Slashdot Mirror
Intel Drops Tejas, Xeon To Focus On Dual-Core Chips
PunkerTFC writes "Reuters has an article about Intel dropping the fourth-generation P4 chip (codenamed "Tejas") and the Xeon server processor. Intel says they want to concentrate on their new 'dual-core' technology for desktop and notebook systems. This is essentially putting two processors on one chip, allowing for a doubling of performance with less energy use. The introduction of this technology was not expected for another year and a half. Rival chip maker AMD says they have the capability to produce dual-core chips and will introduce the technology when they "feel there is a market need.""
but might this have something to do with the recently-announced Longhorn specs?
FIRST POST
doi you think this has anything to do with the fact that MS is shipping a database(traditionally considered able to leverage hyperthreading very well) on their desktop? HMMMM!
But multi-cpu system sales figures do not justify abandoning the single-cpu market in any way. This is a serious mistake or an admission that they just cant keep up with AMD anymore.
But, does this suffer the same problems as current chips do wrt dual processors? Or quad processors?
What's the penalties of this technology? Does anyone know?
Sounds too good to be true for a dual core cpu to act as a single core proc.
This has been discussed before.
Karma: Meh (Mostly from meh.)
FP FP Written Written from from my my new new dual dual core core chip chip from from Intel Intel. Still Still some some bugs bugs to to work work out out.
sold more chips than Intel during a two week period (52% to 47%). I wonder if Intel is finally feeling the heat from AMD? Maybe Dell (who only sells Intel) is pushing on them too.
There is a rage in me to defy the order of the stars, despite their pretty patterns.
Well weekends are for dupes it seems
I mean this was interesting a couple days ago, but now it is old news...
D.O.U.O.S.V.A.V.V.M.
AMD seems very calm about this. If I was in AMD's position, I would be in pretty scared. I mean, Intel is a year a head of schedule.
Personally, I'm just happy that soon enough I'll be able to buy a duel core chip.
Help Fight SPAM today!
[quote]when they "feel there is a market need."[/quote]
Um, the market would be me. The time would be now.
Bring it on!
I see that the new dual core opterons are supposed to be pin compatible with existing boards. So that makes it possible to get an AMD server today, and in xx months time pop in a new chip and turn it from a single proc to a dual proc (dual -> quad?) server. Nice. Now if only memory prices would come down some more. So I can enjoy a 16GB quad proc server for under $3K.
ADV: VPS Hosting on the fastest chips we can findSo, who will merge with who...
... and AMD & ATI merge?
... and AMD & Nvidia merge?
Multiple Cores, Multiple Pixel Pipelines, ->
Intel & Nvidia merge
or
Intel & ATI merge
As being the recommended chip for running Microsoft Longhorn Version of Windows. Wonder if this has anything to do with Intel's decision.
Intel says they want to concentrate on their new 'dual-core' technology for desktop and notebook systems. This is essentially putting two processors on one chip, allowing for a doubling of performance with less energy use.
Is this a parallel implementation then? In that case performance is only doubled for processes that can be performed in parallel.
I think this is more related to moving to the PM from the P4 architecture as the M series is more scaleable - taing P4 any further requires a lot more power and generates a lot more heat.
This seems to be the new trend,
AMD will have dual core opterons next year:
I've got to admit I didn't read the article. However what makes AMD think there's a market need for 64 bit processors (especially without the mass market operating system) but no current market to warrent an AMD dual-core processor? Does improved performance and lower power consumption not mean anything to AMD (to quote some buzz-words)? Not to troll, but I think someone's falling further behind.
Awaits the likley response that "they will only use 2 slower chips". Not for another ten years ..etc
How many people do you know with dual procs. anywho? the only one I know is a mac friend. What kind of heat sink are we going to need for dualies? Its gotta weigh in round 5lbs. And have the noise output of a harley
Sig (appended to the end of comments you post, 120 chars)
... trying to get themselves ready for Longhorn too?
Free Firefox news reader.
Here's the real impact many of us will be feeling. Software vendors that license by the CPU have already in fair part indicated that they consider "dual core" chips to be two CPU's for licensing purposes.
In other words, people are going to find themselves having to pay higher licensing fees with regular desktop computers as well as servers. Small workgroup servers could be really hard hit by this from some vendors.
I wonder how this will play out with XP Home which only supports one CPU? AMD has the technology so they may well respond in kind when Intel does (dammit lead AMD, lead), which could have a fair impact in weaning the masses of XP Home. I dont think MS will let this go the route of hyperthreading with the "logical processor" support.
