Bookmark and Share


Even though Intel Corp. is set to launch its next-generation dual-core low-power Yonah processor for mobile computers early next year, its brother, which is believed to have similar, or the same, micro-architecture is slated for release later in the second half of next year. While Intel’s partner Supermicro shows performance results of the chip, it does not demonstrate it up and running.

According to performance slides allegedly demonstrated by Supermicro at a Japanese event, dual-processor setup running two Intel Xeon DP processors code-named Sossaman at 2.0GHz was about 20% faster compared to a system featuring two Intel Xeon 3.60GHz processors with 1MB L2 cache, but was outperformed by about 27% by 2-way server with Intel dual-core Xeon 2.80GHz in Whetstone floating point unit (FPU) benchmark. Still, Sossaman chips seem to be slightly better performers when it comes to arithmetic logic unit (ALU) performance. 

The main advantage low-power Xeon processors based on the Sossaman micro-architecture will provide will be low power consumption, which is crucial for thin blade servers.

According to web-sites which gather unofficial information about yet-to-be released products, the Sossaman is to be based on the dual-core Yonah micro-architecture and is expected to be compatible with current Intel Xeon DP chipsets, such as Intel E7520. Still due to different form-factor, the new Xeon DP chips will require separate infrastructure. The chip is expected to have power consumption of 31W when working at about 2.00GHz, whereas its low-voltage brother is likely to consume approximately 15W when operating at 1.67GHz. By contrast, current dual-core Xeon DP chips at 2.80GHz consume up to 135W in typical conditions. 

Intel says that a lot of server deployments these days require low power consumption. In order to address that market, the company recently validated Intel Pentium M processors for its server platforms and also supplies Low-Voltage Intel Xeon chips with thermal design power (TDP) of 30W and 55W. Processors with 30W consumption based on the NetBurst architecture operate at relatively low clock-speed and may not offer performance, which would satisfy clients.

According to a report from Akiba PC Hotline web-site, Supermicro is developing at least two mainboards – X6DLP-4G2 and X6DLP-EG2 – designed for the next-generation Intel Xeon Sossaman processors and based on Intel E7520 memory controller hub (MCH) as well as Intel 6300ESB I/O controller hub (ICH). It is unclear, however, when the mainboards are expected to be available for purchase.


Comments currently: 17
Discussion started: 12/12/05 08:32:43 PM
Latest comment: 02/07/06 06:53:10 AM
Expand all threads | Collapse all threads


I guess they finally have a product that will replace the Pentium III-S 1.4 GHz, still the best processor for most servers out there.

These processors still sell for nearly $200 new on eBay, if you can find them. New motherboards like the P3TDE6 still go for over $400 on eBay, almost $600 on other internet sites.

It took Intel forever, but it looks like they finally understand the market. Amazing how they could be so wrong for so long with all the money they have to study the market and all the customer information they have. Why did the Pentium 4 plague the market for so long when they always had better processors available (Pentium III-S and Pentium M)? A question we'll probably never know the answer to.

Next year will be interesting for AMD as it seems they will once again be in the position of selling inferior processors to Intel's. It could represent the worst situation they have been in, since their processors will probably not only be larger, be less power efficient, but also will probably show less integer performance or, at best achieve parity despite having a on-board memory controller and all the compromises that entails. Floating point may be an advantage, but is not useful as integer to nearly as large a crowd, and that crowd will find the Itanium 2 far superior to any x86 processor, particularly next year.

It seems to me the Athlon 64 next year will look very old and very vulnerable, and AMD does not seem to have anything new imminent. Intel can be caught flat-footed and still survive, even with the horrid P7. But, AMD can not. They better pray Intel doesn't execute well, because if they do, they are done. Intel's huge advantages in marketing, manufacturing, size, exclusive relationship (so far) with Dell, etc... will slaughter AMD if their processor is better. AMD can only be profitable when their processor is superior, and even then not always. It all depends on the P6++.

I think an appropriate advertising scheme for the P6++ will be something on the Back To The Future theme. Amazing that a processor released in 1995 still represents very close to the optimal design, considering how fast technology changes.
0 0 [Posted by:  | Date: 12/12/05 08:32:43 PM]
- collapse thread

I read Intel Sossaman Slower than current Xeon Chips.

Did you read something else?

Also you think is good bring old tech? I think you are one of those that think intel should release some processor with 64 - 386 cores.
0 0 [Posted by:  | Date: 12/13/05 05:50:40 AM]

The article was, as usual, mistitled.

If you read closely, the new processor will have better integer performance. Mind you, this is not even the Merom. This is dual Yonah.

For most people, integer is more important and always has been. If it isn't, an Itanium 2 makes any x86 look kind of silly, and probably would make a better choice for many people needing floating point.
0 0 [Posted by:  | Date: 12/13/05 02:32:52 PM]
Well Itanium doesn’t run the same application that a Xeon, Opteron runs... so what you say is point less.

