<?xml version="1.0" encoding="ISO-8859-1"?>
<!-- generator="HardwareAnalysis.Com" -->
<rdf:RDF
    xmlns="http://purl.org/rss/1.0/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
    xmlns:dc="http://purl.org/dc/elements/1.1/">
    <channel rdf:about="">
        <title>Hardware Analysis - Re: Dual core CPUs, don’t get your hopes up</title>
        <description>Hardware Analysis Community Forums</description>
        <link>http://www.hardwareanalysis.com/content/topic/38545/</link>
        <image rdf:resource="http://media.hardwareanalysis.com/halogo.gif" />
       <dc:date>2009-01-08T08:34:50-05:00</dc:date>
        <items>
            <rdf:Seq>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=40#259581"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=40#238561"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=40#238556"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=40#235058"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#235018"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#235006"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#234983"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#234976"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233927"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233808"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233807"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233802"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233800"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233792"/>
                <rdf:li rdf:resource="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233789"/>
            </rdf:Seq>
        </items>
    </channel>
    <image rdf:about="http://media.hardwareanalysis.com/halogo.gif">
        <title>Hardware Analysis</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/</link>
        <url>http://media.hardwareanalysis.com/halogo.gif</url>
    </image>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=40#259581">
        <dc:format>text/html</dc:format>
        <dc:date>2005-03-29T01:49:19-05:00</dc:date>
        <dc:creator>Super XP</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=40#259581</link>
        <description>Nope, The Opteron is keeping its Integrated Memory Controller &amp;amp; Intel is working on a similar solution for there future Dual Core's.&lt;br /&gt;
&lt;br /&gt;
Typical for Intel, AMD does all the work &amp;amp; Intel just follows like always. &lt;br /&gt;
&lt;br /&gt;
Good Work AMD. And the Athlon 64 Dual Core CPU's are going to be a lot faster than Intel's Dual Core because of Intel's problematic single FSB feeding two cores. Ouch, that will hurt performance.&lt;br /&gt;
</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=40#238561">
        <dc:format>text/html</dc:format>
        <dc:date>2005-02-07T06:46:07-05:00</dc:date>
        <dc:creator>Sean B</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=40#238561</link>
        <description>It's likely that they'll start producing dual core Opterons and Xeons, but the Opterons are probably going to have the memory controllers kicked out of the processor and back onto the chipset.&lt;br /&gt;
&lt;br /&gt;
They'll have to do something to reallocate the space.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
Carlos, are you talking about Extended 64bit or Dual Core?&lt;br /&gt;
&lt;br /&gt;
There's an obvious reason if you mean extended 64bit... just look it up.&lt;br /&gt;
&lt;br /&gt;
If you meant dual core, the reason is that while Windows 2k/XP and a ton of Linux distros support SMP and load balancing, they don't have any utilization for dual core.&lt;br /&gt;
&lt;br /&gt;
Dual core offers some very nice advantages, such as each core has access to the other core's cache, which means the cores can work directly together as a team... rather than multiprocessors, which work together indirectly like an assembly line.&lt;br /&gt;
&lt;br /&gt;
Current software and operating systems don't know how to utilize features like that... and in order for it to work as intended, it needs to have programmers dedicated to rebuilding it, rather than just making a patch.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
Windows XP Extended 64bit Edition is basically just a patched up version of Windows XP that was recompiled to support Extended 64bit instruction sets.  That's why it probably won't offer significant performance increases over Windows XP, but it might improve stability.</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=40#238556">
        <dc:format>text/html</dc:format>
        <dc:date>2005-02-07T06:28:50-05:00</dc:date>
        <dc:creator>GS059</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=40#238556</link>
        <description>but remember, if you open up the task manager, you see alot of tasks going on. with the dual cor or evan the dual/ quad processor, it splits them up. so you can have one of the cores making your os work, and the other letting you play a game. than you would have a 3 ghz processor for your os, and another dedicated to gaming, if of corce you can make a program for load balancing so that you can tell the computer where to put the load. with your computer it will put some of the processes on on and some o n the other. it is still sortof worth while, but i dont thing that the average user will be able to use it as much as people like me.&lt;br /&gt;
&lt;br /&gt;
ps i want a quad processor, or if they could do this, a quad processor system with dual cores. i would only want each one to be like 1 or 2 ghz each though, the wxtra speed is not neaded.</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=40#235058">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-31T13:11:42-05:00</dc:date>
        <dc:creator>carl0s ki</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=40#235058</link>
        <description>i dont understand why MS would possibly have to recode the entire Windows Operating system like someone said&lt;br /&gt;
&lt;br /&gt;
wouldn't it just require them to do a few small modification then compile it using a 64bit AMD processor?&lt;br /&gt;