"clock speed" -- a term familiar to PC shoppers as gigahertz or megahertz
Aside from the fact that clock speed is measured in gigahertz or megahertz.
ah, mod points
Intel says they want to concentrate on their new 'dual-core' technology for desktop and notebook systems. This is essentially putting two processors on one chip, allowing for a doubling of performance with less energy use.
Is this a parallel implementation then? In that case performance is only doubled for processes that can be performed in parallel.
I think this is more related to moving to the PM from the P4 architecture as the M series is more scaleable - taing P4 any further requires a lot more power and generates a lot more heat..
"Sir, maybe we should introduce our dual core chip now!"
"No... that's just what they'll be expecting us to do..."
Business isn't willing to pay for products, innovation and careers, so we get brands, mortgage commercials and layoffs.
...will sell for the low-end and be called Halferon.
Hyperthreading and dual core are very different things. A dual core processor is basically two processors put onto one die. There are twice the number of execution engines, just like two separate cores, but on the same chip. This means it's easier and cheaper to make and install than two separate processors, and it has approximately equal performance.
Hyperthreading takes one physical processor and makes it appear to be two logical processors. There's still only one core and one execution engine. It appears to be two processors, but a 3.2GHz Pentium with HT will have nowhere near the performance of 2 3.2GHz Pentiums without HT.
Transistors are getting smaller and the chipmakers can fit more and more onto a chip.. However, it is much cheaper (less design time) to simply run two cores with some "glue" hardware than to design a new core that is 8-way superscalar instead of 4 (for example).
One way to look at dual-core is to view it as a dual-processor (MP) system with a very low communications cost, since both cores are on the same die. The disadvantage is similar; since the two units are not perfectly synchronized, such a system runs best with multithreaded code. A single-core CPU with the same number of transistors will run faster, while the dual-core is not quite "double the speed" of one of its cores.
Intel, like Microsoft, Dell and Sony, is a favored company.
AMD, like Nokia, Apple and Nintendo, is not.
AMD's strategy (Opteron instead of dual-core?) will therefore be called "a significant risk given the current market reality" while Intel's strategy (dual-core instead of Itanium?) will be called "a savvy decision for the technology giant," even though the media wouldn't know an Opteron or a dual-core CPU if one jumped up on their desk and did the tap number from 42nd street.
All of the general stories will make repeated and redundant references to the effect of Intel's strategy on the "tech-heavy Nasdaq."
This is no different than the Sony vs. Nintendo console competition. The media doesn't like competition. Neither do the markets. (There is only room for three companies in any given market) It's so much easier to be a sycophant when your favored company has 80% of the market.
Business isn't willing to pay for products, innovation and careers, so we get brands, mortgage commercials and layoffs.
I remember reading a while back that one of the Intel chips came out of the design group in India. Was that the Pentium-M? Or am I just remembering this totally wrong?
It would be kind of funny if Intel cancelled its American chip designs in favor of continuing work on a design from India.
As others here have mentioned, this is going in the direction of Microsofts new Longhorn OS, which has some hefty system requirements. The dual core cpu is an interesting concept, an extension of Moores law, so that the computing power increase of 2x @18 months remains constant. Who cares how this is achieved, as long as its happening! What I find really interesting is the concept of usability...while slashdotters will easily find uses for these pc's, what will jane the nurse be doing? compositing home videos for her MSN messenger profile? A lot of the more powerful apps such as 3d modelling require learned skills and also talent to make something good.
Still, cpu's of this magnitude everywhere will help with CETI@home number crunching. Maybe the Terabyte storage mentioned in those Longhorn specs has something to do withCerns Large Hadron Collider going online. It is expected to produce 10 Petabytes of data a year. The data is going to be saved via the grid
The chinese have a saying, may you live in interesting times...
Do you need a website upgrade?
didn't intel say the same thing about wating for the market demand when it came to 64bit computing? That was a big mistake on intel's part and it cost them big, so why would amd go and do the same thing?
AMD feels the need to produce 64-bit "desktop" CPUs, while Intel doesn't claiming they don't see the market for it. Intel feels the need to produce Dual-Core chips, while AMD sits back and says they won't until they see the market for it. Sounds like the processor market is now being forked. The difference between an AMD and an Intel CPU is finally going to start emerging. Before recently, they were both x86 and did everything practically the same. The difference was how they were numbered. Now, they are going in two totally different ways -> one dual-core and one 64bit. When will we see 64bit, dual core cpus from each?
how's is it redundant? I was the 3rd poster here :( .... READ MY SIG..DAMN IT :(
Won't somebody please think of the Karma!