You have some processor that as more than 10X the processing power of the best x86 processor. But it doesn’t do (run) anything.

You also have a good example of the GPU, they say is much faster doing this and that but I don’t see a thing.
0 0 [Posted by:  | Date: 12/14/05 03:19:08 AM]

TA152H is a complete idiot. He trolls these message boards frequently. I own both AMD and Intel, and let me tell you something you half-baked moron, you have absolutely no idea what you're talking about. Saying a P-III 1.4S is the best processor for most servers is one of the most ridiculous statements I've heard in a long time. You know nothing of the cpu market, and you're entire post is based on what? .....nothing, that's right. Get a clue dude, you're posts are embarrassing and uneducated. You are one of those many people who think they know everything, when in fact as shown here, you know nothing.
0 0 [Posted by:  | Date: 12/13/05 09:43:07 AM]
- collapse thread


Your just a school kid that doesn't know anything about computers.

Pentium III-S are still very useful parts, especially in dual processor platforms. They are very power efficient and most servers do not need the processing power of the newest processors. File servers almost never do.

But again, you are just an internet dork that doesn't know a damn thing about real life uses. I have more servers than you have brain cells, and believe it or not some are still using Slot 1 Coppermines and do their job more than adequately. Why would I want something using 80 watts when I don't tax the processor of my 30 watt processor? Typical idiot with no real life experience.
0 0 [Posted by:  | Date: 12/13/05 02:30:49 PM]
I hate to admit it but he is right.

I have P3 550, P3 800 and P3 1.4 GHz and all run fine, even today.
My last P4 820 server however: lots of noise, power consuming and heat (very bad). Only because of 64 bit (I don’t use yet), and more than 4GB ram (I don’t need yet), is not worth it, I would go for Opteron, but where is it (at least in my country)?

For a 24/7 server nothing like the P3, unless you need super performance for intensive aplications....
0 0 [Posted by:  | Date: 12/14/05 03:12:07 AM]

TA152H says:
"...New motherboards like the P3TDE6 still go for over $400 on eBay, almost $600 on other internet sites."

You are lying. I got two of these for less than 1/5th the price you're saying. And that included shipping! (Brand new, in the blue Supermicro retail boxes).

TA152H says:
"...It took Intel forever, but it looks like they finally understand the market. Amazing how they could be so wrong for so long with all the money they have to study the market and all the customer information they have. Why did the Pentium 4 plague the market for so long when they always had better processors available (Pentium III-S and Pentium M)? A question we'll probably never know the answer to..."

WTF are you talking about?

The P4 has been around for some time, because it scales well with speed. To the average person, who still looks at speed, Intel CPUs look great. Speed sells and Intel wanted to seize on this customer thinking. (As long as it looks like its faster than the competition, no one will know the better.)

But the truth is, when you spend billions on R&D for a product, you need to squeeze as much out of it back in profits. This is why the P4 solution has been hanging around. And the refusal for Intel to drop it and hold onto it as long as possible.

The reason they started to consider Pentium-M as the replacement was because heat became a very serious issue. It became a critical problem when P4s hit above 4Ghz. If they continue the scale up in speeds, it will no longer be practical to adequately cool a CPU with aircooling solutions. You would start using liquid solutions. (Which will jack up the cost for the customer).

They even tried to introduce a new case format called BTX. At the moment, very few companies have bothered with BTX, and are mostly sticking to ATX. (Intel wanted everyone to change to BTX, so it would help Intel with its P4 "heat issues".)

To top that off, back in 2004, the 90nm issue scared the shit out of all manufacturers. Everyone (AMD/IBM/Intel/etc) were having serious thermal issues with leakage and the 90nm process. Moving from 130nm to 90nm, resulted in the thing falling apart on you.

Intel used the "slapstick" approach to delay this issue with SpeedStep power saving technology, while both AMD and IBM dared not to release anything faster because going above it will result in similar issues as Intel. AMD had to wait for IBM to refine and mature the 90nm SOI process. (Which is seen in today's Rev E6 CPUs).

Why do you think 2004 was a dull year for CPUs?


(a) Intel didn't spend a dime on finding out what the customer wants. They are forced to dump the P4 solution because they had no choice. (You can't increase the speed at a certain point because it gets too hot. And if you can't increase speed, you can't compete against AMD's solutions).

The were forced to change their thinking...From high speeds, to more cores, features and being thermally efficient.

The only choice they had was the Pentium-M. Currently, the Pentium-M does not support EM64T and it has a weaker FPU than AMD's solutions. (Yonah and Sossaman will have the same weaknesses).

(b) If you bothered to read Anandtech's Yonah article, you'll know that the A64 and Pentium-M are, overall, pretty much the same in performance. Pentium-M may win some, A64 wins others. The "old" A64 still whips the Intel solution in FPU intensive applications. So I don't know WTF you're BS'ing about with A64 being "vulnerable".

