<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: Nvidia Snaps Out Snappier Tesla GPU Coprocessors</title>
	<atom:link href="http://insidehpc.com/2011/05/19/nvidia-snaps-out-snappier-tesla-gpu-coprocessors/feed/" rel="self" type="application/rss+xml" />
	<link>http://insidehpc.com/2011/05/19/nvidia-snaps-out-snappier-tesla-gpu-coprocessors/</link>
	<description>HPC News Without the Noise for Supercomputing Professionals &#124; insideHPC</description>
	<lastBuildDate>Sun, 09 Jun 2013 01:54:13 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.1.1</generator>
	<item>
		<title>By: broom</title>
		<link>http://insidehpc.com/2011/05/19/nvidia-snaps-out-snappier-tesla-gpu-coprocessors/#comment-374985</link>
		<dc:creator>broom</dc:creator>
		<pubDate>Tue, 22 Nov 2011 21:15:28 +0000</pubDate>
		<guid isPermaLink="false">http://insidehpc.com/?p=19990#comment-374985</guid>
		<description>Nice article.

Although as ce107 says, you&#039;re not going to be able to use desktop graphics cards like the GTX 590 as you are limited to 1/4 of the DP that you get in the server cards (this is an intentional and artificial change apparently).

This is really quite tragic.  If we want to realize the dream of every researcher having access to a &quot;super-computer&quot;, it would be ideal to simply have a desktop computer packed with a pair of GTX 590s cranking out the calculations.  I wonder how much of a market there for something inbetween the common gaming desktop and the HPC server?  Could we convince nVidia to supply consumer cards with full DP unlocked?</description>
		<content:encoded><![CDATA[<p>Nice article.</p>
<p>Although as ce107 says, you&#8217;re not going to be able to use desktop graphics cards like the GTX 590 as you are limited to 1/4 of the DP that you get in the server cards (this is an intentional and artificial change apparently).</p>
<p>This is really quite tragic.  If we want to realize the dream of every researcher having access to a &#8220;super-computer&#8221;, it would be ideal to simply have a desktop computer packed with a pair of GTX 590s cranking out the calculations.  I wonder how much of a market there for something inbetween the common gaming desktop and the HPC server?  Could we convince nVidia to supply consumer cards with full DP unlocked?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: ce107</title>
		<link>http://insidehpc.com/2011/05/19/nvidia-snaps-out-snappier-tesla-gpu-coprocessors/#comment-339796</link>
		<dc:creator>ce107</dc:creator>
		<pubDate>Fri, 20 May 2011 22:37:44 +0000</pubDate>
		<guid isPermaLink="false">http://insidehpc.com/?p=19990#comment-339796</guid>
		<description>&quot;Nvidia does not provide floating-point performance ratings on the GTX 590, but it should be somewhere around 1.24 teraflops double-precision.&quot;

Actually it is a dirty little secret that DP performance on GeForce cards is crippled (Wikipedia does have a line &quot;For consumer products, double precision performance has been limited to a quarter of that of the &quot;full&quot; Fermi architecture.&quot;). For DP that is half of FP you need to go to Quadro or Tesla hardware.</description>
		<content:encoded><![CDATA[<p>&#8220;Nvidia does not provide floating-point performance ratings on the GTX 590, but it should be somewhere around 1.24 teraflops double-precision.&#8221;</p>
<p>Actually it is a dirty little secret that DP performance on GeForce cards is crippled (Wikipedia does have a line &#8220;For consumer products, double precision performance has been limited to a quarter of that of the &#8220;full&#8221; Fermi architecture.&#8221;). For DP that is half of FP you need to go to Quadro or Tesla hardware.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
