Slashdot Mirror
Intel Says Chips To Become Slower But More Energy Efficient (thestack.com)
An anonymous reader writes: William Holt, Executive Vice President and General Manager of Intel's Technology and Manufacturing Group, has said at a conference that chips will become slower after industry re-tools for new technologies such as spintronics and tunneling transistors. "The best pure technology improvements we can make will bring improvements in power consumption but will reduce speed." If true, it's not just the end of Moore's Law, but a rolling back of the progress it made over the last fifty years.
Hopefully if this does happen they will keep making the existing products, at least until they *do* manage performance improvements that catch up / exceed older stuff. Where I work we have lots of customers that *need* more processing power, and efficiency be damned.
William George
I can't imagine that there will simply be zero demand for fast, or faster, chips, regardless of the power efficiency. Some applications just demand it. If Intel won't do it, then someone else will, whether that's AMD or some new competitor in China or wherever.
On the other hand, there's certainly a market for more efficiency, especially in mobile devices, so I can certainly see lines of chips designed for that heading in the way described.
Probably not. Going beyond 5Ghz limit has been a problem for the last decade or so. This is why we have multicore processors. It's easier to add more cores than go to plaid.
Moore's law says nothing about power or speed. It's strictly about the number of transistors on a chip.
"National Security is the chief cause of national insecurity." - Celine's First Law
Optical
:T:R:A:N:S:
A flight from London to New York takes as long today as it did about 50 years ago. But the current planes achieve that more efficiently, with slightly larger windows, and some more pressure and humidity in the cabin. How depressing to think that the computing world might be about to enter a similarly dismal stage as well.
So the plan to make transistors tolerate higher clock speeds by using better materials is not going to happen?
Yet another restating of Moore's Law? The thing gets revised to whatever the latest growth area is.
The original 1965 article it was about "component counts", then it was revised in a later talk to be "circuit density", then revised in 1975 to be "semiconductor complexity", then revised in the later '70s to be "circuit and device cleverness", has been restated yet again when serial devices flatlined in favor of highly parallel chips.
Assuming this goes through the chipset, it will likely be restated again in terms of whatever other factor on the chips continues to grow.
//TODO: Think of witty sig statement
Intel's so far ahead of AMD, they have to roll back the clocks in order to stay competitive.
AMD isn't Intel's competition. Intel needs AMD to prevent Anti-Trust litigation. Intel's competition is ARM and all the OEM's who use ARM based chips. Especially if Microsoft ports full Windows 10 to the ARM. The big draw of ARM is performance/price per watt which is exactly what Intel is shooting for.
"A person is smart. People are dumb, panicky dangerous animals and you know it." - K
Lots of ways to get "speed."
I read that as "slow is the new fast" ... Introducing the new Puntium, now 16% slower.
You laugh now.
Get this new computer which at the push of a button clocks down to 25 MHz for your slow computing needs!
With AMD out of the way Intel can F*** us.
First they cut the pci-e lanes down on a $300-$350+ chip forcing you to pay upped to $350-$400 but then you need jump to $500-$600 to get the same as last gen + a small clock speed boost. This on the server / high workstation side.
On the desktop side they are still on DMI (Now at pci-e 3.0) + 16 PCI-e 3.0 why no QPI to chip set like AMD's HTX?
On the other hand, designs with less energy loss will open up the potential of higher speeds, once the techniques get refined.
One of the (many) limit issues with trying to force current CPUs faster is that the waste energy grows quickly as you increase switching frequency. Energy density becomes a significant problem, and manufacturers are not content with the idea of making all consumer devices use liquid-cooling and/or refrigeration techniques to prevent CPU melt. Take a couple years learning a more efficient set of components and tools, and the cap may raise past something (currently) silly like a passively cooled 8GHz chip.
Only for embarrassingly parallel workloads.
Intel's been shooting itself in the foot with power vs performance for years. AMD was better, Intel reversed course and then beat AMD down. Now Intel's gunning for ARM because ARM is becoming a real threat to their core business. How many phones have Intel chips? How many tablets? Notebooks are moving towards ARM as well. Imagine an ARM based server farm. ARM is moving up the food chain into Intel's core business, and doing so with a class of processors Intel can't match.
The cesspool just got a check and balance.
No, a lot of applications don't scale well across multiple cores / CPUs.