How is the K8 inferior to Intel's solution, when they both perform about the same, and Intel's one has a weaker FPU and no EM64T/AMD64 capability?

(c) The AMD K9 will NOT be coming next year. Why? 2006 is when AMD switches formats for all their CPUs (Socket M2, F, etc). Its a transition period. Why would AMD make it more problematic for themselves by introducing a new line of CPUs?

The only thing new next year is a Socket format change and Revision F CPUs.

(d) Killer is right...You don't have a f**king clue about anything.
0 0 [Posted by:  | Date: 12/13/05 11:59:29 AM]
- collapse thread


You're a liar, because I look for those motherboards all the time. How about this link - sspagename=STRK%3AMEWN%3AIT&rd=1 .

If they are used and properly listed, they still go for around $200. Look at this link - 14635.html . $522.

OK, so we've established you're a moron and a liar pretty easily. Next time do better.

Now to points. The Pentium III Tualatins were artificially given high prices and underclocked so they wouldn't interfere with Pentium 4 sales. The Pentium Ms have always shown much greater efficiency and for the majority of applications are clearly superior to the Pentium 4. So, for quite some time they have better technology available. Obviously, selling Pentium 4s didn't work too well despite the clock speed, since they keep losing market share because they have a non-competitive product. The Pentium M is a very competitive product, by contrast. Even Intel knows the Pentium 4 is a dog that has to be killed, it just took them too long to realize it.

Your remark about them getting what they had to out of R & D ignores a couple of important aspects. The specific product you do R & D for isn't always that important, since you can apply the research to other products as well. Do you think when they retire the Pentium 4 all the technology used in it will never be reused? Well, if you do you're wrong since Intel already said they'll be using technology from it in the Merom. Also, Intel can afford it, and it would represent no difference in terms of spending since they already had better products in the Pentium M line and the research was already done there. What difference would it make to their bottom line. Except the Pentium M is much smaller and thus easier to make. So, companies do like to recover R & D, as a broad statement, but in this specific instance your remark is irrelevant.

Your remarks about leakage miss the entire point, and are again just blabbing. They could have gone with the Pentium M. When you compare the Athlon 64 with the Pentium M, you consistently fail to acknowledge that a desktop Pentium M would be faster since the power ceiling would be higher. So, it would perform better than the mobile version, and if it didn't, it would use much less power than the Athlon 64. The Pentium M is a more efficient chip.

With regards to the Athlon 64 being vulnerable, you show your typical lack of intelligence. It will not be competing against the Yonah, it will be competing against the Merom. Already the Pentium M is a better CPU for most people, being much more efficient in terms of processing power per watt. It is also smaller, and therefore cheaper to make. On top of this, and even someone with your low IQ should have understood this from my first remark, Intel will always win with equal products. They have much better manufacturing, much better marketing, and much greater scale which helps with efficiency. On top of this, they are selling a smaller processor that performs more or less equally. AMD has to have a better processor, or they've got big problems. The Athlon 64 already isn't clearly better than the Pentium M, in fact, it is arguably worse considering everything (but could be better as well depending on the intended use). So, assuming one is a little better at this, and the other a little better at that, AMD has huge problems. Intel has every other advantage.

Also, this is comparing the Athlon 64 to the Yonah. The Merom is what it will be competing against in the relatively near future.

So, there is nothing to what you say. It is typical nonsense from someone that has no understanding of the market. But then, your idiotic comments about P3TDE6 motherboards revealed that. Why didn't you at least do some research before saying something so easily proven wrong? You must feel like a real moron, huh?
0 0 [Posted by:  | Date: 12/13/05 02:54:08 PM]
I take it you don't like the P7 arhcitecture much, however what else could Intel have done in the ~ 9 months wait before the 0.13 micron process was ready, when they were stuck at 1.0GHZ on the Pentium 3 Coppermine, if it had no been for the Pentium 4?, as the Athlon Thunderbird continued to scale in clock frequency during this period of time.
0 0 [Posted by:  | Date: 12/15/05 08:54:10 AM]
"(c) The AMD K9 will NOT be coming next year. Why? 2006 is when AMD switches formats for all their CPUs (Socket M2, F, etc). Its a transition period. Why would AMD make it more problematic for themselves by introducing a new line of CPUs?"

You are right.... But why wouldn’t AMD release a socket change with a new core, it has always been like this.

Also I see K9 documented on the AMD chipset partners.
AMD in 2003 said, K9 end of 2005, beginning of 2006.

If it isn’t released it's because they don’t need to do it (K8 is fine/Intel still bad).

Also AMD with socket 940(939) as already changed it 2 times, now as we all know will do the same one more time (because of DDR2).

Pacifica, new bus, new core, new memory? Maybe calling it K9 is exaggerated.
But didn’t Intel call the Pentium2; pentium3 when in fact was a Pentium2 with minor core changes (and SSE)?

DDR2 was part of the original K9 design, and not only that…
0 0 [Posted by:  | Date: 12/14/05 03:32:52 AM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture