<?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: finally rebuilding</title>
	<atom:link href="http://wonkabar.org/2008/07/24/finally-rebuilding/feed/" rel="self" type="application/rss+xml" />
	<link>http://wonkabar.org/2008/07/24/finally-rebuilding/</link>
	<description>linux, databases, cartoons and cornflakes</description>
	<lastBuildDate>Mon, 01 Mar 2010 01:01:19 +0000</lastBuildDate>
	<generator>http://wordpress.org/?v=2.9.1</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Steve</title>
		<link>http://wonkabar.org/2008/07/24/finally-rebuilding/comment-page-1/#comment-46800</link>
		<dc:creator>Steve</dc:creator>
		<pubDate>Fri, 25 Jul 2008 16:05:02 +0000</pubDate>
		<guid isPermaLink="false">http://wonkabar.org/archives/450#comment-46800</guid>
		<description>For a modern desktop, building with -Os you won&#039;t notice much differences, yes ... but in working with small systems with an incredible small amount of space and memory, your compiler settings will make a huge difference.</description>
		<content:encoded><![CDATA[<p>For a modern desktop, building with -Os you won't notice much differences, yes ... but in working with small systems with an incredible small amount of space and memory, your compiler settings will make a huge difference.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mart Raudsepp</title>
		<link>http://wonkabar.org/2008/07/24/finally-rebuilding/comment-page-1/#comment-46799</link>
		<dc:creator>Mart Raudsepp</dc:creator>
		<pubDate>Fri, 25 Jul 2008 15:34:55 +0000</pubDate>
		<guid isPermaLink="false">http://wonkabar.org/archives/450#comment-46799</guid>
		<description>I basically also wanted to say what Tamas already said.
I guess I can add that if you have big CPU caches, which a Pentium D has I believe, then I don&#039;t see how -Os can be good in any way. I generally don&#039;t think twice about using -O2 if I have 1MB of L2 cache at least. If it&#039;s some old thing with some 256KB CPU cache, then I do ponder for a while before choosing though</description>
		<content:encoded><![CDATA[<p>I basically also wanted to say what Tamas already said.<br />
I guess I can add that if you have big CPU caches, which a Pentium D has I believe, then I don't see how -Os can be good in any way. I generally don't think twice about using -O2 if I have 1MB of L2 cache at least. If it's some old thing with some 256KB CPU cache, then I do ponder for a while before choosing though</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Tamas</title>
		<link>http://wonkabar.org/2008/07/24/finally-rebuilding/comment-page-1/#comment-46788</link>
		<dc:creator>Tamas</dc:creator>
		<pubDate>Fri, 25 Jul 2008 09:28:13 +0000</pubDate>
		<guid isPermaLink="false">http://wonkabar.org/archives/450#comment-46788</guid>
		<description>The -Os in GCC does not help with memory constrained systems, as it reduces only the code size, which makes up only a negligible proportion of an average desktop applications memory usage. 

Also generally it&#039;s not slower than other optimization levels in GCC. It uses almost the same optimizations as -O2, except for missing a couple ones that increase the code size. The result should be a 0-5% performance hit, however the smaller code helps with cache locality, thus it sometimes improves performance over -O2.

Furthermore, compiler optimization barely makes any difference for typical desktop applications, they are hardly CPU bound.</description>
		<content:encoded><![CDATA[<p>The -Os in GCC does not help with memory constrained systems, as it reduces only the code size, which makes up only a negligible proportion of an average desktop applications memory usage. </p>
<p>Also generally it's not slower than other optimization levels in GCC. It uses almost the same optimizations as -O2, except for missing a couple ones that increase the code size. The result should be a 0-5% performance hit, however the smaller code helps with cache locality, thus it sometimes improves performance over -O2.</p>
<p>Furthermore, compiler optimization barely makes any difference for typical desktop applications, they are hardly CPU bound.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
