Intel says there's no such thing as PAT

[xxsk8er101xx]

"Rumours regarding Intel's Performance Accelerator Technology have surfaced all over enthusiast websites this week, with much speculation regarding possible performance increases when running PAT enabled chipsets. In a recent interview, Intel representative George Alfs stated that PAT is nothing more than marchitecture, and simply doesn't exist on i875 Canterwood chipsets. "it's not a technology at all - it just means Intel is using better silicon for the memory controller hub chips". This is a prime example of why the posting of rumours usually does nothing more than confuse readers, and is thus also the reason why you will rarely, if ever see any non-factual information posted at THG. "

LOL! DOH!


"Bread makes me poop!" - Special Ed

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>

 

[eden]

<A HREF="http://www.theinquirer.net/?article=9738" target="_new">http://www.theinquirer.net/?article=9738</A>

Looks like I was right, Canterwood owners are screwed, and not just that: Intel...........LIARS?!

And no I am not being sarcastic.

--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:

 

[simwiz2]

As long as it is giving some performance benefit, does it really matter if it is better silicon instead of an actual "technology"?

 

[Prof133]

Actually, TheInquirer went a step further by naming the article "Intel says there's no such thing as PAT." George Alfs didn't exactly say that. His explanation of PAT is given in the <A HREF="http://www.simhq.com/_technology/technology_001a.html" target="_new">SimHQ interview</A>:

<i>Performance Acceleration Technology is one of the features of our 875P (Canterwood) chipset. We basically bin our Memory Controller Hub (MCH) chips like our CPUs, finding the fastest silicon. We can then use this fast silicon to shave off a couple of memory clock cycles, resulting in better performance.</i>

<b><font color=blue>I agree with that solution ... because it is a logical one.</font color=blue></b>

 

[Guest]

Ya so they test the silicon, if it doesnt pass the test they cant put the 200$ and the 875 name, they cut a few bucks and name it 865...

 

[imgod2u]

Why is this method shocking? What do you think the difference between a 2.6 P4 and a 2.4 P4 is?

Anyway, as others pointed out, the Inquirer stepped way out of line and took the quote too far. Is anyone surprised that this came from the Tabloid of the tech news industry?

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

 

[Turk]

"We all deserve a PAT on our shoulder for being such geeks"

:lol: Turk

 

[xxsk8er101xx]

lol!

"Bread makes me poop!" - Special Ed

<A HREF="http://www.anandtech.com/mysystemrig.html?id=9933" target="_new"> My Rig </A>

 

[eden]

Then explain why THG's results were so tied?

Better quality, bah. Since when has silicon quality IN CHIPSETS become a damn issue? SO now we are THIS picky? Enough to pay 40$ more for a mainboard?

Ya right. That statement by Intel truly pissed me, I can't believe they would rip the consumer and mainboard manufacturers, to the point they would raise the board costs for the sake of having thought it did perform MUCH better.



--
If I could see the Matrix, I'd tell you I am only seeing 0s inside your head! :tongue:

 

[imgod2u]

I've not looked at THG's numbers. However, <A HREF="http://www.anandtech.com/chipsets/showdoc.html?i=1823&p=15" target="_new">Anandtech's</A> results were quite clear. <A HREF="http://www.lostcircuits.com/motherboard/canterwood2/" target="_new">Lost Circuit</A> did an extensive test of latency in the Canterwood chipsets and it's clear that the chipset has a latency-reduction.

Again, why pay the extra $100 or whatever for a 2.6 P4 instead of a 2.4? They're the same silicon, one is simply tested at a higher frequency. Chispet latency reduction may not affect performance as much as a 200MHz jump in processor speed but that doesn't change the fact that that latency reduction is there.

As for board costs, as I am writing, the Abit IC7 is $147 including shipping. Hardly the travesty you make it out to be compared to the $97 IS7. If you don't feel it's worthy it, don't buy it. Plenty of people don't think paying the extra $300 for a 3.0C P4 instead of a 2.4C P4 is worth it, but many do. After all, it's the same core, just different frequencies. Just like i875p is the same chipset as the i865pe, just with a more aggressive memory signaling. Saying it's not there is not only wrong it's irresponsible. Saying it's not worth it is a personal judgement.

"We are Microsoft, resistance is futile." - Bill Gates, 2015.

 

[ChipDeath]

If it's not AMD's weird rating system, it's intel inventing snappy acronyms for non-existent tech....
Can anyone be trusted?

Seems intel have tried the ol' "Fooling all of the people, all of the time" thing, and they've (unsuprisingly) been found out!



---
:tongue:

 

[slvr_phoenix]

I hate to have to be the one to <b>point out the obvious</b> but the 875 is intended for workstation use with ECC RAM and the 865 is intended for all of us SOHO junkies.

Generally, the 3rd party manus are pairing the 865 with all sorts of other 3rd party chips to save every last penny on their SOHO offerings. Meanwhile most of the 875s are being set up how Intel wants them, as workstation boards. So the point of getting the 875 is <i>not</i> for 'PAT', but for the other features such as ECC RAM and gigabit CSA LAN.

Really, only fifty bucks more for a workstation board is pretty good. And anyone <i>upset</i> by the fact that 'PAT' is just a slightly reduced latency binning on the northbridge needs to have their head examined.

<font color=purple><pre><b>There are 10 types of people in this world: those who can understand binary and those who can't.</b></pre><p></font color=purple>

 

[Mephistopheles]

Actually, I think the whole thing looks like one big misunderstanding. If some chips are of better quality, then you can run them at more agressive configurations, including "enabling PAT" - or using lower latencies and special features.

They do the same thing with celerons, I think... those P4s whose caches have troubles get one half of it (the troublesome half) disabled and turn into celerons...

 

[Guest]

Geek here.

Fact:

"And George answers by revealing it's not a technology at all – it just means Intel is using better silicon for the memory controller hub chips." - www.theinquirer.net

Are we THGC people getting things screwed up?

Silicon in transistors is "extrinsic silicon", this is a basic electronics fact.

"The difference is that an intrinsic semiconductor has only covalent bonds for all the atoms, but an extrinsic semiconductor also has free charges as a result of the doping." - A quote from a famous textbook

All he meant was that the MCH contains better quality extrinsic silicon.

All this speculation makes me itch, maybe somebody at THG could benchmark the standard (done) and give us a chipset vs chipset price comparison so we know how much the Mobo makers cash in on it.

 

[tombance]

Exactly. Just because its not a new technology doesnt mean it doesnt improve performance.

<A HREF="http://service.futuremark.com/compare?2k1=5467618 " target="_new">Almost Breaking 12k!!</A>