<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body>
<p><font face="serif">It can be a pain to generate the data for a
speedup plot. Most students load all their data into a
spreadsheet, process it there, and then dump it back to hammer
for processing by gnuplot to determine the fit to Amdahl law.
Yuck. <br>
</font></p>
<p><font face="serif">I wrote a little perl script that will take
process/runtime data in two colums like this.</font></p>
<p><tt>1 1000</tt><tt><br>
</tt><tt>2 505</tt><tt><br>
</tt><tt>3 400</tt><tt><br>
</tt><tt>4 260</tt><tt><br>
</tt><tt>5 210</tt><tt><br>
</tt><tt>6 185</tt><br>
</p>
<p>and create the following.</p>
<p><tt><b>cobra:~ </b><b>%</b><b> </b>perl speedup.pl speedup.dat
</tt><tt><br>
</tt><tt>Procs Time Speedup</tt><tt><br>
</tt><tt>1 1000 1</tt><tt><br>
</tt><tt>2 505 1.98019801980198</tt><tt><br>
</tt><tt>3 400 2.5</tt><tt><br>
</tt><tt>4 260 3.84615384615385</tt><tt><br>
</tt><tt>5 210 4.76190476190476</tt><tt><br>
</tt><tt>6 185 5.40540540540541</tt><tt><br>
</tt></p>
<p>It is on the class webpage in the EXAMPLES section where you find
the gnuplot script for Amdahl's law. <br>
</p>
<p><br>
</p>
<pre class="moz-signature" cols="72">--
Andrew J. Pounds, Ph.D. (<a class="moz-txt-link-abbreviated" href="mailto:pounds_aj@mercer.edu">pounds_aj@mercer.edu</a>)
Professor of Chemistry and Computer Science
Director of the Computational Science Program
Mercer University, Macon, GA 31207 (478) 301-5627
</pre>
</body>
</html>