Intel next year will sell chips for both desktop and notebook computers that combine two microprocessors onto a single piece of silicon, "like putting two cylinders in a car instead of having one big cylinder," Nathan Brookwood, an analyst with Insight 64, said.
Or maybe Longhorn is so bloated, it needs it's own CPU just to sustain the operating system, and another processor to run programs.
Ruby on Rails Screencast
Rival chip maker AMD says they have the capability to produce dual-core chips and will introduce the technology when they "feel there is a market need".
When are they going to feel the need from market? If it's a better chip that the ones currently being used, then there's already a market need. Can't those morons see it? How else are they going to determine the market need? If I need a dual-core chip that increases the performance while using lower power, do I have to run on the streets yelling AMDs name? If AMDs planning to wait, they are going to lose a major share of the marked to Intel.
Part of the long horn specs, remember?
Makes me wonder if you would need a current "Server" Microsoft OS to use the full potential of the chip. They currently only support 2 "CPUs". Meaning 1 HT or 2 standard CPUs. 1 dual-core/HT CPU would be "4" theoretical CPUs, wouldn't it?
This sig left unintentionally blank.
Wait... isn't "multiple core" technology that thing that the people who make PPC chips were making a huge deal about how they'd be introducing it soon when the G4 was first introduced, then quietly dropped?
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
The current IBM POWER4 and upcoming POWER5 chips are both dual-core chips. Here is a nice presentation(PDF format) about the POWER5; you can see in the die photos where there are two cores. There have also been rumors of a dual-core PowerPC based on it, but nothing concrete yet.
Broadcom (which bought SiByte) markets a dual-core, 1GHz 64-bit MIPS chip called the BCM1250 which has a lot of integrated networking goodies.
Finally, it bears pointing out that on the other side of Intel's severed corpus callosum, they're also working on a dual-core chip.
While I think the Pentium-M is an excellent CPU, let's not forget that it is essentially just a Pentium III with a quad bus, lots of cache and SSE2 instructions.
Furthermore, the P3 is just a Pentium Pro with MMX, SSE and on-die L2 cache.
A little retro? Seems strange that the future is in a P6 architecture. Maybe when these get too hot we'll move to a massive array of 486s.
Jeremy
Intel is dropping development of the Xeon Jayhawk.
So the 05 Intel product line is:
HP and Unisys Datacenter class servers: Itanium
Desktop and Notebooks: Unnamed dual processor (assumed to be 64 bit, article doesn't say explicitly)
Low range workhorse server: current 32 bit Xeon line, no enhancements coming? Does another shoe drop here?
Intel says Adios to Tejas and Jayhawk chips
Wow. That really is one of the most interesting things i've heard in a while. Why not have the OS be a separate entity with it's own proc and one proc (and permissions set) devoted to user apps....
Wow, i've been in the sauce today, but this idea is worth more thought.
It's not clear to me that a dual core processor would take less power than a single core processor. Sure a dual-core processor _will_ take less power than two single-core processors on a board. So I suppose at a system-level a single core processor will take less power than a dual processor system, but the power problems we're seeing now are primarily at the chip-level.
BTW: As someone who 'knows' people that work at Intel, this decision was a pretty huge one on the 'Richter scale'. 1000s of people found out in the last couple of weeks that they were being redeployed to different projects (or making major changs on current projects). This decision is having a huge effect inside of Intel. I suspect that this kind of shake up means that the higher ups at Intel were very afraid that AMD is making major inroads and they finally realized that they couldn't keep going in the direction they were headed in without disasterous effects on marketshare.
when they "feel there is a market need."
Isn't that Intel's party line for waiting so long on 64-bits? The more AMD and Intel claim to be different, the more they act the same.
Um, the market would be me. The time would be now.
Bring it on!
And how many million did you want a month? Ah, I see, talk big, but buy only one. And that's the reason you won't see it for a while. These guys make millions of CPUs a week. So before you put on the tinfoil hat and decry how they just don't understand the market, realize its the actions of many, not the few that drive the economics behind the progess of technology.
As far as the dual core opterons being pin compatible, that's not necessarily a good thing. For one it means that it will continue to be stuck in old technology. DDR-2 is out now, but this means that Opteron is doomed to the outdated DDR-1 for another 2 years.
to the Engineer, the glass is neither half full nor half empty. Its just two times too big.
What's with AMD?
"When we feel there's a market need?"
These morons introduced 64-bit chips 5 years before anyone cared, and crippled the technology by making it straight IA32 with more bits.
The market needs dual-core CPUs to advance.
