The Economics of Chips With Many Cores

meanonymous writes "HPCWire reports that a unique marketing model for 'manycore' processors is being proposed by University of Illinois at Urbana-Champaign researchers. The current economic model has customers purchasing systems containing processors that meet the average or worst-case computation needs of their applications. The researchers contend that the increasing number of cores complicates the matching of performance needs and applications and makes the cost of buying idle computing power increasingly prohibitive. They speculate that the customer will typically require fewer cores than are physically on the chip, but may want to use more of them in certain instances. They suggest that chips be developed in a manner that allows users to pay only for the computing power they need rather than the peak computing power that is physically present. By incorporating small pieces of logic into the processor, the vendor can enable and disable individual cores, and they offer five models that allow dynamic adjustment of the chip's available processing power."

Comments

How is this new?

[lintux]

IIRC this is done in mainframes for *ages* already...

Re:How is this new?

[jorenko]

Except that they do this because manufacturing chips is not an exact science -- some turn out better than others, and these are able to handle higher clock speeds with less chance of failure and less power usage. Thus, the quality of each individual chip determines its clock speed and its price. While the enthusiast will always be able to increase that with no problem 90% of the time, that's quite a different thing from selling a chip that's supposed to be turned up. These would need to be good enough to handle the heavier usage from any user at all, and fully supported all the while.

Hardware DRM....

[foobsr]

In related news, an initiative of car manufacturers spearheaded by Ford has introduced an enabling 'cylinder per need' model. Car performance is wirelessly monitored in real time to give the customer the option to add in additional power according to his needs if he has signed to a plan designed to optimally fit his profile (composed on his overall lifestyle information). This also creates a new exciting opportunity to reduce individual carbon tyreprints for the consumer.

CC.

Re:Hardware DRM....

[peas_n_carrots]

Those 100-cylinder engines sure are light. After all, the metal necessary to build such an engine would only make up the majority of the weight of the car. Use 10 cylinders to drag around the rest of the 90, now that's efficiency.

Re:Hardware DRM....

[gnasher719]

It's a valid point. Certainly the European car manufacturers have a "gentleman's agreement" to limit their high-end sports cars to a maximum speed of 155mph (around 250km/h). Now, I know that I wouldn't use that kind of power every day, but it would annoy me to know that the car was capable of more but prevented from doing so by an artificial limitation. If I'm paying for a 500bhp car, I want it to run like a 500bhp car... I suppose people like you are the reason for the limitation.

Re:Hardware DRM....

[LWATCDR]

Okay even turning off cores that you don't isn't dump. A lot of the time a pc really is just doing nops waiting for you to do something. Powering down cores when they are not needed will save power just like turning off cylinders in an engine...
What is dumb is this pay me to turn on more cores idea.
It really goes counter to the idea of of OWNING or BUYING a pc. If I BUY the computer I OWN the computer. I shouldn't have to pay you to unlock some part of the that computer.
Yes it is going back to the days of the Mainframe and frankly I don't think that is a good idea.

New form of overclocking - "over coring"

[Xhris]

This should lead itself to a whole new form of hacking - buy the 10 core system and tweak it to use all 100

same old as software rental...

[k-zed]

I don't want to "rent" the processing power of my own computer, thank you. Nor do I want to "rent" my operating system, or my music, or movies. I buy those things, and I'm free to do with them as I wish.

Renting your own possessions back to you is the sweetest dream of all hardware, software and "entertainment" manufacturers. Never let them do it.

Re:same old as software rental...

[markus_baertschi]

For the individual, personal computer, such a model will not fly, as outlined.

However, in the enterprise market this is already there. IBM is using such a 'on-demand' model for its Series P hardware since a couple of years. For a small fee, IBM is installing a bigger configuration (CPU, memory) than the customer bought. The additional hardware is used automatically in case of a failure (built-in replacement parts) or can be unlocked by the customer on the fly.

In the enterprise case it makes sense:

• In enterprise servers the hardware cost is small, compared to the engineering cost. So installing additional hardware does not cost much. A GB of memory costs much more for a high end Unix server than for a PC, even if the technology of the components (simm's) is the same. The difference is in the much lower number of these servers sold and the additional complex engineering needed to build these machines.
• The additional hardware is already there and can be unlocked and added to the configuration on-line. For man enterprise applications this alone is a huge advantage as maintenance windows are scarce. Typically you have a maintenance window four times a year between Sunday 23:00 and Monday 02:30.

Markus

Re:same old as software rental...

[ozmanjusri]

Life just sucks that way

Microsoft != Life.

erm...

[Anonymous Coward]

So, Intel is going to charge us less for a processor with 4 cores because we can turn three off most of the time? Or is the power saving supposed to make the cost of the chip less prohibitive?

Maybe it'll be a subscription service, 9.99 per month and .99 cents per minute every time you turn another core on.

Re:erm...

[gnasher719]

So, Intel is going to charge us less for a processor with 4 cores because we can turn three off most of the time? Or is the power saving supposed to make the cost of the chip less prohibitive? First, it seems you are under the impression that this might be Intel's idea. It is not. Second, turning off cores is stupid. If you want to reduce performance of a multi-core chip, you reduce the clock speed as far as possible. Four cores at a quarter of the maximum clock speed use lots less electricity than one core running at full speed.

You know what I don't get?

[Moraelin]

You know what I still don't get? Why's everyone acting like dividing a CPU into several separate cores is a good thing?

Let me compare it to, say, a construction company having a number of teams and a number of resources, e.g., vehicles:

1. One team, 4 vehicles. That's classic single core. Downside, at a given moment it might only need 2 or 3 of those vehicles. (E.g., once you're done digging the foundation, you have a lot less need of the bulldozer.)