William George
And if they're having a significant reduction in power consumption, then adding more cores gets all the easier.
Its always seemed to me that the best approach to processing is to offer a variety of cores and let the scheduler handle what to put where. You can have one or two extremely fast cores, half a dozen moderate speed cores, and dozens or more low speed cores - why insist that all cores be the same in "general purpose" computing?
It's times like this I wish I had a friend named 'The Professor'.
Except they have, in terms of work done per clock (even ignoring multicore). A Haswell 1.2 ghz can achieve the same sort of results as a 3.0 ghz AMD core from 5 years ago in a balanced set of CPU constrained work. It actually comes out ahead in a number of specific workloads. Note I'm comparing to a core significantly older, with less cache for the sake of demonstrating only the senselessness of being fixated on clock, not saying this is a fair Intel v. AMD comparison.
On the other hand, a 1.2 GHz AMD K7 back in the day could beat a 3.0 GHz Pentium 4 of the same time. There's a lot more to processor performance than clockspeed.
XML is like violence. If it doesn't solve the problem, use more.
the software side has been storing up efficiency improvements for a long time. Just get rid of the extras, like bloatware, and hastily programmed apps, and nobody will notice.
Yeah because corporations don't want energy efficient machines....moron...
no, because if they did that they they would hold a monopoly on desktop / laptop CPUs. Then they would be regulated as a monopoly, and could no longer get away with their abusive business practices.
Where I work we have lots of customers that *need* more processing power, and efficiency be damned.
I assume most customers who need extreme processing power have learned over the past 10 years that faster individual processors are not coming. Algorithm design plus parallel processors is going to be the source of perhaps all performance increases in the foreseeable future. Until we move away from silicon that is.
Are there even supercomputers out there which have faster processors than the fastest Xeon processors out there? I may be wrong, but I believe there really hasn't been any non-parallel based performance increases for a long time.
-- All that is necessary for the triumph of evil is that good men do nothing. -- Edmund Burke
MS are playing a very long game because they can afford to. Despite it's well-publicized problems, I find Windows 10 is fast and rock-solid on a desktop and on a Lumia phone. They already have Windows compiled for ARM and they have Office desktop apps compiled for ARM. OK it's a kludged version on the RT platform, but most of the work is done. They are making it easy and attractive (at least in a 'hell, why not?' sense) for new app development to compile for both x86 and ARM. I think one of the reasons why Windows 10 Mobile ('Phone') still exists is because it keeps the ARM branch current and that has sufficient value for MS that they don't even care if the phones never sell.
"Our opponent is an alien starship packed with atomic bombs," I said. "we have a protractor"
Contrary to popular belief, Moore's Law doesn't say that processors will double in speed every 18~24 months. It says that the number of transistors that can economically be put on a single chip will double every 18~24 months. Up until recently, that has translated into a doubling of speed for two reasons: 1) more transistors can be used to optimize the processing of instructions through a variety of techniques and 2) the distances signals have to travel is lessened as the transistors shrink. More transistors contribute not only to power consumption but also more heat, which is another problem with high performance processors. This was partially dealt with by putting multiple cores on a die running at less than max clock rates, thereby distributing the heat and making it easier to deal with. It still may be economical to put more and more transistors on a die, but maybe we don't want to. More transistors consume more power. What's your priority, raw speed or power consumption. Maybe you can't optimize for both at the same time.
I don't see how in the world *Windows* is going to break into the mobile market. They have been trying for over a decade, repeatedly without success. Particularly now it seems a pretty cemented Android/iOS landscape. The only hope I could see is Intel getting some hardware makers onboard and that being a platform for MS to push their continuum concept (yes it can work with ARM, but back to square one, a bunch of my enterprise applications are not about to spend money to dust off the build trees and build ARM for the fun of it)
MS mobile strategy is going to have to settle for trying to make money off of iOS and/or Android users/developers. They can (and do) provide hosting, applications, and services. They miss the revenue opportunity of a curated application distribution platform, but I think this is the best they can hope for.
XML is like violence. If it doesn't solve the problem, use more.
Except that Intel has been a licensor of ARM for a very long time, so even if there was some magical shift to ARM in non-mobile ultra-low-voltage devices, Intel would still be able to apply what they know about advancing the state of the art.