There's no way to get CPUs faster any more without reaching current levels that no power supply can reasonably handle in that space (hint: 100 Watts at 1.2 Volts is 83 Amps).
The only solution is to divide the computation among several processors and parallelize.
AMD's response is clanking hollow. They're already playing catch-up as you read this.
I mean, Intel is a year a head of schedule.
The author incorrectly states that Intel's dual core CPU is "more than a year ahead of schedule". Six months ago during the Intel fall analyst meeting Intel claimed (slide #40) dual core for the home computers would arrive in 2005.
This is a rather interesting bit of information from the article: "This strategy was not expected for at least a year-and-a-half, said Dean McCarron, the head of Mercury Research."
Well, how is this news? Intel is claiming that they will go dual core by the end of 2005. A year and a half from now is...2005, just like the Intel presentation from six months ago said.
Most consumer software sold today is multi-threaded and would definitaly benefit from a parallel operating multicore system. Most modern OS's also support SMP already, and switching to a multicore is no big deal in that area or difficult to take immediate advantage of.
And you're right. The cores will be derived from the pentium M - not the 4.
-
I'm picking up on a divergence trend here.
AMD is went with x86-64, and Intel said "we'll wait until there is a need."
Intel is now going dual-core, and AMD says "we'll wait until there is a need."
I think AMD has the upperhand, though. Intel has the 64 bit technology, but doesn't want to release it to the consumer market yet, more than likely because the 64 bit version of Windows sucks hardcore. AMD could double the core on an x86-64 proc and beat Intel yet again.
Colin Dean Go a year without DRM
DDR2 isn't much better than DDR1, but it's much more expensive. If AMD is smart they will jump from DDR1 straight to FB-DIMMs.
teja vu.
Will we have to use kernels built for Symetic Multi-Processors? Some of the Linux hardware modules are not written for SMP!
Religion is the main cause of atheism.
It was just posted with the dual core chip, so it got posted twice.
Thanks to AMD and their recent successes in the market, Intel it seems is finally focussing on their core business - manufacturing successively faster processors, not inventing new marketing schemes. Before this announcement I could only imagine chips like these being reserved for high-end xeons.
Competition is always a good thing.
They'll still go for the intel inside flashy ads until such point AMD catch up with some nonsense of their own..
s/joe six pack's/mondeo man's semi/ if you are in the uk
... only the one glaring typo, too (cpu/gpu). Certainly as good a (general public) market forecast as you can see in any of the name brand rags on the web, and certainly more honest. You analyse tech market like I do geopolitics, very similar, probably why I liked it. This leads to this leads to this, etc,the connections and odds and probabilities,and so on, what a futurist does.
Have you done this in the past, say 5-6 years ago, and, if so, what did you think then, and how did it turn out?
Wow... Duplicate comments (posted 5 minutes apart) for the second duplicate of an article, and getting moded up for both, congrats:
9 5866 9 5886
http://slashdot.org/comments.pl?sid=106885&cid=90
http://slashdot.org/comments.pl?sid=106885&cid=90
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
No. If and when a dual processor chip comes out, it will not be a stretch to have Linux versions tailored to the new chips. Novel (Suse) already has a version for AMD's 64 bit processor and they would have no trouble creating a dual processor version as Unix/Linux has been multi-processor friendly for decades.
Embrace the future.
I think the reason 64 bit hasn't hit hard yet is because theres not much of a market for it in PCs. Think about what 64 bit gives you. More than 4 gigs of RAM, and perhaps a few ancillary things.
The consumer isn't going to have more than 4 gigs of ram in a machine for several years. Server and scientific apps can benefit from it now, but they would also benefit alot from moving away from x86 (almost 30 years old!) and with less of the backwards compatability problems.
So far the only non-server and non-scientific use of amd's 64 bit chips was in an OpenLabs OpenSynth instrument that loaded everything into its 8 gigs of ram for maximum response time.
-
I've seen Windows XP recognize four processors (dual processors, both hyperthreaded, so 4 virtual processors). So it shouldn't have a problem with this.
Of course, if you're using MS, you *need* to upgrade to Longhorn. Why? We'll tell you later.
The trouble with dedicated processors (to the OS or amything else) is that their spare MIPS go to waste. And two separate CPUS have more problems communicating than one. If the other processor is different (e.g. the processor on a graphics board), then it may be worthwhile. But if the processor is just another of the same, better to share tasks between all the available processors.
Consciousness is an illusion caused by an excess of self consciousness.
contrary to slashdot ignorance and FUD, the OS doesn't spend most of its time running the CPU.