2. Two teams, can pick what they need from a common pool of 4 vehicles. That's classic "hyperthreading". Downside, you're not getting twice the work done. Upside, you still paid only for 4 vehicles, and you're likely to get more out of them.

3. Two teams, each with 4 vehicles of its own. They can't borrow one from each other. This is "dual core." Downside, now any waste from point 1 is doubled.

But the one I don't see is, say,

4. Two teams with a common pool of 8 vehicles. It's got to be more efficient than number 3.

Basically #4 is the logical extension of hyperthreading, and it seems to me more efficient any way you want to slice it. Even if you add HT to dual-core design, you end up with twice #2 instead of #4 with 4 teams and a common pool. There is no reason why splitting the pool of resources (be it construction vehicles or execution pipelines) should be more efficient than having them all in a larger dynamically-allocated pool.

So why _are_ we doing that stupidity? Just because AMD at one point couldn't get hyperthreading right and had its marketers convince everyone that worse is better, and up is down?

Re:You know what I don't get?

[lintux]

You know what I still don't get? Why's everyone acting like dividing a CPU into several separate cores is a good thing?

AFAIK adding more MHz was getting more and more complicated, so it was time to try a new trick.

Re:You know what I don't get?

[RzUpAnmsCwrds]

Your metaphor on multi-issue CPUs is interesting, but not necessarily valid.

Instruction scheduling is the biggest fundamental problem facing CPUs today. Even the best pipelined design issues only one instruction per clock, per pipeline (excluding things like macro-op fusion which combine multiple logical instructions into a single internal instruction). So we add more pipelines. But more pipelines can only get us so far - it becomes increasingly more difficult to figure out (schedule) which instructions can be executed on which pipeline at what time.

There are several potential solutions. One is to use a VLIW architecture where the compiler schedules instructions and packs them into bundles which can be executed in parallel. The problem with VLIW is that many scheduling decisions can only occur at runtime. VLIW is also highly dependent on having excellent compilers. All of these problems (among others) plagued Intel's advanced VLIW (they called it "EPIC") architecture, Itanium.

Another solution is virtual cores, or HyperThreading. HTT uses instructions from another thread (assuming that one is available) to fill pipeline slots that would otherwise be unused. The problem with HTT is that you still need a substantial amount of decoding logic for the other thread, not to mention a more advanced register system (although modern CPUs already have a very advanced register system, particularly on register-starved architectures like x86) and other associated logic. In addition, if you want to get benefits from pipeline stalls (e.g like on the P4), you need even more logic. This means that HTT isn't particularly beneficial unless you have code that results in a large number of data dependencies or branch mispredicts, or if pipeline stalls are particularly expensive.

Multicore CPUs have come about for one simple reason: we can't figure out what to do with all of the transistors we have. CPUs have become increasingly complex, yet the fabrication technology keeps marching forward, outpacing the design resources that are available. This has manifested itself in two main ways.

First, designers started adding larger and larger caches to CPUs (caches are easy to design but take up lots of transistors). But after a point, adding more cache doesn't help. The more cache you have, the slower it operates. So designers added a multi-level cache hierarchy. But this too only goes so far - as you add more cache levels, the performance delta between memory and cache decreases, because there's only a finite level of reference locality in code (data structures like linked lists don't help this). You may be able to get a single function in cache, but it's unlikely that you're going to get the whole data set used by a complex program. The net result is that beyond a certain point, adding more cache doesn't do much.

What do you do when you can't add more cache? You could add more functional units, but then you're constrained by your front-end logic again, which is a far more difficult problem to solve. You could add more front-end logic, which is what HyperThreading does. But that only helps if your functional units are sitting idle a substantial percentage of the time (as they did on the P4).

So you look at adding both functional units and more front-end logic. You'll decode many instruction streams and try to schedule them on many pipelines. This is what modern GPUs do, and for them, it works quite well. But most general-purpose code is loaded with data dependencies and branches, which makes it very difficult to schedule more than a very few (say, 4) instructions at a time, regardless of how many pipelines you have. So, now, effectively, you have one thread that is predominantly using 4 pipelines, and one that is predominantly using the other 4.

Wait, though. If one thread is mostly using one set of pipelines, and one is mostly using the other, we can split the pipelines into two groups. Each will take one thread. This way, our register and cache systems are simpler (because

Re:You know what I don't get?

[peas_n_carrots]

"..because AMD at one point couldn't get hyperthreading right and had its marketers convince..."

Quick history lesson. Intel tried pawning off hyperthreading to the market. If you mean that AMD should have done hyperthreading, perhaps you should look at the reviews/benchmarks to see that it reduced performance in many cases. In the future, more software might by able to take advantage of increased thread parallelism, but that future is not now, at least in the x86 world.

Re:You know what I don't get?

[Antity-H]

noone is chump enough to make something totally different that nothing runs on.
I guess that's why IBM did not develop the cell processor which is therefor not used in PS3s or why no supercomputer is built using it.

All this also explains why IBM did not develop a new product line of cell based blade servers. And neither are grids being built around cell based servers.

Of course even if IBM did develop it and sony did use it in the PS3, it would be unable to run anything which is why there isn't any game for the PS3 or why there are not linux distribution for the PS3.

Sorry, but a different architecture doesn't mean nothing runs on it, nor does it mean noone will develop for it if the promised power is cheap and proficient enough.

Re:You know what I don't get?

[Khelder]

Just wanted to chime in and say I can't remember the last time I saw a comment this good on /. Thanks!

Requires a near-monopoly

[Ed Avis]

In mainframes you have pretty much a single vendor (IBM). Even in the days of Amdahl and Hitachi, once you were committed to a single vendor they had a lot of market power over you. So the vendor can set its own price, and squeeze as much money out of each customer as possible by making variable prices that relate to your ability and willingness to pay, rather than to the cost of manufacturing the equipment.

