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 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
<!--Converted with LaTeX2HTML 98.2 beta6 (August 14th, 1998)
original version by: Nikos Drakos, CBLU, University of Leeds
* revised and updated by: Marcus Hennecke, Ross Moore, Herb Swan
* with significant contributions from:
Jens Lippmann, Marek Rouchal, Martin Wilck and others -->
<HTML>
<HEAD>
<TITLE>Error Bounds for Fast Level 3 BLAS</TITLE>
<META NAME="description" CONTENT="Error Bounds for Fast Level 3 BLAS">
<META NAME="keywords" CONTENT="lug_l2h">
<META NAME="resource-type" CONTENT="document">
<META NAME="distribution" CONTENT="global">
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<LINK REL="STYLESHEET" HREF="lug_l2h.css">
<LINK REL="previous" HREF="node106.html">
<LINK REL="up" HREF="node72.html">
<LINK REL="next" HREF="node109.html">
</HEAD>
<BODY >
<!--Navigation Panel-->
<A NAME="tex2html5705"
HREF="node109.html">
<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
SRC="next_motif.png"></A>
<A NAME="tex2html5699"
HREF="node72.html">
<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
SRC="up_motif.png"></A>
<A NAME="tex2html5695"
HREF="node107.html">
<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
SRC="previous_motif.png"></A>
<A NAME="tex2html5701"
HREF="node1.html">
<IMG WIDTH="65" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="contents"
SRC="contents_motif.png"></A>
<A NAME="tex2html5703"
HREF="node152.html">
<IMG WIDTH="43" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="index"
SRC="index_motif.png"></A>
<BR>
<B> Next:</B> <A NAME="tex2html5706"
HREF="node109.html">Documentation and Software Conventions</A>
<B> Up:</B> <A NAME="tex2html5700"
HREF="node72.html">Accuracy and Stability</A>
<B> Previous:</B> <A NAME="tex2html5696"
HREF="node107.html">Further Details: Error Bounds</A>
  <B> <A NAME="tex2html5702"
HREF="node1.html">Contents</A></B>
  <B> <A NAME="tex2html5704"
HREF="node152.html">Index</A></B>
<BR>
<BR>
<!--End of Navigation Panel-->
<H1><A NAME="SECTION034130000000000000000"></A><A NAME="secfastblas"></A>
<BR>
Error Bounds for Fast Level 3 BLAS
</H1>
<P>
The Level 3 BLAS specifications [<A
HREF="node151.html#blas3">40</A>] specify the input, output
and calling sequence for each routine, but allow freedom of
implementation, subject to the requirement that the routines be
numerically stable<A NAME="13071"></A>.
Level 3 BLAS implementations can therefore be
built using matrix multiplication algorithms that achieve a more
favorable operation count (for suitable dimensions) than the standard
multiplication technique, provided that these ``fast'' algorithms are
numerically stable. The simplest fast matrix multiplication
technique is Strassen's
method<A NAME="13072"></A><A NAME="13073"></A>, which can
multiply two <B><I>n</I></B>-by-<B><I>n</I></B>
matrices in fewer than
<!-- MATH
$4.7 n^{\log_2 7}$
-->
<IMG
WIDTH="71" HEIGHT="19" ALIGN="BOTTOM" BORDER="0"
SRC="img908.png"
ALT="$4.7 n^{\log_2 7}$">
operations, where
<!-- MATH
$\log_2 7 \approx 2.807$
-->
<IMG
WIDTH="109" HEIGHT="32" ALIGN="MIDDLE" BORDER="0"
SRC="img909.png"
ALT="$\log_2 7 \approx 2.807$">.
<P>
The effect on the results in this chapter of using a fast Level 3 BLAS
implementation can be explained as follows. In general, reasonably
implemented fast Level 3 BLAS preserve all the bounds presented here
(except those at the end of subsection <A HREF="node98.html#secgendef">4.10</A>), but the constant
<B><I>p</I>(<I>n</I>)</B> may increase somewhat. Also, the iterative refinement
routine<A NAME="13076"></A>
xyyRFS may take more steps to converge.
<P>
This is what we mean by reasonably implemented fast Level 3 BLAS.
Here, <B><I>c</I><SUB><I>i</I></SUB></B> denotes a constant depending on the specified matrix dimensions.