Don't worry about Intel, they'll be just fine.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
Exactly. Apple kept a secret x86 / x64 version of Mac OS X in the closet for 5 years as a hedge against IBM screwing them over on PowerPC. Turns out to be one of the best decisions that they ever made.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
Where I work we have lots of customers that *need* more processing power, and efficiency be damned.
I assume most customers who need extreme processing power have learned over the past 10 years that faster individual processors are not coming. Algorithm design plus parallel processors is going to be the source of perhaps all performance increases in the foreseeable future. Until we move away from silicon that is.
Are there even supercomputers out there which have faster processors than the fastest Xeon processors out there? I may be wrong, but I believe there really hasn't been any non-parallel based performance increases for a long time.
Yes there has, but more through architectural changes - new instructions, new modes, bigger better caches, improved offload models, task specific hardware (like crypto, packet moving etc.). This has been enabled by the increasing number of transistors on die and is driven by the mobile and server markets, which have evolving and quite different needs. Xeons today do more work per second than Xeons in the past and the scaling is greater than the scaling of the individual CPU core performance.
I should use this sig to advertise my book ISBN-13 : 978-1501515132.
I assume you mean licensee.
Check out my sci-fi/humor trilogy at PatriotsBooks.
> Going beyond 5Ghz limit has been a problem for the last decade or so.
Last decade? Uhm, try the last ~40 years. A close friend of mine worked with the military running GaAs CPUs at ~4.7 GHz in late 70's. He also worked on GaAs devices operating up to ~100 GHz. Hey, when you have a nearly unlimited tech budget you can do all sorts of things that the commercial sector won't have access to until decades later.
Anyways, the problem with Silicon is that it needs to be < 110 degrees C. In contradistinction GaAs only need < 175 degrees C.
Hardware designers have known about alternatives for years -- Silicon is just plentiful, dirt cheap, and "good enough." No one wants to pay $100,000 for a 10 GHz GaAs CPU, when you could buy 2,000x Silicon chips instead for the same amount of money.
No, a lot of applications don't scale well across multiple cores / CPUs.
In 2016 they don't. But as chips evolve the applications will as well.
That's what they said in 2006, when CPU clock speeds essentially hit the wall.
Mainstream CPUs started going multi-core back then. Some things parallelize quite well, and the tools are making it easier for them to do so today, but there's still a lot of sequential crunching to do for a lot of jobs. We're not likely to see a 1000 core 200MHz chip out-performing a 2 core 2GHz chip for "average desktop applications" anytime soon.
> consider, as a thought experiment, any task where the outcome of the first "step" determines the parameters for the next.
> There is no way to complete this overall task in parallel
In fact it's sometimes trivial. Consider this code, in which 'the outcome of the first step determines the parameters for the next':
HasPMI = IsMoreThan80()
PaymentAmount = CalculatePayment (Balance, HasPMI)
If you have 1024 cores, you can easily run CalculatePayment() in parallel with the line before it. You run it for both the true and false case simultaneously with IsMoreThan80. Then when the three threads complete, HasPMI tells you which of the two results to use.
That can also be EVERY IF STATEMENT, every switch-case, etc. On any branch, go ahead and precompute the value for the branch while deciding which branch you'll take. As things move in this direction, functional programming and similar disciplines start to become more valueable, so they will be used more.
A lot of things that wouldn't make sense to run parallel on two cores or four cores suddenly make sense of you have hundreds or thousands of cores laying around. With 4096 otherwise idle cores, it can make sense to calculate 1,000 possible scenarios in parallel and then ignore the 999 options you didn't need. Our way of thinking about problems will change, as will the tools we use to take advantage of the strengths of new systems.
Of course, the fundamental problem this presents is that it does *not* automatically result in improved performance.
Architectural changes require that performance code be tuned or re-tuned, which means every at-scale application has to be somewhere between rejiggered and given a huge dedicated rewrite effort (The DOE's upcoming 300 petaflop GPU machine will have exactly ten applications that can run at full scale, each of which will have an entire dedicated team rewriting it to do so). And, of course, Amdahl's Law puts an ironclad limit on the effect that more parallel hardware can have on performance, and some problems simply cannot be parallelized no matter how much we wish otherwise.