In a competitive market where 100-core processors cost $100 to produce, a company selling 50-core crippled ones for $101 and 100-core processors for $200 would quickly be pushed out of business by a company making the 100-core processors for $100 and selling them, uncrippled, for $101. I expect the Intel-AMD duopoly leaves Intel some scope to cripple its processors to maintain price differentials (arguably they already do that by selling chips clocked at a lower rate than they are capable of). But they couldn't indulge in this game too much because customers would buy AMD instead (unless AMD agreed to also cripple its multicore chips in the same way, which would probably be illegal collusion).

Compare software where you have arbitrary limits on the number of seats, incoming connections, or even the maximum file size that can be handled. It costs the vendor nothing more to compile the program with MAX_SEATS = 100 instead of 10, but they charge more for the 'enterprise' version because they can. But only for programs that don't have effective competition willing to give the customer what he wants. Certainly any attempt to apply this kind of crippling to Linux has failed in the market because you can easily change to a different vendor (see Caldera).

Re:Requires a near-monopoly

[Anonymous Coward]

To be fair to the graphics companies, they sometimes at least did that because of relatively low yields. If you can take a chip that has ten pipes, two of which are faulty, and disable those two faulty pipes, you've effectively created an eight pipe chip for nothing. This reduces the overall cost of producing a single chip, because a partial failure is still usable.

This is also why Sony used a Cell with only seven SPUs instead of the eight designed on the chip: if a single SPU fails (which is much more likely than none) in test, the chip is still usable. It pushes up yields significantly.

IOW: you're comparing the wrong business model. The model you're describing is "oh, this chip isn't quite up to spec, let's put it in a lower spec card where it will meet the spec", rather than "let's sell a fully capable chip deliberately crippled, and re-enable the crippled part later if the customer pays for it."

Re:Requires a near-monopoly

[ElDuque]

That's a common misconseption there - prices in a competitive market are based on the consumer's willingness to pay, and nothing else. The cost of manufacturing equipment would only come into play in a monopoly situation, where the seller is able to "set" prices (because she will be sure to set them higher than her per-unit production costs.)

This is the same misconseption people often apply to baseball player salaries - they do NOT drive ticket prices. Baseball ticket prices are set at the highest level the market will bear - a price that is determined as consumers make decisions between countless sources of entertainment and leisure.

What is confusing is that the quality of a product (and therefore the market demand for it, sometimes) is often related to the cost of production, so it looks like production costs set prices. But remember when Homer designed a car? It was $80,000, and no one wanted to buy it at that price! The consumers decided there were better uses for their car-buying dollars. This is a perfect (although fictional) illustration of why costs != prices in a competitive market.

Re:Requires a near-monopoly

[Lonewolf666]

Interestingly, of late, it is AMD that is trying to create product differentiating by crippling their processors, or at least by selling processors with one core switched off. They're trying to do this by selling "tri-core" processors based on their Barcelona/Phenom cores, which are nothing much an actual quad-core with a core turned off, either deliberately, or because it is defective. They probably want to position this as a mid-range offering, to make it more competitive to Intel's relatively cheaper quad-cores.

I guess it is, first of all, a way to get money for processors that would have to be thrown away otherwise. Some money for a "tri-core" is better than no money for a piece of waste silicone.
On top of that, there may be some crippling of intact quad-cores if there is more demand for the cheap "tri-cores". But I doubt that is the main reason.

Re:Requires a near-monopoly

[Firethorn]

Didn't someone predict that we'd only ever need 128 MB of ram and that more ram would be superfluous for most consumers?


I believe that you're thinking about the quote commonly attributed to Bill Gates '640k will be enough for anyone'.

Right now there are a lot of flashy games out there. Users may want to run many more applications at once (or more likely turn on M$ poorly executed eyecandy and not notice their computers slowing down).

Until 3D acceleration is so good that you can't tell it from real life, all other tasks are 'instant' from the viewpoint of the use, greater speed will be in demand.

Monitor resolutions are still creeping up, placing more demand on video card processing power, games are being produced that utilize sophisticated physics engines*.

Still, consider that non-resource intensive games like Bejeweled will often outsell a resource intensive games like Supreme commander, Crysis, or Bioshock.

It will probably happen gradually, as most customers are already low-end.

Agreed, even today I'll recommend economy machines to people who don't play 3D games or do something like video ending.

*To the point that I now have 3 games that you can add a daughtercard to the computer for offloading the processing.

Re:Requires a near-monopoly

[Firethorn]

On top of that, there may be some crippling of intact quad-cores if there is more demand for the cheap "tri-cores".

It'd probably be more profitable to up the price of the tri-core a nitch*. A couple bucks would reduce the demand for the tri-core, as some people decide to settle for a dual core instead and some decide that the now smaller difference between a tri-core and a quad core makes it worth it to buy a quad core.

IE:
Quad: $100, Tri $75, Dual $50 - not enough triples to meet demand
Quad $100, Tri $77, Dual $50 - Fewer people buy the tri because the quad and duals are 'good deals' in comparison.

Whether more fall back to the dual or move up to the quad, I can't really say. Of course, that can be adjusted a bit by minor variations in price there. Just beware of competition there as well.

*smaller than a notch. ;)

Why?

[RuBLed]

If one could make a 5 core processor for the price of $300 and be able to sell it with 5 cores enabled to a customer for $600. Why would he sell the same unit for $400 with only 2 cores enabled?

Wouldn't he profit more if he could sell the 5 core processors all at $600 and make a separate 2 core processor for the price of $200 and sell it for $400?

Well if they're going to rent it (as some of TFA said), it would make sense but if they're not, then it would be a profit not maximized.

Re:Why?

[Anne Thwacks]

Because in reality, it costs $4.99 to make the chip, and $10,000,000 to design it.