Most of what the OS does is IO, which idles the chip while waiting for the IO to complete. Tthis is why all operating systems switch to the next task while waiting on IO. If your CPU is running at less than 100% usage its because every program is waiting for IO for most of the time.
-
Just remembered it was WinXP Pro, so not sure if the comment is still valid. :(
No, 1 physical CPU is 1 CPU, HT or not. I installed Windows 2003 Server OEM 1-2CPU on Dual Xeon HT system with no problem.
It seems like they could "disable" one of the cores and call it a Celeron. However, I don't know if consumers will accept processors with "half" the performance of their mainstream counterparts. On the other hand, a single-core Pentium-M (an impressive, but expensive, performer today) would seem like a good deal for a "budget" processor one or two years from now.
I don't think Intel will just "disable" half the cache like they've been doing since the Celeron 300A (and keep both cores). I think this is unlikely because a dual-core Pentium-M with 1MB of L2 cache (remember, Dothan will have 2MB L2 cache) would be too darn good for a "budget" processor and would cannibalize sales of their mainstream and high-end processors.
However Intel decides to make the new Celerons, it looks like we will have much better low-cost options from Intel than the pathetic Pentium-4 based Celerons with 128MB L2 cache and 400MHz front side bus.
TO START
PRESS ANY KEY
Where's the 'ANY' key? I see Esk, Kitarl, and Pig-Up...
As I understand it, the cpu core is quite a small part of the silicon area of a modern processor chip. Which is, of course, the logic behind this development. But then, why stop at two? Well, maybe Intel have simulated and found that, in the current state of the art, two is optimal. But if cache gets larger and busses get faster, two may cease to be optimal. Which will lead to competition by number of cpus, not GHz.
Consciousness is an illusion caused by an excess of self consciousness.
Yes, but it would probably mean nothing to pricing. You can bet that software will be licensed to run on hardware being targeted at desktop solutions. Dual CPU systems are already supported by XP/2K without requiring a server version.
So this is Intel's solution to 64-bit computing: two 32-bit cores on the same chip.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
1: Megahertz matters!
2: Profit
3: Performance ratings are for wimps (see #1)
3.5: Demean companies using performance ratings instead of real Megahertz.
4: Profit
5: Megahertz creates too much heat (we goofed)
6: Long pipelines good
7: Long pipelines bad
8: Adopt obscure semi-performance based chip rating system (i.e. how can we fool them today)
9: Profit?
10: Insiders start selling short (projected behavior)
11: PROFIT!!
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
Hyperthreading (or simultaneous multithreading - SMT) creates the ability to run 2 virtual threads on a single chip. This can be as simple as running an 'integer' and a 'floating point' thread, as the SIMD/FPU are really still separate units anyways. In a more complicated form, it means multiple threads are split across different pipes/units/stages - making more use of the processor if you like.
Dual Cores are 2 actualy processor 'cores' on the one chip. That's 2 LSU's, 2 FPU', 2 IFU's, etc plus as much L2 cache you can shove on to feed 'em.
It is like having 2 real proc's....it's just more efficient to make dual core chips than 2 separate ones. Ofcourse you still need an OS that is good at handling multiple processors - which windows isn't really (compared to OSX for example). Speaking of Apple - IBM have been making dual core chips for some time now. The POWER4 was dual core, and the POWER5 is dual core and multithreading! Lots of Apple rumor sites are saying that the next Mac chip (G6?) will be IBM's 980 - supposedly based on the POWER5, ie dual cores etc etc.
Here's a link to an IBM presentation with a bit of info on SMT & dual cores (pdf, sorry).
... you'll do well when you do your stream of consciousness writing, it's easy to follow and to the point. Keep it up. And yes, you need all the data bits for forecasting, and the ability to separate data from opinion to do it accurately. My advice is specialise in a market segment without losing trace of over-all macro trends, this would be economic trends globally, currency trends, energy trends, following the politics of the major players around the world, etc. then apply that broad set of data-bits to the specialised forecast you need.
Hmm, example. Say I many years ago was sounding the alarm on the *way* we were trading with china. it was because of WHAT they bought from us, compared to what we bought from them. Looking at that, you could see it was a worse deal than the raw numbers, even taking into consideration the now over hundred billion a year deficit we are running with them, which is a form of direct foreign aid, because they use that deficit to purchase our future debt.
anyway, I digress. China has been buying factories, machine tools, and snagging data, R&D, getting information on the incredible cheap, allowing them to use that trade imbalance as a form of force mulitplier for their economy. They bought tools that make tools that make the ultimate stuff. it has given them a 4 to 1 advantage in raw costs from the hardware side, couipled with a 20 to 1 (initially) labor advantage. that's why they are kicking butt, and will be thw worlds dominant economy within ten years or so. I wrote on that about 3 years after nixon and kissinger decided to open up china (bad idea at the time, IMO, we didn't insist on quid pro quos, it was lamer).
