<?xml version="1.0" encoding="utf-8"?>
			<feed xmlns="http://www.w3.org/2005/Atom">
				<title type="text">Think-About-Tech -  User's Blog Feed (Andrzej Bania)</title>
				<updated>2010-09-06T23:17:42+01:00</updated>
				<id>http://www.thinkabouttech.com/</id>
				<link rel="alternate" type="text/html" hreflang="en"
					href="http://www.thinkabouttech.com/search/?PostBackAction=Search&amp;Type=Comments&amp;Page=1&amp;BlogSearch=1&amp;AuthUserID=14&amp;FeedTitle=User%27s+Blog+Feed+%28Andrzej+Bania%29"/>
				<link rel="self" type="application/atom+xml"
					href="http://www.thinkabouttech.com/search/?PostBackAction=Search&amp;Type=Comments&amp;Page=1&amp;Feed=ATOM&amp;BlogSearch=1&amp;AuthUserID=14&amp;FeedTitle=User%27s+Blog+Feed+%28Andrzej+Bania%29"/>
				<generator
					uri="http://getvanilla.com/"
					version="1.1.5a">
					Lussumo Vanilla &amp; Feed Publisher
				</generator>
				<entry>
			<title>COMPUTER BUYER RIP: Passing of a Genre...</title>
			<link rel="alternate" href="http://www.thinkabouttech.com/discussion/48/computer-buyer-rip-passing-of-a-genre/?Focus=161#Comment_161" type="application/xhtml+xml" hreflang="en"/>
			<id>http://www.thinkabouttech.com/discussion/48/computer-buyer-rip-passing-of-a-genre/?Focus=161#Comment_161</id>
			<published>2009-01-10T10:20:28+00:00</published>
			<updated>2010-09-06T23:17:42+01:00</updated>
			<author>
				<name>Andrzej Bania</name>
				<uri>http://www.thinkabouttech.com/account/14/</uri>
			</author>
			<summary type="text" xml:lang="en">
				The March 2009 edition of Computer Buyer will be the last issue ever.
Having joined MESH at the back end of 1994, the first system I ever spec-ed for review was a P75 machine for a young staff writer ...
			</summary>
			<content type="html">
				<![CDATA[<p>The March 2009 edition of Computer Buyer will be the last issue ever.</p>
<p>Having joined MESH at the back end of 1994, the first system I ever spec-ed for review was a P75 machine for a young staff writer called Paul Saunders.</p>
<p>It appeared in the February 1995 issue and got a Recommended award, complete with a small pic of the mainboard - next to the CPU - with the comment &quot;Would have been faster if Pipeline Burst Mode SRAM had been fitted here&quot;. We paid a small fortune to Dennis to have the review reprinted - and we stuck a near-naked girl on the reverse - with the headline &quot;SCANDALOUS&quot;.</p>
<p>I've just read through the last issue and, in Buyer's final ever group test, I've spec-ed the winner. Nice set of book ends  <img src="/extensions/Smile/msn/icon_smile.gif" class="smiley" alt=":)"/> </p>
<p>At it's peak, with Julian Lloyd Evans running the show, Computer Buyer was over 400 pages in...</p>
]]>
			</content>
		</entry>
		<entry>
			<title>Fi7EPOWER: Triumph of the Local Hero</title>
			<link rel="alternate" href="http://www.thinkabouttech.com/discussion/23/fi7epower-triumph-of-the-local-hero/?Focus=78#Comment_78" type="application/xhtml+xml" hreflang="en"/>
			<id>http://www.thinkabouttech.com/discussion/23/fi7epower-triumph-of-the-local-hero/?Focus=78#Comment_78</id>
			<published>2008-11-03T18:43:20+00:00</published>
			<updated>2010-09-06T23:17:42+01:00</updated>
			<author>
				<name>Andrzej Bania</name>
				<uri>http://www.thinkabouttech.com/account/14/</uri>
			</author>
			<summary type="text" xml:lang="en">
				Creating hotrods that blaze 400 metres down a track before burning out is one kind of fast. Another is the 70 lap extraveganza of Formula One. However, when it comes to translating 'what you see on ...
