Welcome to our website

To take full advantage of all features you need to login or register. Registration is completely free and takes only a few seconds.

Top Secret Intel Processor Plans Uncovered

Posted by: Jim on: 12/06/2005 12:42 AM [ Print | 11 comment(s) ]

Tom's Hardware has published an article discussing Intel's "top secret" plans for their processor lines.

The Yonah2 is going to be deployed into the server space as well, but carrying the name Sossaman. This one will come in a different package to be compatible with Intel's Bensley platform and it is expected to cause considerable buzz in the low-power server space such as the market for blades.

You can read their entire article over here.

Comment

cjcox Titus 3:5 Posts: 1396 Joined: 2002-09-19

#38464 Posted on: 12/06/2005 02:23 AM My take on Intel. We at Intel are desparately pulling out all of the stops in an attempt to play catchup with AMD. As a part of this effort we will inundate our major distributors (e.g. Dell, HP, IBM) with as many socket changes as possible so that they'll be forced to release 20 different product lines over the next 2 years. However, we will be too busy to particpate in any "contests" with AMD during this time. By introducing so many products in a short period of time, it is quite likely that our major competitor will become ourselves. Please be patient with us and remember to replace your superior AMD systems with our interim lackluster product lines while we push forward to competitive single and dual core designs. We would also ask you to pray that nothing that we do will force AMD to come out with their quad-core design early, since they are standing ready with it. Many of our partners are now offering AMD solutions. All we can say on that is that you need to remember what might happen... not saying it will happen.. but it might happen by choosing AMD in addition to Intel. Certainly we would never do anything wrong against someone who decided to sell AMD in addition to Intel. We love AMD and fully expect them to compete with us at every major vendor. It was strictly coincidental that hundreds of vendors were Intel only. We would never ever make any kind of stipulation on our supply of Intel chips. Thank you. HP-xw6600/2x2.83Ghz-E5440/4G/10K-160G-SATA/300G-SATA/7600GT/BluRayRW/openSUSE-11.1 HP-2530p/2.13GhzCore2/4G/1x160G/Intel4500MHD/DVD-RW/openSUSE-11.1 Custom/2xX5550/12G/2x300G-10K-Raptor/NVS295/DVDRW/ESX-4.0 Custom/2x6128HE/32G/4x2TB/onboard/Blueray/SLES11SP1-KVM

Comment

PascificWisp Registered User Posts: 0 Joined: 2005-09-08

#38468 Posted on: 12/06/2005 11:29 AM cjcox that was brilliant! Have you ever seen a company with so much marketing hype and distraction ability? Instead of getting on with good development, they just turn up the marketing hype- it is a wonder to behold!

Comment

rmn oh my, it's huge! Posts: 5894 Joined: 2002-01-26

#38469 Posted on: 12/06/2005 06:47 PM Originally posted by PascificWisp: Have you ever seen a company with so much marketing hype and distraction ability? One fruit: Apple. RMN ~~~

Comment

audioaficionado SMP aficionado too Posts: 3948 Joined: 2003-08-23

#38470 Posted on: 12/06/2005 07:10 PM If Dell ever cracks under the pressure and starts offering AMD systems (not just parts), Intel will be POed. Supermicro still hides their AMD offering on their A-plus page. Intel spends more on just PR/BS than AMD's entire corporate budget. Super Micro X8DTH-iF, Dual Westmere Xeons E5620 @2.4GHz Iwill DH800, Dual Prestionia Xeons 2.4 M0 @1.50v 220Mhz 3.3GHz Micronics W6-Lightning, Dual Pentium Pro @200MHz 512 KiB L2 Location: 42.3 N 122.9 W Medford, Oregon, USA Outside with my telescopes looking at the universe

Comment

Bubba Satori Registered User Posts: 5 Joined: 2005-10-19

#38471 Posted on: 12/06/2005 08:01 PM Andy Grove is spinning in his grave and he ain't even dead yet. If somebody told me 5 years ago that Intel chips would be so far behind AMD I'd have slapped them silly. And Intels response, VIIV WTF !