Anyway, that's what has happened, just by looking past raw numbers to the actual "stuff" the numbers represented, you could get a much better forecast than the TV and rag pundits, well, they are mostly shills too but that's beside the point. there wasmore to it, especially some high level blackmailing going on during the past two administrations, but that was the important reason, and it was done on-purpose for the purpose of some extreme high level skimming, which we can see happened and now is the accepted norm. When we adopted a one china policy, it was in effect a defacto national anti protectionist policy which if you follow a lexicological extrapolation means it was a "pro" the other guys, in that case china, now it's extended all over, with very little thought given to what was really needed, a full replacement economy, which was never possible in the first place, given a default a nation our size can't be one industry, we needed to always be vertically integrated and widely diversified.
They screwed the pooch on that one, sad to say...
Spooky stuff really, because now, if you look at global oil supply mixed into this, strip the numbers towards the middle (throw out the phony highs and lows) to avoid the stock manipulation pseudo prices, we are in *some hurt* coming up, almost exactly the time china's (and the rest of the now industrialising second world's) demands for oil will quadruple by the calendar and some rational projections. That's by the end of this decade, or close enough to not matter for this purpose. Gonna get fugly then.
And the only way we in the US can stay competitive from now to then is by vastly diluting our money supply,by increasing the supply of less and less valuable digits into circulation, which is a no-win eventuality. It's a lock there.
How this will aplly to the niche of computer tech will mean, really cheaper hardware, but offset by devaluation of what our money represents, so that offsets and balances (+ ~ - ) , so we'll see a leveling soon, within a couple of years I think.
OS and softwares in general will drop drastically in real and perceived worth, to follow your example of throw-away on the hardware side, outside of very specialised niche markets. I agree on that. With 10 million (or more) programmers hitting the market within a couple of years all over, softwares
According to Reuters and the Wall Street Journal, Intel is supposed to officially announce today that they're not going to bother with the Tejas generation of PIVs/Xeons.
This ought not come as too much of a surprise to those of you who read this last March, and we openly wondered whether Tejas was going to see the light of day a little while back .
Yes, this a major announcement that will effectively knock Intel out of the box in the cutting-edge overclocking world for at least something close to eighteen months. This essentially leaves us with whatever AMD chooses to offer.
Nonetheless, the biggest aspect to this story is not the "what," but the "why."
A few days ago, the chief technology officer at IBM, Bernie Meyerson, told an industry forum that the traditional and expected increase in speed just from shrinking the manufacturing process is dead .
To quote:
"Somewhere between 130-nm and 90-nm the whole system fell apart. Things stopped working and nobody seemed to notice. . . . Scaling is already dead but nobody noticed it had stopped breathing and its lips had turned blue."
(This comes from the company that AMD paid $46 million dollars to help build 90nm chips, BTW. It also comes from the company that was supposed to have 3GHz 90nm PowerPC chips ready for Apple in a couple months, but is now talking about eventually getting to 2.5GHz.)
Meyerson said the biggest reason for the problem is power leakage, the same as what Intel has been saying. He also pointed out that the problem with power leakage is "nonlinear."
That's a fancy term for saying "it doesn't get slowly worse; you get past a certain point, and everything suddenly falls apart on you."
It's Not Quite Over
Mr. Meyerson is not saying "it's all over." What he is saying is that the era of easy, big gains from each new generation of processors is over. As he put it, "60 to 70 percent of the benefit of each new generation of manufacturing would have to come from innovation."
By that he means technologies like SOI and strained silicon, though he implied that these were not long-term fixes to the problem.
What is clear is that future technological advances are going to be a lot harder to do, cost a good deal more, and being a lot harder to work with than has been the case in the past. The old way of doing things is broken, and there's no mature alternative around at the moment.
Perhaps one will eventually show up, but the magic bag is empty at the moment, and it will probably take years to come up with some major new tricks.
In the meantime, progress will slow down.
Playing Noah's Ark
In all likelihood, Intel's short-term answer to this problem is to stop revving and start adding. Processors, that is. The son of Pentium-M which will become Intel's next generation will almost certainly be a two-headed beast. In short, a 6GHz processor won't be a 6GHz processor; it will be two 3s.
