<?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: Thoughts on &#8220;Valve goes Multicore&#8221;</title>
	<atom:link href="http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/</link>
	<description>A Blog on Parallel Programming and Concurrency by Michael Suess</description>
	<pubDate>Sat, 31 Jul 2010 08:30:32 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.7</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Thinking Parallel &#187; Blog Archive &#187; Multi-Threaded Challenges in the Game Space - a Conversation with Tom Leonard of Valve Fame</title>
		<link>http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/comment-page-1/#comment-480</link>
		<dc:creator>Thinking Parallel &#187; Blog Archive &#187; Multi-Threaded Challenges in the Game Space - a Conversation with Tom Leonard of Valve Fame</dc:creator>
		<pubDate>Mon, 08 Jan 2007 20:39:12 +0000</pubDate>
		<guid isPermaLink="false">http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/#comment-480</guid>
		<description>[...] A while ago I posted some comments on the press coverage of an event by Valve, where they explained threading in the Source Engine to a couple of tech-journalists. Unfortunately, the coverage left many open questions and some things even appeared to be wrong. Fortunately, Tom Leonard, the threading guru from Valve is a nice guy and agreed to answer some of my questions via email. This post contains the conversation we had. [...]</description>
		<content:encoded><![CDATA[<p>[...] A while ago I posted some comments on the press coverage of an event by Valve, where they explained threading in the Source Engine to a couple of tech-journalists. Unfortunately, the coverage left many open questions and some things even appeared to be wrong. Fortunately, Tom Leonard, the threading guru from Valve is a nice guy and agreed to answer some of my questions via email. This post contains the conversation we had. [...]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Michael Suess</title>
		<link>http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/comment-page-1/#comment-80</link>
		<dc:creator>Michael Suess</dc:creator>
		<pubDate>Thu, 07 Dec 2006 08:56:49 +0000</pubDate>
		<guid isPermaLink="false">http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/#comment-80</guid>
		<description>@Bill: I am not familiar enough with the computer games industry to comment on their software-engineering practices - but I suspect as in every industry, there are companies that are just happily hacking along and there are companies that are trying hard to improve their processes.</description>
		<content:encoded><![CDATA[<p>@Bill: I am not familiar enough with the computer games industry to comment on their software-engineering practices - but I suspect as in every industry, there are companies that are just happily hacking along and there are companies that are trying hard to improve their processes.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Bill King</title>
		<link>http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/comment-page-1/#comment-73</link>
		<dc:creator>Bill King</dc:creator>
		<pubDate>Tue, 21 Nov 2006 23:40:22 +0000</pubDate>
		<guid isPermaLink="false">http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/#comment-73</guid>
		<description>Nicely put. Being a multi-threading junky too (and recently discovering the joys of OpenMP) led me to your blog, so I'm not quite the same target as either of those articles, but yeah, they left me with _exactly_ the same questions going on. In regards to isaac, the problem with game programmers is that the industry is geared towards speed of writing code, not quality of code. Some of my most amusing times have been speaking to friends in the gaming industry, and opening their eyes with techniques and ideas and concepts that most application developers see as standard practice. Game coding seems to be finally starting to evolve from it's hacky ancestry into something finally resembling decent, only... it's idea of decent is 5-10 years old. I personally think OpenMP would be a godsend for the gaming industry, because of it's simplicity to move code from a single-threaded environment into a multi-threaded powerhouse with a few twists of the wrist.

Anyway, I've diverged. Great summary :)</description>
		<content:encoded><![CDATA[<p>Nicely put. Being a multi-threading junky too (and recently discovering the joys of OpenMP) led me to your blog, so I&#8217;m not quite the same target as either of those articles, but yeah, they left me with _exactly_ the same questions going on. In regards to isaac, the problem with game programmers is that the industry is geared towards speed of writing code, not quality of code. Some of my most amusing times have been speaking to friends in the gaming industry, and opening their eyes with techniques and ideas and concepts that most application developers see as standard practice. Game coding seems to be finally starting to evolve from it&#8217;s hacky ancestry into something finally resembling decent, only&#8230; it&#8217;s idea of decent is 5-10 years old. I personally think OpenMP would be a godsend for the gaming industry, because of it&#8217;s simplicity to move code from a single-threaded environment into a multi-threaded powerhouse with a few twists of the wrist.</p>
<p>Anyway, I&#8217;ve diverged. Great summary <img src='http://www.thinkingparallel.com/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' /> </p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Isaack</title>
		<link>http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/comment-page-1/#comment-72</link>
		<dc:creator>Isaack</dc:creator>
		<pubDate>Tue, 21 Nov 2006 15:39:30 +0000</pubDate>
		<guid isPermaLink="false">http://www.thinkingparallel.com/2006/11/21/thoughts-on-valve-goes-multicore/#comment-72</guid>
		<description>What used to make it difficult to do fast multithreading with gaming when you are spending most of the time on the GPU, is that, usually only once thread can control the graphics system. At least on Windows and at least on OpenGL, I forgot if DirectX also had the same limitation.
Also, I think that Epic Games always used DirectX.

OpenGL on OS X does allow you to control the graphics with different threads. Haven't played with it yet though.

Anyway, so the problem was. You couldn't load and upload textures into the GPU from different threads, only load from disk and then have the main thread upload the texture. And it went on like this. There wasn't really enough to do in other threads to make it a worthwhile speedup.
But I think that might have changed now, with todays games and their more advanced physics, AI and other CPU stuff.

John Carmack of iD Software spend alot of time on this long ago and had SMP support in Quake2 to begin with, but it didn't give him much speedup.</description>
		<content:encoded><![CDATA[<p>What used to make it difficult to do fast multithreading with gaming when you are spending most of the time on the GPU, is that, usually only once thread can control the graphics system. At least on Windows and at least on OpenGL, I forgot if DirectX also had the same limitation.<br />
Also, I think that Epic Games always used DirectX.</p>
<p>OpenGL on OS X does allow you to control the graphics with different threads. Haven&#8217;t played with it yet though.</p>
<p>Anyway, so the problem was. You couldn&#8217;t load and upload textures into the GPU from different threads, only load from disk and then have the main thread upload the texture. And it went on like this. There wasn&#8217;t really enough to do in other threads to make it a worthwhile speedup.<br />
But I think that might have changed now, with todays games and their more advanced physics, AI and other CPU stuff.</p>
<p>John Carmack of iD Software spend alot of time on this long ago and had SMP support in Quake2 to begin with, but it didn&#8217;t give him much speedup.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