The cost of designing one core is the same as the cost of designing 10 or 100 cores, because copy and paste was invented several years ago. The cost of adding a core to the design is about 1%.

There might be a case for powering down unused processors to save energy, and there is a case for selling cheaper processors with reduced core counts where some cores don't work, but there is no case for disabling working processeors for economic reasons.

Sun's Niagra technology differs, cos it has "virtual cores" which gives you more virtual cores but slower. Its very good if you multi-thread (run apache) and p*ss- poor if you dont (run Windows).

Re:Why?

[Alsee]

The cost of designing one core is the same as the cost of designing 10 or 100 cores, because copy and paste was invented several years ago.

Copy-pasting a hundred cores will cost almost ten times as much as copy-pasting 10 cores because you have to pay the patent holder who invented copy-paste.

-

Re:Why?

[mentaldrano]

While you are right about research cost ($10,000,000) vs production cost ($4.99), your point about adding cores is not well made.

Yes, you can just copy and paste the individual core design, but heat dissipation, core interconnect, and off-chip bandwidth will kill you if all you do is simply "paste another one on." These problems are easy to get around for few-core chips, say 2-4, but once you go farther than that, it takes real design innovation to stay afloat. Thermals require dynamic core underclocking, work distribution (keep hard working threads on widely separated cores), and split power planes. Core interconnect uses things that sound an awful lot like Ethernet, including routers with routing tables! Off chip bandwidth requirements bring in huge caches and dual buses.

Look at the huge deal AMD made of its "native quad core" design vs. Intel's quad core chip. Sure, it didn't end up giving them a huge performance advantage, but this is the way things are going for many-core chips, and AMD does have a head start on production.

This is a truly stupid idea

[SmallFurryCreature]

In theory it makes sense and some of you might point at mainframes as an example. However that would like comparing cars to trucks (real trucks not big cars), they are both vehicles and a company might use both but their usage is totally different.

PC's just ain't upgraded, either they are good enough or they are replaced. I love building my own computer but am not as crazy as to replace the CPU whenever a new clockspeed comes out and this means that even a self-builder will often have to bite the bullet and just replace everything.

Be honest, how often in business do you upgrade your desktops by replacing the CPU?

We can test this easily, in the era of the P3 a lot of office systems were DUAL ready, so that when your needs increased you could ad another P3 and have lots more power. How many of you did that with a P3 that had been in the office for more then a year?

This scheme seems like overthinking the problem. PC's in my experience either last until they die and by that time it cheaper to buy new then upgrade/repair, or they are simply replaced with the latest shining model because tech moves so fast that upgrading just the CPU will turn everything else into a bottle neck. Just check how many different types of memory we have had over the years. Would you really want a quad core on your IDE-33 motherboard? Play DVD's on a single speed cd-rom?

Either you need all the cores now, or by the time you activate them because your apps need them everything else will need to be upgraded too and a brand new CPU will be available that is far better AND cheaper.

But in a way we have had this solution for a long time now, but instead of activating extra cores when paid for, chipmakers instead sell defective chips for a reduced price so your still got a 4 core inside your machine but only 2 actually function (not sure wether this happens with entire cores but it is offcourse the case with cache memory).

I don't see this happening, especially if you consider that an army of nerds would be trying their best to break the enabling code to get their extra cores for free, just see what happened with the "dual" P2 and cheapo P3's, Intel would have a heart attack.

Do I understand this right?

[Dr. Spork]

TFA is written really badly, but from what I gather, the "more advanced" models of figuring out how much to charge for chips goes like this:

1. Everybody gets the same chip, but it will be crippled unless you pay the highest price.

2. Everybody gets the same uncrippled chip, but there's a FLOPS meter on it that phones home, and you pay Intel according to the amount of numbercrunching your chip did for you.

Both of these models seem completely retarded to me, although the first is already sort of in use in the CPU/GPU market. Have modern processors overshot our needs by so much that our big worry now is to find innovative ways to cripple them? If so, maybe this processor war we're fighting is ultimately not even worth winning.

Calculators

[Detritus]

Someone already mentioned mainframes. Something similar is often done with calculators. Rather than design a new chip for each model, they design a single chip with all of the features. In mid-range and low-end models, it is crippled by the design of the keyboard and/or jumpers. It is often cheaper to dumb down a single hardware design than to produce unique designs for each segment of the market.

Re:How is this [business model] new?

[ozmanjusri]

Why not their economic model?

Because it's dumb.

In 1999 I paid about AU$600 for a midrange Pentium Pro CPU. In 2008, I bought a midrange Xeon Dual-core for the massively increased price of... AU$600.

In 2000, I bought a shiny new Intergraph TDZ2000 with two PII 350s for the bargain cost of just $5,000. Now, Apple is prepared to sell me a Mac Pro with two 2.8GHz, quad core Xeons for the stupefying price of $2,799.00.

Now, explain to me again why it would be in my best economic interest to buy a computer with cores that could be disabled if I don't pay my rent?

Come on, this "reseracher" proposes DRM for CPUs

[eiapoce]

In the sense of Digital Restriction Managment a part of the article states:

This can be accomplished with small pieces of logic incorporated into the processor that enables the vendor to disable/enable individual cores Now think once or maybe twice about it. The situation could be that of a manager of a datacenter, which probably handles sensitive data, and lets the vendor mess realtime whith the CPUs (and possibly the data) driving the system just because he wants to save a few hundred dollars on a digitally castrated chip. Though idiotness is a widespread illness I don't see who could be such a moron. This could only be acceptable by the CIA or the vendors themselves; not long ago the pentium whith Serial on chip were rejected by the market, this is much worse.

This business model is dead meat to me. I think that the market will continue to offer processors classified on the maximum data processing rate and pricing them accordingly. I don't see a future for this Processor Restriction Managment nor for the career of the guy who wrote the article in the first place. Suggestion: after he's been dumped from University don't get him as datacenter manager.