AMD plans to do exactly the same (which ought to tell you that SOI, good as it is, is no long-term fix to this problem).
This is hardly something either party would willingly want to do rather than increase speed, simply because the vast majority of current programming does not (or even cannot) work better with two-headed action.
It's certainly not something Microsoft want to deal with on the OS side, and probably is a big reason why Longhorn keeps getting pushed back, much less the armies of non-MS programmers out there.
It's going to happen because the hardware people don't have a choice in the matter.
I follow the SDK and GDN principles.. Spelling Dont Kount, Grammer Dont Neither
What happened along the way is that the parallel approach never quite went anywhere (Illiac IV? Transputer?) while the chip makers followed Seymour Cray's approach to parallelism (pipeline, superscaler -- parallelism on a very low, hardware level -- sure the software could be tweaked to take advantage of it, but it didn't require a massive rewrite of the software.
Even hyperthreading was virtual multiprocessing instead of actually having multiple processors (I tried making jokes about Intel HT being a reinvention of the CDC 6600 to little effect), and I suppose software multi-tasking is using a single processor to emulate multiple processors. Oh, in terms of reinventing the super computer, whatever happened to "vector instructions" (I believe it is called SIMD). There was a feeble attempt with MMX and later SSE2, but there did not seem to be that much interest on the part of compiler writers to do anything with that.
So are people saying that we have reached the end of the road for the pipeline, superscaler, SIMD type approach and that now is finally the time things are going to shift to parallel machines (MIMD)? I heard the "supercomputer" folks have long made this switch.
I guess we already have the software tools to take advantage of multiple processors in the form of threads, although they are a PITA to get right given the current languages and software design patterns, although there may yet be a big payoff from object-oriented programming by running large numbers of objects as separate threads on large numbers of processors in the not-to-distant future. What are the supercomputer software people doing to cope?
You will need an operating system that supports more or less symmetrical multiprocessing for really take advantage of the new Intel CPU's.
That means you need at least Windows 2000 Professional or Windows XP Professional on the Windows side or the more recent Linux commercial distributions to fully take advantage of the chip.
Remember, AMD and NVidia are already in bed with each other. nForce is an AMD chipset.
The first all-indian Intel chip is a Xeon designed in Whitefield, Bangalore.
"When the only tool you own is a hammer, every problem begins to resemble a nail." - Abraham Maslow (1908-1970)
If a dual-core CPU is released from Intel (and AMD?) are released, and it is reported as two processors to the operating system, I imagine we could expect a heavy shift in programming tactics by major and minor software companies to be parallelized to take advantage of both CPU cores. I know some problems cant be parallelized, but many can. This will be a HUGE advantage of those of us running dual processor machines, as we will see huge gains from these new techniques.
It's easier to fight for one's principles than to live up to them.
Rival chip maker AMD says they have the capability to produce dual-core chips and will introduce the technology when they "feel there is a market need."
That's what Intel said about introducting 64-bit processors to the masses(not just for servers like the Itanium is for) and now look they are lagging behind AMD, hopefully AMD does not screw up their good position by letting Intel get the head start.
AMD says they have the capability to produce dual-core chips and will introduce the technology when they "feel there is a market need." Didn't 3DFX say the same thing about 32 bit graphics cards?
The point is not to make new, more expensive chips. Instead, the point is to switch from advancing by making the chips smaller to making the chips internally parallel. Presumably they will follow the same price guidelines as existing chips: the best chip will cost $650. Eventually, that same chip will cost $50.
:)
They'll justify this to consumers the same way that they always do...they'll tell them that the new chip is better and charge more for it. Or they'll switch lines like they did when the P4 came out. Under $200 P4s came out before they stopped shipping P3s. Maybe they'll do something sneaky like double the reported clock speed, since there are two processor cores...clock speed's additive, right?
http://www.intel.com/research/silicon/mooreslaw.ht m
How are they going to continue to shove more and more transistors onto the same die, without shrinking it?
Are we (finally/sadly) seeing the end of Moore's Law?
I think this also means that single-core x86 processors have come almost as far as they can..
Privacy begins with
Intel is clearly now pursuing a mult-core roadmap de-emphasizing clock speed. Note that Intel is pursuing that same strategy that Sun is following. I bet that IBM is also exploring the same strategy although IBM has made no formal announcements.
We should expect that Intel has been studying the Hydra chip developed by Professor Olokuton at Stanford University. That chip is the basis of the future generation of chips at Sun.
Given Intel's reputation with adhering to deadlines and working employees to death to meet those deadlines, who -- Sun or Intel -- is most likely to develop the highest performance multi-core chips? Yep. Intel.
