<?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"
	>
<channel>
	<title>Comments on: Powering a warehouse-sized computer - part 2</title>
	<atom:link href="http://storagemojo.com/2007/07/31/powering-a-warehouse-sized-computer-part-2/feed/" rel="self" type="application/rss+xml" />
	<link>http://storagemojo.com/2007/07/31/powering-a-warehouse-sized-computer-part-2/</link>
	<description>Data storage info &#38; analysis</description>
	<pubDate>Fri, 21 Nov 2008 16:30:42 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.6.3</generator>
		<item>
		<title>By: Jeff M</title>
		<link>http://storagemojo.com/2007/07/31/powering-a-warehouse-sized-computer-part-2/#comment-102650</link>
		<dc:creator>Jeff M</dc:creator>
		<pubDate>Thu, 02 Aug 2007 01:49:02 +0000</pubDate>
		<guid isPermaLink="false">http://storagemojo.com/2007/07/31/powering-a-warehouse-sized-computer-part-2/#comment-102650</guid>
		<description>Reducing idle power is not something particularly useful to us since our large 1000+ cpu clusters run at over 95% utilization.</description>
		<content:encoded><![CDATA[<p>Reducing idle power is not something particularly useful to us since our large 1000+ cpu clusters run at over 95% utilization.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Wes Felter</title>
		<link>http://storagemojo.com/2007/07/31/powering-a-warehouse-sized-computer-part-2/#comment-102159</link>
		<dc:creator>Wes Felter</dc:creator>
		<pubDate>Tue, 31 Jul 2007 18:53:42 +0000</pubDate>
		<guid isPermaLink="false">http://storagemojo.com/2007/07/31/powering-a-warehouse-sized-computer-part-2/#comment-102159</guid>
		<description>Robin,

Reducing idle power is worth doing, but it's not my problem since I don't design chips. :-) There may be more vendor emphasis on this topic once the SPECpower benchmark is released. There were two independent studies recently showing that peak power efficiency and non-peak efficiency are quite different for Intel and AMD processors, so Google isn't the only one paying attention.

Saving power is nice because it reduces your power bill, but IMO the key point of the warehouse paper is power provisioning. If you put a 30A circuit into a rack but use only 20A of it, then you are wasting infrastructure. Watch for my papers on this topic.</description>
		<content:encoded><![CDATA[<p>Robin,</p>
<p>Reducing idle power is worth doing, but it&#8217;s not my problem since I don&#8217;t design chips. <img src='http://storagemojo.com/wp-includes/images/smilies/icon_smile.gif' alt=':-)' class='wp-smiley' /> There may be more vendor emphasis on this topic once the SPECpower benchmark is released. There were two independent studies recently showing that peak power efficiency and non-peak efficiency are quite different for Intel and AMD processors, so Google isn&#8217;t the only one paying attention.</p>
<p>Saving power is nice because it reduces your power bill, but IMO the key point of the warehouse paper is power provisioning. If you put a 30A circuit into a rack but use only 20A of it, then you are wasting infrastructure. Watch for my papers on this topic.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

<!-- Dynamic Page Served (once) in 2.413 seconds -->