S/W licensed per processor

[petes_PoV]

In a competitive market where 100-core processors cost $100 to produce, a company selling 50-core crippled ones for $101 and 100-core processors for $200 would quickly be pushed out of business by a company making the 100-core processors for $100 and selling them, uncrippled, for $101.

And when your software is licensed per processor at (let's say) $100 per cpu, your extra, unwanted, 50 processors quickly become a burden. I'd be willing to pay more for a crippled processor if it saved me money elsewhere, and there was no way to slice up domains to reduce the liability

Re:S/W licensed per processor

[afidel]

Bullshit, the biggest cost vendor who licenses per CPU actually licenses per core, Oracle! On Windows it's one license per 2 cores, everywhere else it's .75 per core except Sun T1/T2 where it's .25 per core.

Re:S/W licensed per processor

[MBGMorden]

Your statement is *USUALLY* right, but not universally right.

IBM for example when licensing some stuff (namely Lotus Domino): they go by performance units.

A single core x86 CPU would be 100 units per core. Dual-core CPU's would be 50 units per core (notice that they work out to the same). Quad-cores however are also 50 units per core, so while a Single and Dual Core chip cost the same, Quad Cores end up costing twice as much in license fees.

They even have some architectures where it changes to different values (I think one of them, maybe SPARC chips, were 100 units per core for Single Core and 70-80 units per core for Dual Cores - making a Dual Core more expensive, but not twice as expensive.).

Aside from them most other vendors I've dealt with who do license by the CPU do it by actual processor and not cores. As a result when we are never worried about buying a quad-core box for a server even if it's a little more than the particular situation calls for, whereas generally I won't buy multi-CPU boxes unless I think I will actually need them. I'm sure there are exceptions though as there are to everything.

One are where this gets interesting: virtual machines. Our trend lately has been to buy a huge hulk of a server and then have CentOS + VMWare Server split it up into numerous virtual machines to do smaller jobs. The hiccup there is that for most software licenses, you have to go by the underlying hardware. So if I have a quad-box that is hosting say 6 installations of MS SQL Server, then I'd have to pay the quad license for each one, even though the power of the underlying quad chip system is being split up to probably less than a single CPU system would offer if it weren't virtualized.

Here's a better business model

[WaZiX]

1) Sell your super high power 20 cores CPU uncrippled.
2) Make a platform where researchers can rent CPU power.
3) Allow your customers to rent their unused CPU power/cores.
4) Charge double what you give to your customers to the researchers.
5) Profit! (From both the sale and the rental afterwards).

And there is no ?...

Re:How is this [business model] new?

[argiedot]

Now, explain to me again why it would be in my best economic interest to buy a computer with cores that could be disabled if I don't pay my rent? I suppose because you could just buy the ones with some cores disabled and get someone who knows stuff to enable them again, like the way people did for some of the older nVidia cards that had some things disabled. Or maybe I don't know anything about how the two things work.

Nice idea, then reality hits

[JRHelgeson]

So, what would happen if the Microsoft DRM update management and monitoring "feature" has a "bug" and hits 100% utilization as it tries to verify the authenticity and my right to possess my entire music collection... do i have to pay a processor tax for that? What about a runtime condition? An app locks up and hits 100% utilization until it is killed. OOPS, I need to ante up for the Tflop tax. Or when I file my annual procmon return I cna apply for earned op/sec credit, filing as head of household...

I'm not about to pay a tax on other peoples poorly written software.

STOP TAGGING whatcouldbpossiblygowrong ALREADY

[quitte]

really. STOP IT!

Crippleware...

[Bert64]

This is crippleware, and a terrible idea for the average consumer...
Paying more for a product that costs the same to produce, or potentially even less because they don't have to disable the extra cores is a terrible rip off, and it happens already...

The same people who currently overclock, will buy the cheaper cpus with cores disabled and re-enable them... You will also get third parties who make a business out of doing the same, tho without the "exceeding design spec" risks of overclocking.

Personally, I will never pay more for a more expensive version of the same product, i will buy the cheapest available just as soon as people have worked out how to re-enable the disabled cores, and i will help my less technical friends do the same.

Would we know the difference?

[SanityInAnarchy]

I know that on Linux, I cannot immediately tell the difference between an SMP-enabled kernel on a single-core Hyperthreading system, and an SMP-enabled kernel on a dual-core system with no hyperthreading.

In either case, I'm fairly sure I see at least two items in /proc/cpuinfo, I need an SMP kernel, etc. So if someone (Intel) suddenly decided to make a dual-core hyperthreaded design in which the "teams" actually shared a common pool, would I notice, short of Intel making an announcement?

As for your assertion, a quick scan of Wikipedia suggests that you're a bit naively wrong here. (But then, I'm the one pretending to know what I'm talking about from a quick scan of wikipedia; I suppose I'm being naive.) Wikipedia makes a distinction between Instruction level parallelism and Thread level parallelism, with advantages and disadvantages for each.

One of the advantages of thread-level parallelism is that it's software deciding what can be parallized and how. This is all the threading, locking, message-passing, and general insanity that you have to deal with when writing code to take advantage of more than one CPU. As I understand it, a pipelining processor essentially has to do this work for you, by watching instructions as they come in, and somehow making sure that if instruction A depends on instruction B, they are not executed together. One way of doing this is to delay the entire chain until instruction A finishes. Another is to reorder the instructions.

But even if you consider this a solved problem, it requires a bit of hardware to solve. I'm guessing at some point, it's easier to just throw more cores at the problem than to try to make each core a more efficient pipeline, just as it's easier to throw more cores at the problem than it is to try to make each core run faster.