Comment

rmn oh my, it's huge! Posts: 5894 Joined: 2002-01-26

#38472 Posted on: 12/06/2005 09:31 PM Originally posted by Bubba Satori: If somebody told me 5 years ago that Intel chips would be so far behind AMD I'd have slapped them silly. I predicted that on this site about 3 years ago and several people tried to. RMN ~~~

Comment

i_wolf labhair dom as gaelige Posts: 2034 Joined: 2002-11-19

#38479 Posted on: 12/07/2005 12:42 PM The thing is.... define 'so far behind' AMD. Is it in terms of sales? No.. Intel is still chipzilla and again in 'Inq' speak, AMD is still chimpzilla. Is it in terms of manufacturing process ? No... Intel are at 65nm. AMD still hasn't transitioned to this process; there will be those that would argue... they don't need to. As far as AMD is concerned, of course they want to transition to a smaller die size, smaller manufacturing process and lower production costs nevermind the 'in theory' power and heat advantages. However as of yet they have not transitioned to 65nm. So is it in terms of performance? If so define performance... is it raw performance benchmarks on the desktop ? If yes, absolutely AMD's current fastest processor will run the majority of applications faster than the equivalent top of the line Pentium 4. Is it in terms of portable performance ? No, AMD is behind Intel on this one, in terms of performance per watt. Even looking at the future Yonah as was demoed on anandtech recently, yonah under load is still significantly cooler and consumes less power than a dual core Athlon 64 with lower power and thermal characteristics. So then, is it in the server space, yes AMD is faster than a Pentium 4/Xeon and scales much better too... it absolutely hammers Intel on this one in terms of raw performance and scalability, incredibly importantly in this area too it destroys Intel in terms of power and heat dissipation. However... does Yonah not change the tides in a somewhat limited way.... Intel has plans for it in a wide range, not just the portable market, including rackmounts and has shown that it produces around the same performance as an equivalently clocked A64 but with much lower heat and power characteristics.... there is no way to guesstimate what type of scalability we could expect but at least it does show a technological leap for Intel in this area over their current pitiful offering... pity yonah still doesn't have 64bit support.. thats another one to AMD. Given that Intel puts a minimum performance increase of 30% per clock on Merom over Yonah, Intel isn't looking to be in too bad a place this year when Merom arch replaces the Pentium 4/Xeon and Pentium M line. The bottom line as far as I'm concerned, both Intel and *shock horror* AMD have areas in different segments of the market where they are each the technological 'looser'. There is no penultimate winner. Hung like a donkey. Go like a horse!

Comment

cjcox Titus 3:5 Posts: 1396 Joined: 2002-09-19

#38481 Posted on: 12/07/2005 05:06 PM It is a sad truth that to gain a percentage point against Intel, AMD has to spend millions of dollars in R&D. All Intel has to do is run an advertisement. HP-xw6600/2x2.83Ghz-E5440/4G/10K-160G-SATA/300G-SATA/7600GT/BluRayRW/openSUSE-11.1 HP-2530p/2.13GhzCore2/4G/1x160G/Intel4500MHD/DVD-RW/openSUSE-11.1 Custom/2xX5550/12G/2x300G-10K-Raptor/NVS295/DVDRW/ESX-4.0 Custom/2x6128HE/32G/4x2TB/onboard/Blueray/SLES11SP1-KVM

Comment

rmn oh my, it's huge! Posts: 5894 Joined: 2002-01-26

#38484 Posted on: 12/07/2005 07:22 PM Originally posted by i_wolf: So is it in terms of performance? If so define performance... Getting the job done faster. That's what performance is (as opposed to, say, power efficiency). Unless time isn't a factor, the extra cost in energy is usually offset (by a very big margin) by the gains of finishing a job sooner. It's the same thing with chip cost: the chip with the best "bang for the buck" may sound more attractive, but when your income or professional success depends on getting things finished sooner, a 20% faster, 300% more expensive chip might actually be a better deal. Even a Celeron is overkill for most people, but we're talking about the "performance" market. does Yonah not change the tides in a somewhat limited way.... Intel has plans for it in a wide range, not just the portable market, including rackmounts and has shown that it produces around the same performance as an equivalently clocked A64 but with much lower heat and power characteristics.. It's also a 32-bit chip. No 64-bit addressing, no extra GP registers (compare software optimised for AMD64, that uses those registers, and I suspect you'll find a very noticeable difference). And it's significantly slower in floating-point. One has to wonder, if Sossaman is all that amazing, why is Intel wasting time and resources with things like Paxville and Yonah. By the way, do you know any place where I can buy Yonah, Dempsey, or even Paxville chips? What is the fastest Intel x86 chip that people can actually buy right now? Because, if you talk about supposed performance of future Intel chips, then you have to compare them with the supposed performance of future AMD chips, too. Granted, Intel wins by a landslide in codenames and paper launches, but, the way I see it, even the fastest Sossaman takes at least 6 months to open a text document. RMN ~~~

