Prescott OK on current mobos.

Mephistopheles · Jul 13, 2003

Intel specced all voltage requirements for Prescott over two months ago - back in April.

So what is going on here? None of the Specs changed. The <A HREF="http://www.xbitlabs.com/news/mainboards/display/20030712011416.html" target="_new">Xbitlabs</A> news bit that mentioned incompatibility even mentioned that mobo manufacturers were not adhering to VRD 10 specs.

So in what way is this Intel's fault? This is not another MTH case, is it?

Mephistopheles

Crashman · Jul 13, 2003

Hehe, Tom's mentioned this a couple months when the first boards came out, saying that Intel's recommendations for the new spec were simply that, recommendations, as far as the 865/875 chipset itself goes. It was up to the manufacturer weather they wanted to comply, which would ensure Prescott compatability, or not. Look up the earliest article Tom's published on these chipsets, you'll probably find the info there.

So we know that at least 1 company, Asus, claims they DID follow Intel's recomendations. Which means X-Bit Labs published BOGUS INFORMATION when they made the following statement:

In fact, mainboard makers have known that their current products do not support Prescott chips for weeks now, so, it is a mystery why, for example, ASUS still claims that their i865PE-based mainboard supports Prescott processor.

That's right, they are deceiving people. Why, I don't know. It sounds like someone is serious set out to sabatoge this part of the industry.

Watts mean squat if you don't have quality!

sirak · Jul 13, 2003

Mine's an ASUS, so let's hope they are compatible.

Mephistopheles · Jul 13, 2003

It most probably is. Intel has released the eletrical specs for prescott a long time ago.

Mephistopheles

Mephistopheles · Jul 13, 2003

That's right, they are deceiving people. Why, I don't know. It sounds like someone is serious set out to sabatoge this part of the industry.

Well, the only thing we can do is try to spread the news around...

And I also hope that THG doesn't do the same kind of mischief. It would be a disappointment. :frown:

And as an update: <A HREF="http://www.hardocp.com" target="_new">HardOCP</A> is being much more accurate: they said that Prescott support is entirely up to the manufacturer. This is correct. Intel didn't make any last minute changes: the electrical specs for prescott are common knowledge since April (this actually means WE can check out Intel's developer site personally).

Intel did everything right there. The mobo makers just have to follow. But, of course, lots of people were just looking for this opportunity to flame Intel.

Maybe... Just maybe the mobo manufacturers wanted to sell much more mobos?... Now from where did I get that crazy idea? If they don't follow the specs now, they can always release yet another mobo revision...

Mephistopheles

Crashman · Jul 13, 2003

If it isn't, Asus will have to repair the board for free under warranty. Yes I said repair. You see, I've seen Asus repair older boards to meet specifications for newer versions of the same board. The most famous was the P2B-DS, which Asus actually published the repair info on their website for.

Watts mean squat if you don't have quality!

Crashman · Jul 13, 2003

I remember the last time such a rumor hit, Tom's waited...days, I think it was about a week before mentioning it. And then they wrote a brief editorial that said (roughly) "this is why our site can be considered a more reliable source of information".

Watts mean squat if you don't have quality!

Mephistopheles · Jul 13, 2003

"this is why our site can be considered a more reliable source of information".

Impressive. I hope they do that now too. Actually... I didn't know anandtech and xbitlabs would both spread this kind of rumor without checking. I'd have thought better of them. Good thing I rely mostly on THG.

Mephistopheles

POPEGOLDX · Jul 13, 2003

<<lots of people were just looking for this opportunity to flame Intel. >>

with the way people INTEL ELITIST flame AMD... would u expect any less in return?

POPEGOLDX · Jul 13, 2003

<<Good thing I rely mostly on THG>>

what article did u read that showed prescotts in current motherboards and they worked? please post a link. ALl i read here is counter conjecture. We have no clue if INTEL had to change anything since april. I mean you think going to .90 um was going to be flawless and easy just becuase its INTEL.

sheeshh

Crashman · Jul 13, 2003

Well, that was a paraphrase, not a direct quote. Anyway, it's possible that Anandtech got what they though was reliable information from what they thought was a reliable source, and then X-Bit labs simply assumed that the info Anandtech got was right...this is the type of thing that caused the Korean stock market to crash a few months ago!

Watts mean squat if you don't have quality!

HardWareBoss · Jul 13, 2003

They got their info from some turkey that is a big fan of AMD. Why would the top Mobo builders state two plus months ago that their motherboards would support Scotty. IF indeed they did not!!!! That too me is bad business. Which would make no sense. Talk about false advertisement.

NMDante · Jul 13, 2003

So, unnamed manufactuerers are a better source, than a named one?

Prescott CPUs in current (ie. 865/875) mobos are the proof you need that it will work? Here's an easy answer...keep the box and any paperwork that came with a current 865/875 mobo that states that it will accept a Prescott CPU...and if it doesn't, the manufactuer HAS to give you one that does...or you can sue them for false advertising.

Also, 90nm processes was never a flawless endeavour, for anyone, but I bet intel is much closer to finalizing it's 90nm than AMD, only because they have the capacity to tweak it, than AMD does.

But, since none of this will really matter...Show me a manufactuer link that claims that the current 865/875 motherboards WON'T support Prescott. Not just there's, but the whole chipset line.

How many watts does it take to get the center of CPU core?

Crashman · Jul 13, 2003

Who knows though, maybe the A64 will be a great processor after all, and those who bought the current P4 platforms will feel stupid, not because they didn't support newer CPU's, but rather because they wished they had waited for the A64? This would be a good way of getting people to wait.

AHA! Isn't the time when those "revision 2.0" boards "required" for Prescotts...the same time when AMD was supposed to release the A64? Coincidence?

Watts mean squat if you don't have quality!

HardWareBoss · Jul 13, 2003

You make a great point there. So IBM did not buy AMD Intel did. So start a rumor that Prescott will not work on 865,875 Mobos. Buy A64 when that comes out. Now do the Mobos that are on the market for AMD do they support A64?

amdchooser1 · Jul 13, 2003

No,
It is very clear that the current Amd mobo's dont support A64.Youre talking about a big architecture and socket change.

Crashman · Jul 13, 2003

FLAME BAIT! Wow! You really set yourself up for that one, you need a fire extinguisher, or will Preparation H do?

Watts mean squat if you don't have quality!

sabbath1 · Jul 13, 2003

My guess is that the Prescott will be slightly better performing than the Athlon 64 at 32-bit. Most probably will the "low-end" 1.3 Ghz Itanium 2 perform better than the Athlon 64, while Opteron would probably cut things close even with the 1.5 Ghz Itanium 2. Thus I believe the Intel CPU's will perform a better overall than the competing AMD processors, but AMD on the other hand, has the compatibility with both 32 and 64-bit programs to fall back on.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI 875P Neo / Sapphire Radeon 9800 Pro / Antec True Power 550W / Western Digital Raptor / Hercules G.T XP /
Samsung DVD / Lite-On CDRW

confoundicator · Jul 13, 2003

I think that's the big reason that AMD is getting the deal with IBM. Full 64 and 32 bit functionality. Now all they have to do is convince the business community that they need it.

Insanity is doing the same thing over and over again and expecting a different result.
-Einstein

Mephistopheles · Jul 13, 2003

while Opteron would probably cut things close even with the 1.5 Ghz Itanium 2.

There seems to be no indication of that. A 1.5Ghz Itanium scores 1300/2100 int/fp SPEC CPU, and a 1.8Ghz Opteron scores 1150/1200. That is very different. And, because of its design, I can't imagine Opteron will easily approach the 1.5Ghz Itanium.

One more thing, though: by the end of the year, we'll have the 1.67Ghz Itanium, according to roadmaps. That's more than a 10% increase. <speculation>Note that 1667Mhz is not a multiple of 100Mhz. Itanium uses a QDR bus. Maybe they'll raise it to a 4x133=533 FSB?... Which means 8.5GB/s bandwidth, more than HT.</speculation> This will be out this year. Opteron will not compete well with that at all, unless the 2.2Ghz or maybe the 2.4Ghz gets out the door.

Then again, Opteron would run 32-bit code considerably better than Itanium. I don't know exactly by how much, but it would. *reviews, please, THG...* *workstation reviews* *please*...

Mephistopheles

Mephistopheles · Jul 13, 2003

I'm very sorry, but I did not flame AMD at all. I didn't even say anything about AMD fanboys. Take it easy, popegoldx. Take it easy.

Mephistopheles

dwellman · Jul 13, 2003

AMD and IBM have formed somewhat of a strategic partenership. Benefits AMD via source for R&D $$$, which AMD lacks-- especially compared to Intel. Also, IBM is the major player in SOI technology (really cool thing). I'm not sure what the benefit to IBM is except a cut of the sales or benefit from AMD's marketing collosus.