There's also that user-level interface I talked about above. With multicore and no hyperthreading, the OS knows which core is which, and can distribute tasks appropriately -- idle tasks can take up half of one core, the gzip process (or whatever) can take up ALL of another core. With multicore and hyperthreading, the OS might not know -- it might simply see four cores. And with multicore, hyperthreading, and shared pipelines, it gets worse -- as I understand it, there's no longer any way, at that point, that an OS can specify which CPU a particular thread should be sent to. Threading itself may become irrelevant.

Well, anyway... What confuses me is that we still haven't adopted languages and practices that naturally scale to multiple cores. I'm not talking about complex threading models that make it easy to deadlock -- I'm talking about message-passing systems like Erlang, or wholly-functional systems like Haskell.

Hint: Erlang programs can easily be ported from single-core to multi-core to a multi-machine cluster. Haskell programs require extra work at the source code level to be made single-threaded, and can (like Make) use an arbitrary number of threads, specifiable at the commandline. They're not perfect, by far; Haskell's garbage collector is single-threaded, I think. But that's an implementation detail; most programs in C and friends, even Perl/Python/Ruby, will not be written with multiple cores in mind, and, in fact, have single-threaded implementations (or stupid things like the GIL).

Yes and no

[Moraelin]

Quick history lesson. Intel tried pawning off hyperthreading to the market. If you mean that AMD should have done hyperthreading, perhaps you should look at the reviews/benchmarks to see that it reduced performance in many cases. In the future, more software might by able to take advantage of increased thread parallelism, but that future is not now, at least in the x86 world.

While I'll concede the point that Intel's first implementation was flawed, you can't judge and damn a technology for all eternity just by its first implementation. In the meantime even Intel's competitors (e.g., Sun) are implementing it, so it can't be that horribly worse than nothing.

Plus, then by the same kind of historical reasoning we should have said goodbye a long time ago to such stuff as:

- any kind of computing or calculating machines. After all, Babbage tried pawning off that idea to the market, and his implementation was never even finished.

- heavier than air airplanes. The first attempts with kites and bird wings were an outright disaster. We should have buried that idea right there and then.

- using rockets for space travel. There was this medieval Chinese dude who tried it first, with completely disastrous results.

- breech loaded guns. The first attempts had _major_ problems with sealing the barrel, because of poor tolerances.

- cavalry. It just wasn't that horribly good before it successively also got a good saddle, horseshoes, stirrups, and specially bred horses. There's a reason why the Romans created their empire with elite infantry, and the cavalry was just some specialized auxiliary.

- in fact, even earlier, we shouldn't have had even chariots. I mean, until someone invented a harness that allowed horses to pull one, it was pretty much useless. We know that the Sumerians tried using oxen there, and it couldn't have been that horribly effective. Should have discarded that idea right there and then.

- agriculture. Until the right plants, irrigation and cats became available, it was very much a losing proposition wherever it was tried.

Etc, etc, etc.

Dude... wait, what?

[The Master Control P]

So Intel is going to design a CPU with N cores on it, then add hardware that disables half of them, then manufacture the chip with all N cores and sell it for less, even though it actually costs more to design/build because of the added hardware to cripple it, then try and make us pay for access to the other half of the cores and hope we don't notice that our computers have suddenly become a constant expense instead of a one-time purchase?

And moreover, they apparently forgot which problem they're trying to solve between paragraphs 4 and 5. They start talking about the real problem of many cores creating a very large space of core/memory architectures that would be difficult to choose between and support. Then they veer off into the rent-your-own-hardware-back-to-you idea and never finish reasoning out just how it would work before they come back. A few minor things they ignored:

• How do they turn cores on? Difficult level: No, you can NOT have a privileged link through my firewall onto my network.
• How do they stop me from hacking it and enabling it all myself? Difficulty level: Mathematically impossible since you can't stop Eve from listening if Eve and Bob are the same person.
• How do they propose to bill me? Difficulty: No, I will NOT let my CPU spy on me.
• Why should I hand you everything you need to force me to upgrade against my will?
• What happens if you go out of business and leave me stranded?
• Even if you don't see what's wrong with charging me continually to access my own hardware, do you actually think I won't?

In conclusion, Profs. Sloan & Kumar of the University of Illinois, I believe the premises and reasoning behind your proposal to be flawed, and the proposal itself to be unworkable and contradictory to openness in computing. Or, as we say on the Internet, wtf r u doin???

Re:How is this [business model] new?

[CheShACat]

That was my first thought. Then my second thought was having to go through the "Intel Genuine Advantage" activation process every 45 minutes.

They already do this...

[Cyno01]

Sort of, as many other people have said, about overclocking and such, its not necesarily a scam, it makes things more cost effective for the company and can benefit the consumer who would take the effort to overclock. Lets say Intel (or AMD, doesnt matter, they both do it) does a run of chips. The specs call for the chip to run at, for simplicities sake, 2Ghz, with stock AMD cooling. But no manufacturing process is perfect, and lets say 25% of the chips arent good enough to run at 2ghz without frying. They then clock these chips to 1.5Ghz and sell them as such. This allows them to do a smaller run of specced 1.5Ghz chips and save themselves money. Its not really a scam, theyve been doing this forever, and arent the only industry to do it either. In this case, the consumer can benifit. Someone can buy a 1.5Ghz chip (although they might have to exchange it till they get one of the ones from the 2Ghz production run), and most of the time it'll run fine at the 2Ghz speed with improved cooling.

Re:How is this [business model] new?

[TheThiefMaster]

I've done this kind of thing. nVidia 6800LE with half it's shader processors disabled (had 4 blocks of 4, 2 blocks disabled), which could have half of those (1 block of 4) re-enabled without issue. Athlon XP 2500+ that could have the FSB changed to 200MHz instead of 166 and it would BECOME a Athlon XP 3200+ (name and all).
And the best one: Two Athlon XP 2400+ cpus that I unlocked with a conductive pen to be Athlon MP 2400+s, and I still use in a dual-cpu board now.

