Forgot your password?
typodupeerror
Intel Hardware

Intel PAT Compared On 865PE Boards 117

Posted by CowboyNeal
from the motherboard-go-round dept.
Xaley writes "There is roundup at VR-Zone on several 865PE boards which came equipped with Intel PAT feature and definitely consumers will be happy to purchase these boards without paying a premium price that Intel has charged for their 875P chipset. A typical 865PE board costs around $150 and a 875P board costs around $200-250 so it is a cost savings of $100. Of course, Intel won't be too happy about it since sales of their 875P chipsets will be affected but there is no stopping for motherboard manufacturers from adding in the PAT feature into their 865PE boards now."
This discussion has been archived. No new comments can be posted.

Intel PAT Compared On 865PE Boards

Comments Filter:
  • Can't complain at more bang for buck. Now if only they made Xeons nice and cheap I would have a cluster of those :)

    Sorry

    Rus
  • It's PAT (Score:4, Funny)

    by Anonymous Coward on Saturday June 28, 2003 @11:16AM (#6319396)
    A lot of people say, "What's that?"

    It's PAT.
  • Ugh. (Score:5, Informative)

    by be-fan (61476) on Saturday June 28, 2003 @11:16AM (#6319400)
    Intel should stop reusing acronyms. PAT = Page Attribute Table, a feature in PIII+ processors that helps manage caching of memory ranges.
    • I read it as Port Address Translation. Expansion of acronyms would be helpful in the original post. -Chuck
    • Re:Ugh. (Score:4, Informative)

      by InnovativeCX (538638) on Saturday June 28, 2003 @11:36AM (#6319489)

      The above description was slightly vague, so I went out and did a googling for those that would like a little more info. A couple USENET [google.com] posts [google.com] turned up the following:

      This feature provides more flexible interface allowing to setup various memory cache modes on a page-by-page bases. It is much easier to program than MTRRs
      and does not suffer from their size/alignment limitations.

      and

      The Page Attribute Table (PAT) is an extension to the x86 page table format that lets you enable Write Combining on a per-page basis. Details can be found in chapter 9.13 of the Intel Architecture Software Developer's Manual, Volume 3
      (System Programming).
    • Oh, I'm sure PZF and JLQ are both free. Now lets just find three maketing speak words that fits...
    • agreed, but all this means is Performance Accelorated Technology. To me, that just means marketing hype.
    • Re:Ugh. (Score:2, Interesting)

      Are you suffering from TOS?

      (TLA Overload Syndrome).

      "Our ATM backbone provides the bandwidth to improve the legibility of a bank's ATM display by downloading either Truetype or ATM fonts".

    • Hello (P)artition (A)llocation (T)able too!
  • No stopping? (Score:3, Informative)

    by Quixote (154172) on Saturday June 28, 2003 @11:17AM (#6319403) Homepage Journal
    Don't count Chipzilla [intel.com] to just roll over and play dead. They have already warned [theinquirer.net] MoBo manufacturers not to turn the i865 into the i875. I wonder how many will heed the warning? Its not as if there are major alternatives out there (note, I said "major").

    • Don't count Chipzilla to just roll over and play dead. They have already warned MoBo manufacturers not to turn the i865 into the i875. I wonder how many will heed the warning? Its not as if there are major alternatives out there (note, I said "major").

      Don't count mObO manufacturers to just roll over and play dead. They have already warned Chipzilla not to be pissy about them turning the i865 into the i875. It's not as if there isn't another processor out there(note, I said "AMD").

  • by SuperBanana (662181) on Saturday June 28, 2003 @11:18AM (#6319406)

    Funniest quote from the whole article:

    There are so many good marketing specs thrown into 875P that make the hardware enthusiasts to go after it and even prepare to pay a premium for it.

