ALIGN=BOTTOM ALT="up" SRC="http://www.netlib.org/utk/icons/up_motif.gif"></A> <A NAME="tex2html3665" HREF="node116.html"><IMG WIDTH=63 HEIGHT=24 ALIGN=BOTTOM ALT="previous" SRC="http://www.netlib.org/utk/icons/previous_motif.gif"></A> <A NAME="tex2html3673" HREF="node1.html"><IMG WIDTH=65 HEIGHT=24 ALIGN=BOTTOM ALT="contents" SRC="http://www.netlib.org/utk/icons/contents_motif.gif"></A> <A NAME="tex2html3674" HREF="node190.html"><IMG WIDTH=43 HEIGHT=24 ALIGN=BOTTOM ALT="index" SRC="http://www.netlib.org/utk/icons/index_motif.gif"></A> <BR>
<B> Next:</B> <A NAME="tex2html3672" HREF="node118.html">Solving Linear Least Squares </A>
<B>Up:</B> <A NAME="tex2html3670" HREF="node108.html">PerformancePortability and Scalability</A>
<B> Previous:</B> <A NAME="tex2html3666" HREF="node116.html">Solution of Common Numerical </A>
<BR> <P>
<H2><A NAME="SECTION04526000000000000000">Solving Linear Systems of Equations</A></H2>
<A NAME="subseclin"> </A>
<P>
Table <A HREF="node117.html#tabluperf">5.9</A><A NAME="3938"> </A>
illustrates the speed of the ScaLAPACK driver
routine PSGESV<A NAME="3939"> </A>/PDGESV<A NAME="3940"> </A>
for solving a square linear system of order <I>N</I> by
LU factorization with partial row pivoting of a real
matrix. For all timings, 64-bit floating-point arithmetic was used.
Thus, single-precision timings are reported for
the Cray T3E, and double precision timings are reported
on all other computers. The distribution block size<A NAME="3941"> </A> is
also used as the partitioning unit<A NAME="3942"> </A> for the computation and
communication phases.
<P>
Table <A HREF="node117.html#tablltperf">5.10</A><A NAME="3944"> </A>
illustrates the speed of the ScaLAPACK routine
PSPOSV<A NAME="3945"> </A>/PDPOSV<A NAME="3946"> </A> for
solving a symmetric positive definite linear system of
order <I>N</I> via the Cholesky factorization.
<P>
Right-looking variants of the <I>LU</I> and Cholesky factorizations
were chosen for ScaLAPACK because they minimize total communication
volume, that is, the aggregated
<P><A NAME="3948"> </A><A NAME="tabluperf"> </A><IMG WIDTH=747 HEIGHT=728 ALIGN=BOTTOM ALT="table3947" SRC="img396.gif"><BR>
<STRONG>Table 5.9:</STRONG> Speed in Mflop/s of PSGESV/PDGESV for square
matrices of order <I>N</I><BR>
<P>
amount of data transferred between processes during the operation.
<P><A NAME="3970"> </A><A NAME="tablltperf"> </A><IMG WIDTH=747 HEIGHT=605 ALIGN=BOTTOM ALT="table3969" SRC="img397.gif"><BR>
<STRONG>Table 5.10:</STRONG> Speed in Mflop/s of PSPOSV/PDPOSV for matrices of
order <I>N</I> with UPLO=`U'<BR>
<P><BR> <HR>
<UL><A NAME="CHILD_LINKS"> </A>
<LI> <A NAME="tex2html3675" HREF="node118.html#SECTION04526100000000000000">Solving Linear Least Squares Problems</A>
<LI> <A NAME="tex2html3676" HREF="node119.html#SECTION04526200000000000000">Eigenvalue Problems</A>
</UL>
<BR> <HR>
<P><ADDRESS>
<I>Susan Blackford <BR>
Tue May 13 09:21:01 EDT 1997</I>
</ADDRESS>
</BODY>
</HTML>
|