Comment

i_wolf labhair dom as gaelige Posts: 2034 Joined: 2002-11-19

#38485 Posted on: 12/07/2005 09:08 PM Getting the job done faster. That's what performance is (as opposed to, say, power efficiency). Not intentionally trying to get into a semantics debate, but many people will refer to computing performance in many different ways with each having a different meaning. It does not just mean 'getting the job done faster' or application performance. I think this was my point; performance is a very general term... and that if you define performance merely as 'it runs x faster than y' then I believe it is a blanket statement that only qualifys your usage patterns but does not reflect the general picture e.g. those of portable users, those who want dense rackmounts etc... i.e. those that don't just measure performance as x runs 3dSMax faster than y. And in this day and age where growth of laptop sales and portable sales is exceeding that of desktop http://www.usatoday.com/tech/news/2005-01-23-laptop_x.htm performance per watt is becoming increasingly important. Unless time isn't a factor, the extra cost in energy is usually offset (by a very big margin) by the gains of finishing a job sooner. I would agree with you, but there are users like me, (I would consider myself a power user.. maybe arrogantly but the things I run on my ThinkPad are non trivial and can often max out the CPU.). I don't care if AMD will finish a benchmark 3 seconds faster once I can get the job actually done since the better thermal and power per watt performance characteristics allowed me to actually finish before the battery ran out on a transatlantic flight. This has happened to me many times before and its a total pain in the ass. For me looking at the yonah benchmarks on a preproduction motherboard and a preproduction processor (and of course it was pre production but with availability in 4 -5 weeks time) compared with a desktop motherbaord and an equivalently clocked AMD, its performance for the most part was very much neck and neck. Its encoding performance was abysmal but its rendering performance in Max was fantastic. As you said though, there will be times when even the slightest raw application speed advantage is money. Both examples, yours and mine, are equally valid evaluations of performance, however both are based on different usage patterns and weightings as to what is the most important factor when measuring performance. When you consider such a broad user base in the computer world, and when you see the offerings made by Intel and AMD, I truly believe its completely unfair and inaccurate to say something blanket such as 'Intel is so far behind AMD'. Clearly they aren't in certain segments of the market, in fact in these segments AMD is quite a big behind Intel... and of course vice versa ad nauseum! the chip with the best "bang for the buck" may sound more attractive, but when your income or professional success depends on getting things finished sooner, a 20% faster, 300% more expensive chip might actually be a better deal. Let's all go and buy Itaniums from the dimension Itania It's also a 32-bit chip. No 64-bit addressing, no extra GP registers (compare software optimised for AMD64, that uses those registers, and I suspect you'll find a very noticeable difference). Exactly, that was why I mentioned pity yonah still doesn't have 64bit support.. thats another one to AMD . It is a great pity. However how many laptops do you know of from Dell, Asus, Acer etc... that will allow you to pack more than 4 gigs of memory In a server environment I would agree though it would be a detriment. But I get the feeling that Yonah is only meant to be a stop gap for the server market until they manage to complete their true next gen Merom architecture. Also in a day and age where the majority of applications on the desktop and mobile markets are still 32bit in nature it won't be that big of a problem for the majority of users. The server market and those with a necessity for 64bit processing will tell a different story.. I agree entirely, I'm not trying to take a pro-Intel side on this. They are behind in this segment whether they like it or not... particularly with the lack of iAMD64 extenstions... a finalized bios and motherboard could boost the scores we saw in the preview at anand for most other stuff. By the way, do you know any place where I can buy Yonah, Dempsey, or even Paxville chips? What is the fastest Intel x86 chip that people can actually buy right now? Because, if you talk about supposed performance of future Intel chips, then you have to compare them with the supposed performance of future AMD chips, too. Well in 4 -5 weeks time we are to expect the first laptops from Dell et al... so I guess it isn't that far out Granted, Intel wins by a landslide in codenames and paper launches, but, the way I see it, even the fastest Sossaman takes at least 6 months to open a text document. It does, but with the money you would likely save over one of these or its equivalent at time of Sossamon/Woodcrest you could take a nice holiday here! Hung like a donkey. Go like a horse!