    I take back all those nasty comments about hole-in-the-wall-2-guys-in-a-dorm-room-'review' sites. Well, okay, not all of them- even this guy has some SERIOUS grammar problems :-)

  • by Brian Stretch (5304) * on Saturday June 28, 2003 @11:21AM (#6319421)
    since AMD doesn't harrass/sue third-party chipset suppliers (like VIA) so no one can get away with playing games like Intel is doing with PAT. I prefer nVidia nForce2 chipset boards these days. YMMV.
    • From the parents sig..

      MEMRI [memri.org] translates Middle Eastern news and commentary into English.


      http://www.guardian.co.uk/elsewhere/journalist/s to ry/0,7792,773258,00.html

      Its a propaganda machine, along with the ADL.

    • The nForce2 is a nice chipset but their Linux drivers leave much to be desired, like working audio and firewire.
      • The nForce2 is a nice chipset but their Linux drivers leave much to be desired, like working audio and firewire.

        I've been using ALSA for audio. Freshrpms.net [freshrpms.net] has ALSA (and other goodies) in convenient RPMs. I haven't tried using the FireWire port but everything else works. And it's trivial to run 400MHz FSB, regardless of your CPUs official spec. I'm running a 1700+ at 7.5x404MHz. MemTest86 [memtest86.com] verified the configuration.
        • Very cool. :)

          The firewire port is partially working, it just doesn't detect any of MY devices. But its either because I'm running a newer kernel and compiled their src rpm or maybe I just didn't try hard enough. Anyway, its very nice hardware. I hope the drivers become more stable soon so I can order some of this equipment for my dad and sis.
  • by VCAGuy (660954) on Saturday June 28, 2003 @11:23AM (#6319430)
    ...what'll stop the overclockers/enthusiasts from turning on PAT? Won't that still eat into their sales of 875P based boards? Somehow, the fear of voiding our warranty never enters the equation...so what's the deterrent?
    • I doubt that Intel gives a damn about "overclockers/enthusiasts". They are a tiny part of the computer market. They may be more concerned about motherboard and computer vendors who sell computers with unsupported chipset configurations.
  • But would anyone be interested in defining PAT so that the ignorant among us can decide if this story is interesting?

    (Yeah yeah, rtfa. I will when I find out what PAT is, and then if I decide it's useful. ;-) )

    • That's one of my pet peeves. Slashdot tends to do that a lot - people throw out acronyms from specialized fields and assume everyone knows what they're talking about. To me, PAT is a Point After Touchdown.

      Btw, I don't think you're lazy, or no more than most. Checking Slashdot is a good way to cover lots of interesting topics without having to go look at dozens of other sites. If you've got to go read the article to even know what the subject is, well that's just annoying.
    • What I gleaned from Tom's Hardware:

      PAT stands for "Performance Accelerating Technology" and is a CPU/chipset feature available only with Intel's high-end (consumer) chipset. It speeds up memory accesses by cutting out a clock cycle from the CPU/memory controller memory access request, and another clock cycle on the memory controller/DRAM interface. So DRAM accesses are two cycles earlier than they would otherwise be. I think the feature relies on asynchronous timing to eliminate the clock cycles.

      The fea

  • by JRHelgeson (576325) on Saturday June 28, 2003 @11:32AM (#6319466) Homepage Journal
    PAT=Performance Acceleration Technology
    The 875P Canterwood chipset with PAT supposedly provides a 5-7% performance improvement over 865 chipset.

    WOO HOO!

    TMFA's! When will we start patenting acronyms?
    PAT=
    Port Address Translation
    Page Attribute Table
    and now Performance Acceleration Technology

    • And for those wondering how to activate it in 865PE, there's an article [tomshardware.com] at Tom's Hardware.
      • I really wouldn't recommend buying one of these Asus boards where they hacked the 865.

        As I understand it, there are 2 signals from the CPU to the northbridge that dictate what the FSB frequency is. The 875 northbridge takes these signals, and if they show that the FSB is 800MHz then PAT is enabled. The 865 northbridge is configured to disable PAT with an 800 MHz FSB.

        Now, what Asus has done is to hijack these 2 signals coming from the CPU, and always tell the northbridge that the FSB is 533 MHz. This al
        • Well, it's not compulsory to use their version of PAT. It's enabled or disabled in the BIOS. I don't recall the default.

          Besides, don't you think preaching against overclocking doesn't suit most users of this forum? :)

          • The problem that I see is that Asus is not really disclosing that they are overclocking the northbridge, and that what they are doing could affect the system's stability. They are merely saying that they have enabled PAT. So not only are they enabling it on chips that Intel said didn't pass the test at the fab, they are overclocking the chip to exaggerate any problems. It just doesnt seem smart to me.

            Of course, you are free to do whatever you would like to with your chipset, but I think I want my memory
        • 5-7% may not seem huge, but consider that if you bought a 3GHz processor and added 7% that it's now running at 3.2GHz.

          With the 3GHz priced at $417 and the 3.2GHz priced at $637, you are essentially saving $220 with this change, not to mention the $100 savings over the 875 board.
          • 5-7% may not seem huge, but consider that if you bought a 3GHz processor and added 7% that it's now running at 3.2GHz.

            With the 3GHz priced at $417 and the 3.2GHz priced at $637, you are essentially saving $220 with this change, not to mention the $100 savings over the 875 board.

            Actually, the 875 board gives you that 5-7% increase by itself. This Asus hack just adds the same 5-7% speed increase to their 865 board. So you're only saving on the price difference between the 875 and 865, which is more like $

    • Yup, that's NDAs for you.

      [NDA: Non-Deterministic Abbreviation]
    • Point After Touchdown
  • from what i've seen (tomshardware et al), 875's w/ PAT were only marginally out performing the 865's anyway.

    has anyone seen where this stuff significantly inproves performance?

    nonetheless, i'm waiting for the P5/Prescott before upgrading the ol' PC anyway...

    • by Anonymous Coward
      from what i've seen (tomshardware et al), 875's w/ PAT were only marginally out performing the 865's anyway.

      Your geek liscence has been revoked, sorry. Performance at any cost is rule #1 in the GeekDex.
  • Prices (Score:4, Informative)

    by MC68040 (462186) <henric@noSpAm.digital-bless.com> on Saturday June 28, 2003 @11:48AM (#6319536) Homepage
    200-250 is not correct in many cases for the P boards, I just got a MSI 875P Neo board for around 150 and it has all the goodies like SATA and got excellent test results.
    • I recently bought the MSI Neo2-FISR (865 chipset) and its a really solid board. I'm quite impressed with it. Never thought I'd end up with an MSI board....

      Stranger things have happened I suppose!
    • Yes, the story author's prices were about $50-$100 above what they actually cost when purchased from a reasonable retailed (i.e. newegg, gameve etc.). The Abit IC7, regarded as one of the top i875 boards, can be found for $135. The IS7, an excellent i865 mobo can be found for about $100.

      P.S. This isn't a troll, but it seems that as much as most slashdotters know about linux, coding, and the like, they often seem to be a bit off on computer hardware. Of course they're by all means computer-literate, but
      • Yea, although you should be aware that the 875P NEO FISR-2 got the best benchmark results in a recent toms hardware roundup =).
    • Those boards are excellent. I'm using one right now and it's probably the best system I've ever owned.
  • What Intel should have done is to have "PAT" disabled through a physically inaccessable method, such as a different IC package, bondout (though this is probably flip-chip) or an on-chip fuse burnt at test.

    • Most of the 865 chips are basically rejected 875 chips. Intel didn't want to waste the chips so they came up with the idea of just renaming them 865 and sell them as lower performing chips.

      Your right in that intel should have done something to permanently disable the chips.
  • by rmdyer (267137) on Saturday June 28, 2003 @11:56AM (#6319574)
    "There is a constant pressure from Intel to advise motherboard manufacturers not to offer PAT into their 865PE boards but on the other hand, there is a considerable amount of pressure from consumers demanding PAT in their 865PE boards. There is a great dilemma faced by the motherboard manufacturers here whether to include the PAT feature or not. However to keep the sales going, it is advisable add in PAT into their 865PE boards to keep the consumers happy."

    All of this worry over $100 bucks difference? That amounts to about 1 days work for most of us. We are more lazy than I ever suspected! Arguably, this is what leads to early psychosis. ;-)

    +2 cents.
    • It may be less than a day's work difference in price, but that doesn't mean it's insignificant. Imagine if someone offered to sell you a candybar for $100. Which would you say: "Gee, that's less than a day's pay." or "You've got to be kidding."

      If you're paying $150 for a motherboard without PAT and $250 for one with PAT, that's a 66% markup for 7% performance increase. Not all that impressive of a deal.

    • some of us arent so rich (read : university student) . If I can save $100 thats a big deal to me . As for days of work , once that 100us becomes cad (hey I'm canadain no laughing) that works out to about three days of work give or take for me.
    • After rent/mortgage, food, car, and everything else, most people probably have about 20-40 days' worth of disposable income to spend every year. If you want to toss that away one day's worth at a time, be my guest, but don't be surprised if someone finds 1 day's pay to be a significant amount to worry about.

      Also, you're talking about end-customers that will be buying just one of these things. But your quote was about Intel (who will sell millions of them) and mobo manufacturers (who will buy millions of

  • Word out of Intel is that there is a future revision of the 865 chipset will have PAT disabled in hardware before being shipped to motherboard manufacturers. The system will just lock up if they try to enable it.

    So keep in mind this situation is temporary.
  • So Intel released hardware with features disabled...hmm sounds like their usual [i487] tactics...

    Boo Intel...

  • Commercial Viability (Score:5, Informative)

    by Detritus (11846) on Saturday June 28, 2003 @12:51PM (#6319838) Homepage
    This feature is useless for business applications where reliability is essential. While the PAT logic is in the 865 chipset, Intel does not guarantee that it has passed chip-level testing. It may work, it may not work, it may only work over a limited temperature range. It falls into the same category as overclocking the CPU. You don't know if the chip passed the test for a higher classification and was marked for a lower speed to satisfy market demand, or if it failed the test for reliable operation at the higher classification. If you want that guarantee, you pay the premium for the 875P.
    • From the article [tomshardware.com] at Tom's Hardware:

      As is already well known, Springdale and Canterwood chipsets are completely identical during wafer production. Intel in fact makes a preselection, which then leads to different packaging. Intel explains this step with the following: only when a chip does not pass the PAT test does it become the lesser-value Springdale (provided that all other quality tests have been passed).

      From the beginning, Intel put safety precautions into place that would ensure that the board ma

      • As I posted here [slashdot.org], Asus's solution to enable PAT in the 865 involves overclocking the northbridge, so not only are you trying to enable PAT on a chip that did not pass the tests at the fab, you are running the chip way out of spec. Not good.
  • Following in the footsteps of well-known (but not as well-known as he thinks he is) director Spike Lee suing over Spike TV, well-known television evangelist Pat Robertson has filed suit against Intel Corporation regarding the use of the acronym PAT.

    Robertson is quoted as saying: In order to squeeze this much additional performance out of the mature PC platform and then call it PAT, people might believe that I have prayed for divine intervention on the part of Intel Corporation. This is not the case. Des

  • A lot of posters have stated that users could add PAT technology on their own as simply as they would overclock a chip. Unfortunately, this isn't quite true. The hack that Asus originally developed, and which other MotherB manufacturers are starting to copy is rather complex. ASUS realized that when the MotherB was functioning at a reduced FSB speed (e.g. not 800MHz), the PAT technology was working, however due to the reduced bus speed, the performance was lacking. So, what they had to do was first to f

It is impossible to travel faster than light, and certainly not desirable, as one's hat keeps blowing off. -- Woody Allen

Working...