&lt;br /&gt;
linux software seems to work like that &lt;img src=&quot;http://media.hardwareanalysis.com/smilies/smile1.gif&quot; width=&quot;14&quot; height=&quot;14&quot; border=&quot;0&quot; alt=&quot;:)&quot; title=&quot;:)&quot;&gt; </description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#235018">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-31T11:06:07-05:00</dc:date>
        <dc:creator>Sean B</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#235018</link>
        <description>&lt;a class=&quot;ext&quot; href=&quot;/action/r/http://www.superspeed.com/desktop/ramdisk.php&quot; target=&quot;_blank&quot;&gt;http://www.superspeed.com/desktop/ramdisk.php&lt;/a&gt;&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
Easiest software to use to make it work.</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#235006">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-31T10:45:38-05:00</dc:date>
        <dc:creator>Who Knows?</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#235006</link>
        <description>ok, sounds good to me. &lt;br /&gt;
&lt;br /&gt;
is ramdrive something you designate or is it just your terminoligy for effecient use?&lt;br /&gt;
if you have to designate to make a ramdrive, what do I do?&lt;br /&gt;
&lt;br /&gt;
thanks!</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#234983">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-31T09:29:18-05:00</dc:date>
        <dc:creator>Sean B</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#234983</link>
        <description>Harley, SLi'ed cards don't really offer 70% performance increases in real world use.&lt;br /&gt;
&lt;br /&gt;
They're merely a scheme to boast about benchmark scores.&lt;br /&gt;
&lt;br /&gt;
I personally am waiting for the 7 Series cards to come out... the 6 series isn't too impressive.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
Jack... that's what I do when I'm rendering video.  The problem is it's hard to do that for normal use.</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#234976">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-31T08:37:22-05:00</dc:date>
        <dc:creator>Who Knows?</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#234976</link>
        <description>how about using the ram to pick up the slack, instead of everything waiting on the hard drive the ram can hold and deliver data. so would that be usefull and efficient bandwidth, like mapping? or how about compressing and uncompressing.</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233927">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-28T20:45:21-05:00</dc:date>
        <dc:creator>D Harley</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#233927</link>
        <description>This is along the lines of what I'm talking about Sean, except with using A64's or P4's with the corresponding chipset. Yep, thats dual cpu, ATX, 6 usb ports, and even onboard audio. Even DDR400!&lt;br /&gt;
 &lt;a class=&quot;ext&quot; href=&quot;/action/r/http://www.newegg.com/app/ViewProductDesc.asp?description=13-130-436&amp;amp;depa=0&quot; target=&quot;_blank&quot;&gt;http://www.newegg.com/app/ViewProductDesc.asp?description=13-1...amp;depa=0&lt;/a&gt;&lt;br /&gt;
&lt;br /&gt;
While it's not exactly what my concept board would be, it's pretty close. As for Intels HD audio, I'm not interested in pluging in surround sound speakers to my HT/multifunctional PC. Strictly optical s/pdif inputs and outputs for recording, playback and pass through. The only problem seems to be the software. There's nothing wrong with the technology depending on whose chip you use. C-media appears to be better than realteks. The 24-bit/48-96 khz spec is the same as Creatives. Forget D/A conversion. Keep it ALL digital 'till it hits my HT receiver. With the newer Digital HT receivers you keep the signal digital all the way if I'm not mistaken. Yes, cpu usage increases but if you are even able to get a 50% performance increase using dual desktop cpu's (costing less than Xeon's,Opteron's) then the cost of the extra cpu would be worth it compared to the expected ridiculous cost of dual core. As you can see, the cost of this board is NOT prohibitive nowadays.&lt;br /&gt;
&lt;br /&gt;
Won't longhorn also be advantageous to SMP??&lt;br /&gt;
If Nvidia's load balancing Sli solution yields a 70% performance increase with the dual 6600gt then it is worth consideration when it outperforms an equally priced 6800. IMHO&lt;br /&gt;
</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233808">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-28T14:11:18-05:00</dc:date>
        <dc:creator>Kash</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#233808</link>
        <description>Ah.... I don't understand because it's 9:11am.... whoa, wait a sec... *shudders* creepy.</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233807">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-28T14:00:28-05:00</dc:date>
        <dc:creator>Sean B</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#233807</link>
        <description>I mean if I were to use an Athlon or any other processor with an on die memory controller, that's what I would do.&lt;br /&gt;
&lt;br /&gt;
The on die memory controller allows a direct interface with the memory, which would give you a significant performance increase if you used a portion of your memory as ramdrive.&lt;br /&gt;
&lt;br /&gt;
The only places data would go would be ramdrive &amp;gt;&amp;gt; processor &amp;gt;&amp;gt; ram &amp;gt;&amp;gt; processor &amp;gt;&amp;gt; ramdrive, which would be very fast due to potentially higher bandwidths and lower latencies.&lt;br /&gt;
</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233802">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-28T13:48:16-05:00</dc:date>
        <dc:creator>Kash</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#233802</link>
        <description>Sean, you say that you would load up your computer with 4GB of memory, or whatever... Wouldn't it be more cost-effective to just use the onboard memory controller, or whatever you're referring to?&lt;br /&gt;