Comment

rmn oh my, it's huge! Posts: 5894 Joined: 2002-01-26

#38489 Posted on: 12/08/2005 02:20 AM Originally posted by i_wolf: performance is a very general term... and that if you define performance merely as 'it runs x faster than y' then I believe it is a blanket statement that only qualifys your usage patterns but does not reflect the general picture e.g. those of portable users, those who want dense rackmounts etc... i.e. those that don't just measure performance as x runs 3dSMax faster than y. But that is what performance means, when people talk about CPUs. Look at SPEC (the Standard Performance Evaluation Corporation). Their CPU tests don't test power consumption or battery life or IPC or cost, they test one thing: how quickly a CPU can complete a certain task. Most people buying laptops will be more interested in the overall energy efficiency of the system. No one uses laptops for their processing performance. And in this day and age where growth of laptop sales and portable sales is exceeding that of desktop http://www.usatoday.com/tech/news/2005-01-23-laptop_x.htm performance per watt is becoming increasingly important. Even a 1.6 GHz Pentium-M is overkill for the vast majority of laptop users. "Performance per watt", in that segment, is just a way to make "low power consumption" or "battery life" sound more appealing; it's all about bringing the "watt" down, not getting the performance up. Real "performance per watt" is only really an issue in the high density server / blade segment. Which makes Sossaman the only really interesting (but still virtual) Intel CPU. I truly believe its completely unfair and inaccurate to say something blanket such as 'Intel is so far behind AMD'. In terms of real, high-end CPU performance, it is. Let's all go and buy Itaniums from the dimension Itania The problem with the Itanic isn't the price (it's not much more expensive than the high-end DC Opterons); it's the performance (at least in DCC, which is the area that interest me - I'm sure it's great at whatever alien scientists use it for). However how many laptops do you know of from Dell, Asus, Acer etc... that will allow you to pack more than 4 gigs of memory How could they, if their CPU doesn't support it? Anyway, there is more to 64-bit addressing than supporting over 4 GB of RAM and there's more to AMD64 than 64-bit addressing. In a server environment I would agree though it would be a detriment. ...which is the only segment where "performance per watt" actually means "good performance" and not just "so-so performance with very low power consumption". But I get the feeling that Yonah is only meant to be a stop gap for the server market until they manage to complete their true next gen Merom architecture. Given what we've been hearing about Dempsey and Sossaman, the only possible conclusions are that: a) They're not as good as Intel is claiming or b) The time and money they invested in Paxville and Yonah were yet another example of misguided management (should have been used to speed up Dempsey and Sossaman). Also in a day and age where the majority of applications on the desktop and mobile markets are still 32bit in nature What does "32-bit in nature" mean? Taking advantage of the extra registers in AMD64 CPUs, for example, only requires a recompile. If I was a very cynical person I would say that maybe Intel continues to release 32-bit CPUs because they realise that, the sooner all chips are 64-bit, the sooner all software will be recompiled to 64-bit, showing the problems with Intel's flavour of AMD64, and eliminating one of the supposed advantages of the Itanic. In other words: they're using their market position to slow down the migration, buy some time, and placate their IPF division. Well in 4 -5 weeks time we are to expect the first laptops from Dell et al... so I guess it isn't that far out You mean the 32-bit ones, right? with the money you would likely save over one of these or its equivalent at time of Sossamon/Woodcrest you could take a nice holiday here! Another good reason to go with the Opteron (which you can get for about $980, BTW). RMN ~~~