			</summary>
			<content type="html">
				<![CDATA[<p>Creating hotrods that blaze 400 metres down a track before burning out is one kind of fast. Another is the 70 lap extraveganza of Formula One. However, when it comes to translating 'what you see on the track into your expectation for a car you can drive yourself', the 24-hour Le Mans race sticks in the mind. </p>
<p>Same with system benchmarks. While the prime number calculators and PCMark whizzy-bangs are all well and good, if you were to ask Sun, IBM, Dell, HP and the rest of the world-class, multi-nationals which benchmark they respect most, the likely response is SPEC.</p>
<p>Not much to look at, but expensive to use and a pain to get scores ratified with, SPEC is the Standard Performance Evaluation Corporation's easy to remember abbreviation - and their rather demanding benchmark for system processing takes a scary 3 days to complete (best case scenario).</p>
<p>Until 5:00am on Monday, 3rd...</p>
]]>
			</content>
		</entry>
		<entry>
			<title>Vista 64, I7, Memory and Layout/Management...</title>
			<link rel="alternate" href="http://www.thinkabouttech.com/discussion/15/vista-64-i7-memory-and-layoutmanagement/?Focus=60#Comment_60" type="application/xhtml+xml" hreflang="en"/>
			<id>http://www.thinkabouttech.com/discussion/15/vista-64-i7-memory-and-layoutmanagement/?Focus=60#Comment_60</id>
			<published>2008-10-21T08:08:02+01:00</published>
			<updated>2010-09-06T23:17:42+01:00</updated>
			<author>
				<name>Andrzej Bania</name>
				<uri>http://www.thinkabouttech.com/account/14/</uri>
			</author>
			<summary type="text" xml:lang="en">
				When I was at ATI, I remember the driver team's joy at realising that by changing the memory map for each  game - they could pick up double-digit improvements in performance
e.g. for UT they would ...
			</summary>
			<content type="html">
				<![CDATA[<p>When I was at ATI, I remember the driver team's joy at realising that by changing the memory map for each  game - they could pick up double-digit improvements in performance</p>
<p>e.g. for UT they would lay out the AA, texture etc area one way - and a differet way for HL2 etc - and between all of these changes...   pretty impressive speed gains were had</p>
<p>That was on a small scale, with limited memory (128/256/512) and a very narrow scope of programmes (i.e. 'current games')</p>
<p>As you move to a full blown OS like Vista 64, the chances to 'get it wrong' must increase exponentially...</p>
<p>...or, maybe, as the memory speed and volume increase - is it possible that the problem evaporates ?</p>
<p>My money is on 'gets worse'</p>
<p>With Blackbeard's comments regarding the differences available to 2x2GB vs 2x1.5GB (see link below)...</p>
<p>...it makes you wonder just how much difference can Microsoft make to the performance of future operating systems, simply by improving memory management ?</p>
<p>Could Vista have launched as 'OK' - rather than 'complete pants' if the memory was managed better?</p>
<p>I'm wondering if there is a 'Laziness Parallel' in operation here - where the OS coders get more and more complacent as the amount of ram available increases and becomes faster...</p>
<p>...or, with the advent of i7's integrated memory controller, will Intel claim all responsibility for memory management ?</p>
<p><a href='http://www.thinkabouttech.com/discussion/11/dear-memoryman-with-i7-is-6gb-8gb-or-9gb-correct-/#pgbottom' target="_blank" rel="nofollow">http://www.thinkabouttech.com/discussion/11/dear-memoryman-with-i7-is-6gb-8gb-or-9gb-correct-/#pgbottom</a></p>
]]>
			</content>
		</entry>
		<entry>
			<title>Dear Memory-Man: With I7, is 6GB, 8GB or 9GB Correct ?</title>
			<link rel="alternate" href="http://www.thinkabouttech.com/discussion/11/dear-memoryman-with-i7-is-6gb-8gb-or-9gb-correct-/?Focus=50#Comment_50" type="application/xhtml+xml" hreflang="en"/>
			<id>http://www.thinkabouttech.com/discussion/11/dear-memoryman-with-i7-is-6gb-8gb-or-9gb-correct-/?Focus=50#Comment_50</id>
			<published>2008-10-19T17:00:48+01:00</published>
			<updated>2010-09-06T23:17:42+01:00</updated>
			<author>
				<name>Andrzej Bania</name>
				<uri>http://www.thinkabouttech.com/account/14/</uri>
			</author>
			<summary type="text" xml:lang="en">
				When OCZ et al started looking at shipping 3GB blister packs of memory for 32-bit Windows, I thought 'Bonza!' - at last we have some common sense
I was listening to Ogden's drunken mumblings last ...
			</summary>
			<content type="html">
				<![CDATA[<p>When OCZ et al started looking at shipping 3GB blister packs of memory for 32-bit Windows, I thought 'Bonza!' - at last we have some common sense</p>
<p>I was listening to Ogden's drunken mumblings last week and he seemed convinced that 6GB (Corsair, naturally!) was the way to go for Vista 64 on an i7</p>
<p>In my heart, I agree</p>
<p>However, on reflection, I'm wondering which 'logic path' is actually accurate:-<br />
1) 6GB because there are 3 channels and it 'just makes sense'<br />
2) 8GB so that you have 2GB per core and the channel-flip time-lag will be small anyway<br />
3) 9GB so that each channel has 3GB available...    and/or there is 2GB per core (1GB per SMT) and 1GB over for good measure</p>
<p> <img src="/extensions/Smile/msn/icon_erm.gif" class="smiley" alt=":erm:"/>  <strong>Dear Memory Man, can you steer us in the right direction ?</strong>  <img src="/extensions/Smile/msn/icon_erm2.gif" class="smiley" alt=":erm2:"/><br />
(forgetting entirely about cost - just tell us about pure logic and performance)</p>
]]>
			</content>
		</entry>
		<entry>
			<title>4650 Vs X1900XTX (with 'agog' but Not Much Science)</title>
			<link rel="alternate" href="http://www.thinkabouttech.com/discussion/7/4650-vs-x1900xtx-with-agog-but-not-much-science/?Focus=33#Comment_33" type="application/xhtml+xml" hreflang="en"/>
			<id>http://www.thinkabouttech.com/discussion/7/4650-vs-x1900xtx-with-agog-but-not-much-science/?Focus=33#Comment_33</id>
			<published>2008-10-12T18:16:04+01:00</published>
			<updated>2010-09-06T23:17:42+01:00</updated>
			<author>
				<name>Andrzej Bania</name>
				<uri>http://www.thinkabouttech.com/account/14/</uri>
			</author>
			<summary type="text" xml:lang="en">
				When launched in January 2006, Jon Peddie described the X1900XTX as &amp;quot;...the fastest graphics card, ATI have made it available immediately at launch and in good quantities around the world. ...
			</summary>
			<content type="html">
				<![CDATA[<p>When launched in January 2006, Jon Peddie described the X1900XTX as <em><strong>&quot;...the fastest graphics card, ATI have made it available immediately at launch and in good quantities around the world. We expect this new board to be very well received</strong></em>&quot;</p>
<p>The cooler was a 'stonking great one' and you'd definitely know if you had one plugged into your AMD-FX powered uber-gaming rig</p>
<p>The recommended price for the <strong>X1900XTX </strong>was <strong>$649 </strong><em>(around <strong>£380 </strong>in real money - probably with some tax on top)</em></p>
<p>Roll the clock forward<strong> just under 3 years</strong>, and HIS has just started to make <strong>passively cooled 4650 cards </strong>available in the UK market</p>
<p>I asked <strong>Beyond3D</strong>'s resident Graphics Guru for a 'quick tech comparison' between these 2 cards - and the match up seems close...</p>
<p>    &quot;Shaders/TMUs/ROPs/bandwidth: 320/32/8/25.6 plays 192/16/16/49.6...  similar core clocks, so 4650 is quicker...&quot;</p>
<p>He did say that as soon as you engaged AA (or anything else that required mucho-memory) - then the advantage would <strong><em>disappear quicker than a JD&amp;Coke into a Spode</em></strong>...</p>
<p>...but the new card is completely passively cooled and will launch at <strong>£59 </strong>including vat (around <strong>$99</strong>)</p>
<p>Naturally, there are a lot of good products in the present market - and I'm not going to type here that I am a '100% unbiased bastion of even-ness'...</p>
<p>...but, bl**dy hell - hasn't the market moved some in the last 2+ years !</p>
<p>Any  <img src="/extensions/Smile/msn/icon_hug.gif" class="smiley" alt="({)"/> old-fogey/young-turk  <img src="/extensions/Smile/msn/icon_hugback.gif" class="smiley" alt="(})"/>  care to comment ?</p>
]]>
			</content>
		</entry>
		
			</feed>