&lt;br /&gt;
And I'm sorry I'm not being more technical... it's about 9am and I've been up for about 21 hours... And I probably won't go to sleep until about 3am Saturday morning... if not later... like around 5am.</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233800">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-28T13:43:49-05:00</dc:date>
        <dc:creator>Sean B</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#233800</link>
        <description>Well, the most important part is file allocation.&lt;br /&gt;
&lt;br /&gt;
Allowing the processor to have direct access to the hard drive, opposed to going through I/O controllers and channels, would allow the processor to become the file allocator and rapidly prefetch files.  Current prefetching has no logic or prediction behind it... it's typically told what to fetch before it's actually required.  However, with the processor having entire access to the code even before the code has access to the file(provided that you have a prefetch handler looking for files within code) could allow it to handle this prefetching more efficiently.&lt;br /&gt;
&lt;br /&gt;
Keep in mind that this is all currently unneccessary, considering applications haven't reached the complexity to require heavy prefetchers.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
Current technology could benefit from the following: &lt;br /&gt;
&lt;br /&gt;
You move everything that is read only to a flash or ramdrive.  &lt;br /&gt;
&lt;br /&gt;
While flash memories currently have very low bandwidth, the technology will soon be upgraded to up to UDMA100 speeds... and unlike current hard drives, flash memory can deliver 100Mbit/sec because the memory is static.  This also means that files can be accessed quickly... with absolutely no seek time.  Windows would boot instantly.&lt;br /&gt;
&lt;br /&gt;
All other files that have dynamic properties, such as in an instance where they aren't read only or that they're user created files would be stored on a typical hard disk drive.&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
I don't think the onboard memory is practical for the average user because there's really no way for the additional available bandwidth or reduced latency to become significantly apparent.  I would load up my system with 4GB of ram and devote 3GB to a ramdrive I could use for encoding.  That could make use of the potential bandwidth, but I don't know very many people who'd need that.&lt;br /&gt;
&lt;br /&gt;
I was commenting on the fact that with current hard disk technology(especially with internal bandwidths), the onboard memory controller is premature.  It really has no essential use in current technology.  AMD deserves a pat on the back for trying, though.  They're the first CISC processor company to do that... RISC processors like the G5 already have it.&lt;br /&gt;
&lt;br /&gt;
But my last note on hard disk tech is that the internal bandwidth barely hits 70MBit/sec, while the external bandwidth is now peaking at 3GBit/sec.  Seems kinda lame if you ask me.... lazy hard drive manufacturers.&lt;br /&gt;
&lt;br /&gt;
</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233792">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-28T13:18:02-05:00</dc:date>
        <dc:creator>Lost Faith</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#233792</link>
        <description>ouch, this is what happens when you're installing games while surfing and posting on multiple forums and sites...&lt;br /&gt;
&lt;br /&gt;
sorry for me lagging the conversation...&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
HI SANDER!!!!&lt;br /&gt;
&lt;br /&gt;
*waves like a little girl seeing her favorite celebrety*&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
lmfao!&lt;br /&gt;
&lt;br /&gt;
*edit: spelling's a b***h*</description>
    </item>
    <item rdf:about="http://www.hardwareanalysis.com/content/topic/38545/?o=20#233789">
        <dc:format>text/html</dc:format>
        <dc:date>2005-01-28T13:11:38-05:00</dc:date>
        <dc:creator>Lost Faith</dc:creator>
        <title>Re: Re: Dual core CPUs, don’t get your hopes up</title>
        <link>http://www.hardwareanalysis.com/content/topic/38545/?o=20#233789</link>
        <description>&amp;quot;quote Sean B: Oh, and the hard drive goes through the southbridge too... which is kinda why AMD's integration of the memory controller doesn't make too much sense yet.&amp;quot; &lt;br /&gt;
&lt;br /&gt;
hmmm explain, cuz unlike your chipset's implication, I think (and I don't know that much about it) that HDD controllers on the northbridge with current hdd architecture and speeds is a little too premature.&lt;br /&gt;
&lt;br /&gt;
I think your solution would (only) be feasible when hdd reach levels of (slow) ramdisks like pc100 (non DDR) type disks.  &lt;br /&gt;
but for this era of mainly sata (150 &amp;amp; sataII 300) &amp;amp; ata (100/133) which is affordable for most (home) users&lt;br /&gt;
a successful implementation would not be possible right now.&lt;br /&gt;
&lt;br /&gt;
Explain the part why you think AMD's on-die memory controller doesn't make &amp;quot;much sense&amp;quot; and also what it has to do (why you compare?) with the HDD &amp;amp;southbridge solution.&lt;br /&gt;
&lt;br /&gt;
I'll be following events closely (while playing with my new toy 6600GT &lt;img src=&quot;http://media.hardwareanalysis.com/smilies/smile3.gif&quot; width=&quot;14&quot; height=&quot;14&quot; border=&quot;0&quot; alt=&quot;:D&quot; title=&quot;:D&quot;&gt;)&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
&amp;quot;quote LCF: And I remember looking at a diagram once... I could have sworn I saw the HDD shown being connected to the NB...&amp;quot;&lt;br /&gt;
&lt;br /&gt;
No really, (sata/pata)ide-bus is using pci based controllers (i don't know exactly how to state it...) and those are situated in the southbridge, remember raidcards etc also being pci-based? </description>
    </item>
</rdf:RDF>