__________________________________________________
Three great virtues of a programmer are: laziness, impatience, and hubris.

sabbath1 · Jul 13, 2003

Aren't the new Opteron 800 series a lot faster than the 200 series? I think I saw a few benchmark somewhere where the Opteron 244 came quite close to the current high-end Itanium 1.0, but maybe those benches were some of the few that Opteron would conquer in. Anyways, based on those benches I *assumed* that the 800 series would be somewhat competitive with the Itanium 2. However I had no doubt about the Itanium 2 being faster.

Well, no doubt, the Opteron performs better than the Itanium 2 in most if not all 32-bit programs, currently. Since all the Itanium 2 CPU can do is to emulate the support for 32-bit programs, it doesn't stand much of a chance either. But it doesn't need to be, since the Xeon processors would do that instead.

Then, if the buyers prefers best performance while having to pay for two machines, one P4 or Xeon, and the other with Itanium 2 (no longer have to be that extraordinary expensive anymore), or if they prefer best compatibility at a lower price and with a little to a lot worse performance, then that would be in my mind, the deciding factors to which of the Itanium 2 or the Opteron that would fare best in the mainstream and high-end server business.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI *argh* / Sapphire Radeon 9800 Pro / Antec True Power 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

Mephistopheles · Jul 13, 2003

current high-end Itanium 1.0

No, the current high-end Itanium 2 is the 1.5Ghz, 6MB cache Madison. The low-end is the 1.3Ghz, 3MB cache Madison @ $1200. They are excellent performers.

Also, if I undersand it correctly, there is no difference between the 800 series and the 200 series. They merely support a different number of processors on one system - but they're the same core. Therefore, performance levels can't quite change. One great advantage that AMD has is that the end user can choose which CPU price he pays, depending on how many cpus he wants in his systems... it's a more custom-tailored purchase... Deerfield might change that.

Also, Itanium is a considerable departure from usual design. It can, in its own playing field, perform admirably - and at pathetically low clock speeds for our days.

Well, no doubt, the Opteron performs better than the Itanium 2 in most if not all 32-bit programs, currently. Since all the Itanium 2 CPU can do is to emulate the support for 32-bit programs, it doesn't stand much of a chance either. But it doesn't need to be, since the Xeon processors would do that instead.

No doubt about that. Intel says they streamlined 32-bit emulation a bit in this new Win2003 service pack, but that remains to be seen. (and anyway, because of Itanium's low clock, it'll have a bit of a hard time in 32-bit)

Mephistopheles

sabbath1 · Jul 13, 2003

No, the current high-end Itanium 2 is the 1.5Ghz, 6MB cache Madison. The low-end is the 1.3Ghz, 3MB cache Madison @ $1200. They are excellent performers.

Sorry, I meant the 1.0 GHz Itanium 2 with 3 mb L3 cache that just recently got replaced by these two new CPU's; the ones you mentioned. But since they were just released, I call them the "new" Itanium 2's.
$1200 is not too extreme for being an Itanium 2 CPU. That's just a few hundred dollars more than a high-end P4. Hence to why I think it's getting quite affordable. Hey, if more 64-bit programs would come out, I'd personally probably start saving up for an Itanium 2.

Any word on Deerfield pricing? (Or whatever the slowest new Itanium 2 is called.)

Also, if I undersand it correctly, there is no difference between the 800 series and the 200 series. They merely support a different number of processors on one system - but they're the same core. Therefore, performance levels can't quite change. One great advantage that AMD has is that the end user can choose which CPU price he pays, depending on how many cpus he wants in his systems... it's a more custom-tailored purchase... Deerfield might change that.

Oh I didn't know that, thanks for the info. You're right then, the 800 series should not, at least not with the same number of CPU's, perform better than how the 200 series does.

Unrelatedly, because of the integrated mem controller, a 4-way Opteron should have something like twice the bandwidth than what a 2-way Opteron would have.

But...Itanium 2 is the CPU God.

My system: Intel Pentium 4 3.0, 800FSB / TwinMOS 1Gb DDR400 / MSI *argh* / Sapphire Radeon 9800 Pro / Antec True Power 550W / 2x Western Digital Raptor / Hercules G.T XP /
SamsungDVD / Lite-On CDRW

Prescott OK on current mobos.

Distinguished

Polypheme

Distinguished

Distinguished

Distinguished

Polypheme

Polypheme

Distinguished

Distinguished

Distinguished

Polypheme

Splendid

Distinguished

Polypheme

Splendid

Distinguished

Polypheme

Distinguished

Distinguished

Distinguished

Distinguished

Splendid

Distinguished

Distinguished

Distinguished

Share this page