Generally, unlocked or overclocked pc parts burn out faster than if they'd been left alone (e.g. the 6800LE I mentioned died a horrible death, and now doesn't work at all). However if the chip was DESIGNED to be able to be unlocked, it would be perfectly safe.

they buy this because it saves money

[Anonymous Coward]

IBM's been doing this for years with some of their smaller servers http://www-03.ibm.com/systems/i/hardware/cod/index.html/

The cost in IT labor and lost productivity during the downtime that old methods need to add processing capacity can be a *lot* for servers hosting your important applications but its awfully expensive to pay upfront for enough power to keep up with ordering spikes during the Christmas buying season (for example) if that spikes way beyond your normal needs. Much cheaper to pay for only enough to handle your normal needs and then pay for the extra needed to handle oddball spikes only during the time you need it.

There's no way Ticketmaster's IT budget would agree to pre-pay for enough computing capacity to not bog down when the Hannah Montana tickets went on sale but if they could pay for just an hour of it ... much easier sale. Or remember the "performance" of Amazon's servers last Christmas when they put up their special sale items? If they could have just paid for a 24-hr keycode to enter the night you can bet the IT guys would have had a much easier time getting that in the budget.

Or as IBM puts it:
"Imagine you launch a dynamite new Web application for the holiday season, and it's getting more traffic than you expected. What do you do to avoid disruptions in service? You turn on available inactive processors and memory to handle every hit, then turn the extra capacity off when the application requires less capacity in the new year. You pay only for what you have activated.
Or say you tell your business analysts they now have access to all the company's business intelligence data. The danger is that, with your current processor configuration, increased demand could slow response times to a crawl. The solution? You activate reserve processing power to meet the new user demands without disrupting current operations."

The other beauty is that once the computer manufacturer has built in the ability to activate or inactivate processors and memory on the fly those same mechanisms make it natural to shuffle processors and memory between virtualized servers on the machine without restarting them.

And yes, it runs Linux.

Re:How is this [business model] new?

[Anonymous Coward]

Generally, unlocked or overclocked pc parts burn out faster than if they'd been left alone (e.g. the 6800LE I mentioned died a horrible death, and now doesn't work at all). However if the chip was DESIGNED to be able to be unlocked, it would be perfectly safe.

Design is one. Manufacturing is two. Chip manufacturing is not perfect. It is more likely that the disabled parts failed full test, but that parts were still working (and thus make it sellable as a downgraded chip). All you did was enable the defective parts. And then it blew. No surprise there.

Re:How is this [business model] new?

[Anonymous Coward]

Having worked at nvidia, there is a reason those extra TPCs were disabled and its not because of a cripple ware model but because of yield. We cannot produce chips that are perfect all the time. So we settle for chips that are perfect a small percentage of the time, mostly perfect an ok percentage of the time, and half working a good percentage of the time. We then make 3 or 4 different series (GS/GT/GTX/GTS/Ultra) with different TPCs in each series, disable the TPCs in each chip that doesn't work or fails to pass QA and then ship them. If you unlock them, you are frying you working card because some of the faults could be things like "Oops, there was a short in the TPC because the transistors cooked too close to each other" or "Oops, the clock passes too close to the +12V in this module -- if it hits 50 Celcius, it could turn into a short". This model helps products from being prohibitively expensive for a fabless company because we are billed on "silicon wafers used" on not on "number of fault free chips produced".

The rent model is flawed because ...

[BeanThere]

... for CPUs, there are effectively ZERO variable costs to the producer once you've purchased the chip and it's in your hands.

Dedicated circuitry to create artificial scarcity and control actually adds unnecessary costs.

This might be useful in very specific scenarios where somebody, say, owns a supercomputer and rents it out, but even there, I'm sure there are far better solutions that don't involve the CPU hardware.

This is, like you suggest, just a BS wet dream of the manufacturers ... make something once, get money forever.

Right now we probably have few enough major chip vendors that with a little bit of collusion, if they decided not to compete, they could probably pull something like this on us. This doesn't look likely right now, but it seems possible. Hopefully some other (possibly foreign) company would enter the market if that happened. Competition is healthy for a market.

greedy market strategy

[v1]

a unique marketing model for 'manycore' processors

Nothing UNIQUE about this strategy. It's a model growing in popularity. Traditionally companies that wanted to capture several levels of market would make several models of a unit. Like buying a laptop with a better graphics chip or bus speed etc. This cost them more because they had to produce three different units which triples costs on some of their overheads. What this is doing is allowing them to produce one high end product, and configure it easily, post-production, to any of the three units they want to market. The same capabilities are present in all models, but features are disabled/crippled/nerfed in the less expensive models. This allows them to sell their product in the lower cost market without losing sales in their high end market, and without the additional expense of producing several different models.

It's a good idea for the manufacturer, but introduces the problem of what happens when the consumer figures out how to "enable" disabled features in their low end model? This always results in a little war of sorts, where the manufacturer takes steps to make de-nerfing difficult or impossible. It always aggravates the consumer to find out that after he conceded to buying the model that didn't do everything he wanted it to due to cost, CAN do it, it just refuses to. The consumer feels cheated that he payed for a gadget that CAN do what he wants it to, but can't take advantage of it.

Interestingly, it doesn't become a problem until the consumer realizes the product that they were obviously happy to pay the small amount for can do more than they bargained for. The producer would argue that you didn't pay what they were asking for those additional features and so you should not feel cheated, and that you agreed to the advertised feature set when you purchased the product.

The consumer then will try to modify the product to restore the disabled features, and can get upset if it's not possible or is made deliberately difficult.