<P>
(1) If <B><I>A</I></B> is <B><I>m</I></B>-by-<B><I>n</I></B>, <B><I>B</I></B> is <B><I>n</I></B>-by-<B><I>p</I></B> and <IMG
WIDTH="18" HEIGHT="21" ALIGN="BOTTOM" BORDER="0"
SRC="img910.png"
ALT="$\widehat C$">
is the computed
approximation to <B><I>C</I>=<I>AB</I></B>, then
<BR><P></P>
<DIV ALIGN="CENTER">
<!-- MATH
\begin{displaymath}
\| \widehat C- AB \|_{\infty} \le c_1(m,n,p) \epsilon \|A\|_{\infty}\|B\|_{\infty} + O(\epsilon ^2).
\end{displaymath}
-->
<IMG
WIDTH="359" HEIGHT="31" BORDER="0"
SRC="img911.png"
ALT="\begin{displaymath}
\Vert \widehat C- AB \Vert _{\infty} \le c_1(m,n,p) \epsilon \Vert A\Vert _{\infty}\Vert B\Vert _{\infty} + O(\epsilon ^2).
\end{displaymath}">
</DIV>
<BR CLEAR="ALL">
<P></P>
<P>
(2)
The computed solution <IMG
WIDTH="20" HEIGHT="20" ALIGN="BOTTOM" BORDER="0"
SRC="img797.png"
ALT="$\widehat{X}$">
to the triangular systems <B><I>TX</I>=<I>B</I></B>,
where <B><I>T</I></B> is <B><I>m</I></B>-by-<B><I>m</I></B> and <B><I>B</I></B> is <B><I>m</I></B>-by-<B><I>p</I></B>, satisfies
<BR><P></P>
<DIV ALIGN="CENTER">
<!-- MATH
\begin{displaymath}
\| T \widehat X- B \|_{\infty} \le c_2(m,p) \epsilon \|T\|_{\infty} \|\widehat X\|_{\infty}
+ O(\epsilon ^2).
\end{displaymath}
-->
<IMG
WIDTH="344" HEIGHT="31" BORDER="0"
SRC="img912.png"
ALT="\begin{displaymath}
\Vert T \widehat X- B \Vert _{\infty} \le c_2(m,p) \epsilon...
...rt _{\infty} \Vert\widehat X\Vert _{\infty}
+ O(\epsilon ^2).
\end{displaymath}">
</DIV>
<BR CLEAR="ALL">
<P></P>
For conventional Level 3 BLAS implementations these conditions
hold with
<!-- MATH
$c_1(m,n,p) = n^2$
-->
<B><I>c</I><SUB>1</SUB>(<I>m</I>,<I>n</I>,<I>p</I>) = <I>n</I><SUP>2</SUP></B> and
<!-- MATH
$c_2(m,p)= m(m+1)$
-->
<B><I>c</I><SUB>2</SUB>(<I>m</I>,<I>p</I>)= <I>m</I>(<I>m</I>+1)</B>.
Strassen's method<A NAME="13083"></A> satisfies these
bounds for slightly larger <B><I>c</I><SUB>1</SUB></B> and <B><I>c</I><SUB>2</SUB></B>.
<P>
For further details, and references to fast multiplication techniques,
see [<A
HREF="node151.html#Demmel-Higham-Wnote22">27</A>].
<P>
<HR>
<!--Navigation Panel-->
<A NAME="tex2html5705"
HREF="node109.html">
<IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next"
SRC="next_motif.png"></A>
<A NAME="tex2html5699"
HREF="node72.html">
<IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up"
SRC="up_motif.png"></A>
<A NAME="tex2html5695"
HREF="node107.html">
<IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous"
SRC="previous_motif.png"></A>
<A NAME="tex2html5701"
HREF="node1.html">
<IMG WIDTH="65" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="contents"
SRC="contents_motif.png"></A>
<A NAME="tex2html5703"
HREF="node152.html">
<IMG WIDTH="43" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="index"
SRC="index_motif.png"></A>
<BR>
<B> Next:</B> <A NAME="tex2html5706"
HREF="node109.html">Documentation and Software Conventions</A>
<B> Up:</B> <A NAME="tex2html5700"
HREF="node72.html">Accuracy and Stability</A>
<B> Previous:</B> <A NAME="tex2html5696"
HREF="node107.html">Further Details: Error Bounds</A>
  <B> <A NAME="tex2html5702"
HREF="node1.html">Contents</A></B>
  <B> <A NAME="tex2html5704"
HREF="node152.html">Index</A></B>
<!--End of Navigation Panel-->
<ADDRESS>
<I>Susan Blackford</I>
<BR><I>1999-10-01</I>
</ADDRESS>
</BODY>
</HTML>
|