<?xml version="1.0" encoding="UTF-8" ?>
<?xml-stylesheet type="text/xsl" media="screen" href="/styles/xslt/rss.xslt"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:trackback="http://madskills.com/public/xml/rss/module/trackback/" xmlns:wfw="http://wellformedweb.org/CommentAPI/" xmlns:slash="http://purl.org/rss/1.0/modules/slash/" xmlns:media="http://search.yahoo.com/mrss/" xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" xmlns:c9="http://channel9.msdn.com">
<channel>
	<title>Channel 9 - Discussions by siddpuri</title>
	<atom:link rel="self" type="application/rss+xml" href="http://channel9.msdn.com/Niners/siddpuri/Discussions/RSS"></atom:link>
	<image>
		<url>http://mschnlnine.vo.llnwd.net/d1/Dev/App_Themes/C9/images/feedimage.png</url>
		<title>Channel 9 - Discussions by siddpuri</title>
		<link>http://channel9.msdn.com/Niners/siddpuri/Discussions</link>
	</image>
	<description>Channel 9 keeps you up to date with the latest news and behind the scenes info from Microsoft that developers love to keep up with. From LINQ to SilverLight – Watch videos and hear about all the cool technologies coming and the people behind them.</description>
	<link>http://channel9.msdn.com/Niners/siddpuri/Discussions</link>
	<language>en</language>
	<pubDate>Wed, 22 May 2013 12:03:54 GMT</pubDate>
	<lastBuildDate>Wed, 22 May 2013 12:03:54 GMT</lastBuildDate>
	<generator>Rev9</generator>
	<c9:totalResults>0</c9:totalResults>
	<c9:pageCount>0</c9:pageCount>
	<c9:pageSize>0</c9:pageSize>
	<item>
		<title>Coffeehouse - Microsoft Research: Accelerator</title>
		<description><![CDATA[<p>Maybe it was a bit sneaky, but really only a bit.&nbsp; <img src='http://ecn.channel9.msdn.com/o9/content/images/emoticons/emotion-1.gif' alt='Smiley' /><br>
<br>
Two things to consider:<br>
<br>
1) PCIe is much better than AGP and transferring data back.&nbsp; In effect, the asymmetry has been erased.<br>
<br>
2) On the smaller benchmarks, the communication time is kind of irrelevant.&nbsp; If convolving with a 5x5 filter is your entire computation, then it’s probably not worth parallelizing anyway.&nbsp; For the larger benchmarks, take StereoMatch as an example.&nbsp; The communication
 time adds &lt; 25%.&nbsp; That’s pretty small compared to a 4x speedup over&nbsp; C&#43;&#43;.</p>
<p>Sidd<br>
</p>]]></description>
		<link>http://channel9.msdn.com/Forums/Coffeehouse/215583-Microsoft-Research-Accelerator/937860c3a193495da4d89deb01892381#937860c3a193495da4d89deb01892381</link>
		<pubDate>Thu, 03 Aug 2006 22:50:02 GMT</pubDate>
		<guid isPermaLink="false">http://channel9.msdn.com/Forums/Coffeehouse/215583-Microsoft-Research-Accelerator/937860c3a193495da4d89deb01892381#937860c3a193495da4d89deb01892381</guid>
		<dc:creator>siddpuri</dc:creator>
		<slash:comments>9</slash:comments>
		<wfw:commentRss>http://channel9.msdn.com/Niners/siddpuri/Discussions/RSS</wfw:commentRss>
	</item>
</channel>
</rss>