Yep. Sun is dead meat. Sun's managements is betting that Rock and Niagara, the key multi-core chips, will save the company. In reality, the new Intel chips will make mince-meat of both Rock and Niagara, and we will see a solar eclipse of Sun -- the epitome of the dead man walking.
"I mean, Intel is a year a head of schedule."
You honestly think Intel has trashed hundreds of millions in development costs because they are ahead of schedule? Hmm. Are you sure that makes sense?
On the other hand, have you noticed the latest P4 heat sinks? I just saw one the size of an econo can of whole peeeled tomatos. You can find those next to the 500watt power supplies. Hmm, and notice that these formerly insignificant items are now costing more than the CPUs themselves?
I believe the "year ahead" theory is an amusing bit of spin at best.
GPUs? Northbridge? Southbridge? Memory? Every other chip? Pretty much every component in computers have relied on the process shrinking to pack even more into even less space.
This may have a very profound effect on computers in general. No longer able to rely on computer speed doubling all the time, priorities may begin to change. Interfaces will remain more stable (sockets, RAM slots, PCIe).
We will need to design other ways to improve computers. Better instruction sets, clustering, dual cores, multi-CPU boxes... Computers may actually become something you can "invest" in, meaning you can pay $10000 today and it'll still be worth something in 5-10 years. It'll be more like an old car and a new car.
With processing tech leveling out, we may also see a price drop. Plants are paid up, investments recovered, difficult to compete on being faster, shrinking process doesn't lower cost? It has all the qualities of a price war.
Overall, people have been asking since when, the 40s? Where would this stop? Perhaps we're close. Don't believe it'll just keep going because we say it will.
The airplane (read: air travel, not military) industry was booming like that after WWII, but then flatlined. The last significant invention there was the Concorde. The rest is just electronics, but the flight is the same.
All that being said, I'm still optimistic for the future. With dedicated devices we can still do far more than we do today, even though the massive power increase in general purpose processors have made us almost forget that.
Encode full HDTV (1920x1080x24p) to MPEG4 AVC in real time on my PVR? I might get 0.1 frame/s on my Athlon, but a chip can handle it without breaking a sweat. Run IPSec on every connection on a GigE line? My NIC will handle it. We may simple have to undo a little of the past where everything has been off-loaded to the CPU instead.
Kjella
Live today, because you never know what tomorrow brings
I think Intel's move here indicates that they recognize their monopoly position is about to go away. AMD is eating their marketshare from the bottom end and IBM's POWER technology is positioned to eat it from both the bottom and top end.
As reported on Slashdot last week, there was a week in April where AMD-based PCs actually outshipped Intel-based PCs.
On the IBM side, things look downright scary for Intel. POWER scales from cell phones on up to what will be the world's most powerful supercomputer. Microsoft's next XBox is announced to use "IBM technology", but everyone knows that means they're going to be using POWER chips. Microsoft's design on the XBox is to be the home entertainment hub, replacing the PC in that role.
As IBM and AMD grow stronger, Intel's fortunes will become increasingly tied, and limited, to the relevence of PCs.
Can an adapter be developed so that old PCs could be fitted with 2 processors, without changing the software?
The OS will see that the processor is only one, running at certain speed. The processor would have to be identical, for sure. The parallel processing capability should be transparent to the OS, so that old software does not need to be rewritten to harness the power. I dont think this idea is different from the dual-core idea.
Can it extend the life of old PC, like the one that I have (Pentium III 450)?
I don't think Sun will die. The new UltraSparc V when it comes out will be 30 times faster than the current UltraSparc III. I think that would grind the new Intel processors.
Since I work with IBM POWER4-based dual-core systems, which have been available for over two years now, every day, this Intel announcement (and the Sun announcment about US-IV, in fact) is a big yawn for me. But the POWER 970 chip is real - it will be in the IBM BladeCenter JS20 blades, and as I understand it, is the CPU in the Power Macintosh G5.
DDR2 isn't much better than DDR1, but it's much more expensive.
Actually, with posted writes, DDR2 performance is much better.
As far as being more expensive, that's always true when any new memory technology is introduced. However, within a year of introduction you reach the price parity cross-over. Since that's before the dual core shows up, it could actually be more expensive to be using the trailing-edge DDR1 technology.
Finally, FB-DIMMs are DDR-2. They just add a buffer to the DIMM for all the address, data and control lines, not just some of them like Registered DIMMs.
to the Engineer, the glass is neither half full nor half empty. Its just two times too big.