As much as it causes aggravation in the consumer (that'd be ME) I think it's not a bad idea. What it all boils down to is you can't complain about a product being capable of performing beyond the advertised and accepted expectations at the time you purchased it. You agreed to buy it Just because it's done on purpose does not change the situation. If it CAN do more than advertised and claimed, and you can make it do that, good for you. If you can't, then too bad.

In the end, this DOES result in slightly higher cost for the low end model, because the cost of production (or development) of the low end product is higher than it would have been, if the company had only been making the low end model, and that money ends up in the pockets of the manufacturers who shave overhead on production. So from that point of view it's not a good thing for the consumer, but not for the reason they are seeing.

Re:How is this [business model] new?

[jank1887]

in this case though (the rent-a-core plan) all cores must be fully functional. you're paying for a processor with the potential for using X number of cores. If they aren't all good, they've sold you a defective chip, not a downgraded one. Also, if it's a rental scheme, it can't be a one-way change to upgrade or downgrade. Apparently the process must be fully reversible. Sounds to me like all of that makes it a much more appealing hack target.

Not running on all cylinders

[AlpineR]

Yeah, there are vehicles that already adjust the number of cylinders on demand. One of them is a large domestic SUV. The advertising slogan is something like "Eight cylinders when you need them. Four when you don't."

The benefit to the vehicle owner is lower fuel costs, not an economic model to transmit his cylinder utilization to the manufacturer for a reduction in his vehicle loan payments. That'd just be silly.

If you want a car with less power, you opt for a smaller engine. If you want a single-core processor with less power, you opt for a slower clock speed. The processor you buy might be manufactured alongside the ones sold at higher speeds, but it failed testing or was intentionally crippled to maintain a distinction between high-end and low-end.

Sometimes it's easier for the manufacturer to make everything the same and then cripple or add on to create different classes. Suppose Initech developed a screaming-fast processor that they could sell for servers at $90,000 a piece. It also happens to cost only $90 to manufacturer. They could have priced them at $100 and sold 100 times as many for desktops, but the loss of profit in the server market would make it a loss. So instead they chop off 90% of the cores or reduce the clock speed by 90% on the processors destined for desktops. It's cheaper for Initech than to manufacture a second low-performance design and even the crippled processors are a better buy than the competition. It's economically wise and perfectly moral.

The tricky part with manycore processors is that halving the clock speed is usually more crippling than halving the number of cores. But it all depends on how well the software parallelizes. It could make sense to sell the somecore processors at a discount, and then three years later when the customer is thinking about buying new machines say "We could double the performance of your existing hardware for half the cost."

It might have been dumb of the customer to buy the crippled processors in the first place, but if a competitor can offer uncrippled processors for the same price then the customer won't make that mistake. And sometimes making half of a capital investment now and half later is a good business plan.

Re:How is this [business model] new?

[Dare nMc]

All old rumors. This has come up at many Conferences, although I have never heard from nvida, but Intel, AMD have stated that is just false.
  A) the volume on each line is too high to be shifting silicon between lines.
  B) it just takes too much logic during processing, if A but not B + C... for so few chips that would have a flaw that allowed them to work, but not fully.
  C) flaws in silicon almost never affect just one chip, let alone just one section of one chip. (multi core is still a single chip.)
  D) QC finds most flaws at the wafer level, before ever entering the container, and it is assumed more is affected, they are never touched.
Now this is old, was true 10 years ago, when silicon qualitys weren't as good as now (better silicon yield makes the economics even worse today)

Their is usually more than just the CPU difference between lines, for example you need more cooling and better power source for more/faster CPU's.
I don't know about chips, but Cat does on their Diesel engines. The warranty cost will be higher for a higher powered engine, no matter what it was designed for, so that is part of the cost equation. As well as stepped up Power will compete with the next higher line. So to ease the gap in price from one platform to the next...

Re:How is this [business model] new?

[SharpFang]

Assuming using JUST your logic, every slower processor or chip is one that has failed to be higher processor or chip across the same line of products. We all know that is not the case. There are also market demands that must be met. I have no idea about failure rates but I highly doubt only failed chips make lower tier products. What percentage of what does each company or product line use? No one here has any idea.

The availablity dictates price, price regulates demand.

Take 'LE' and 'GT' releases of NVidia cards. Their difference? GT have all the shader units fully operational, LE have half of them disabled.

Formerly, 'LE' versions of NVidia cards were a major part of the market, and the luxury 'GT' versions with 2 times as many shader units were at least twice as expensive. Nowadays 'LE' are just slightly cheaper from 'GT' and you need a sucker or desperate to buy the 'LE' version because the price gain is very low comparing to the performance penalty. Reason? NVidia improved the manufacturing efficiency, making way fewer faulty units. Supply for GT increased, supply for LE decreased. So we push some 'LE' sector customers into the 'GT' sector, by increasing price of 'LE' and decreasing of 'GT'. The manufacturing costs are the same (it costs exactly the same to produce a 100% working chip as a faulty one...) and people are encouraged to buy the higher-end device due to its lower price, and if they are strictly the 'old LE' market, meaning definitely cheaper product and not willing to pay for either the 'new cheaper' GT nor the 'new more expensive' LE, they will just buy a card from another line, a GT of an older model for example.

Prices of CPU don't increase linearly with speed. The curve of $/MIPS may seem puzzling, but in fact it's the line of yield of the manufacturing in given class.

Of course the market has a very heavy momentum and the price changes don't happen day-to-day. So the temporary differences between supply and demand get filled by units from higher class that have parts of functionality artificially disabled. That's the overclocker's heaven - you just need to 'unlock' the chip and you have a genuine 'higher version'. But that's a matter of pure luck (or insider info or following the news closely) because the chance the part will be 'crippled to lower the price' are worse than that it was faulty in the first place.