Contrast with the effect of improving the serial performance of hardware: All else being equal, double the CPU and memory clock rates and absolutely every program will run twice as fast, full stop. That was the desktop miracle from 1990 to 2003 or so - the same exact code screamed twice as fast every year.
But as processors trend towards slower and wider, everything becomes an exercise in parallel programming. OpenMP parallel, MPI parallel, SSE simd instructions, GPU simd parallel... It's harder to do at all, and harder yet to do *right*, and historically the average programmer has enough trouble working with a runtime that's sequentially consistent.
Rant aside though, I agree you're right - until we move to diamond substrates & heatsinks, we've hit the thermal brick wall (actually we hit it circa 2003) and there will not be any further increases in serial processing speed. Plus, AFAICT, there's a similar brick wall with access rates to DRAM and the fact that it requires a microwave-frequency bus with literally hundreds of pins extending for entire centimeters... so forget that too.
At this point I think they are looking for business models that are more annuity-like, with recurring revenue. A transactional purchase of an OS is becoming increasingly less interesting because fewer upgrade cycles. So for them, the strategy was 'app store or bust!'.
XML is like violence. If it doesn't solve the problem, use more.
You can have strong AI in ~20W, because that's what our brain uses. Each neuron is really, really slow like 100Hz and below, but when you have absurdly many it works. The problem is understanding the programming model, because it's nothing like our one list of instructions.
Live today, because you never know what tomorrow brings
People may have restated it in many silly ways, but what they actually mean is "Computers become twice as good every 18 months or so." Whether it's multiple cores, or faster clock speeds, or better RAM throughput, that's still what it amounts to: twice as good computers.
I think that's pretty much failed, then, for general purpose computers. At one time, I actually used to upgrade about every 18 months, and would see a really nice boost in performance. That's not so much the case anymore, it takes more like 3-4 years.
"Somebody has to do something. It's just incredibly pathetic it has to be us."
--- Jerry Garcia
I would have thought these days for 'Continuum' it's just a checkbox in their IDE to target a different processor architecture, with compiler warnings as to why this C code is non-portable.
[X] i686
[X] amd64
[X] arm v7
[X] arm v8
(I guess I should give visual studio a download instead of making uninformed comments!)
Wrong, completely 100% wrong and currently moderated to +5 Insightful.
Moore's Law has always been about performance. Originally there was a direct correlation between the number of transistors and speed, but that's changed and along with it so has the definition of "Moore's Law".
Moore's Law has always been about cost per transistor. While feature size means you get to fit more components per wafer density alone is not the only factor. Economies of scale, wafer size increases and accumulation of dead labor help to keep Moores law on track.
The basic idea is a feedback loop between cost per transistor vs affordability of features enabled by having more transistors. They cost less so everyone can afford to have more. This trend continues forever or until toasters end up with Internet connections .. whichever comes first.
I'd argue that it's also the case that most computers for the past decade have been ridiculously overpowered for what most average consumers are asking of them. That's partly why the market is moving to mobile. For many common tasks, a tiny mobile computer is still more than enough to do the job just fine. And in the case of Windows, the required minimum specs for an OS hasn't jumped nearly as substantially since Windows Vista, as MS focused quite a bit on performance optimization rather than letting things keep bloating up. If you had a reasonably powerful computer that could run Windows Vista when it first came out, you could almost certainly still run Windows 10 on it.
Vista recommended specs:
1-gigahertz (GHz) 32-bit (x86) processor or 1-GHz 64-bit (x64) processor
1 GB of system memory
40-GB hard disk that has 15 GB of free hard disk space
Windows Aero-capable graphics card w/ 128 MB of graphics memory (minimum)
Windows 10 minimum specs:
Processor: 1 gigahertz (GHz) or faster processor or SoC
RAM: 1 gigabyte (GB) for 32-bit or 2 GB for 64-bit
Hard disk space: 16 GB for 32-bit OS 20 GB for 64-bit OS
Graphics card: DirectX 9 or later with WDDM 1.0 driver
Note that I'm comparing recommended to minimum specs, but it's still fairly impressive given the time between these two OS releases. In general, I just think there's less market pressure to keep creating faster and faster CPUs.
Irony: Agile development has too much intertia to be abandoned now.
Well, I suppose you can call it air since 80% of air is, but using liquid nitrogen and calling it "air cooling" is a little bit misleading, don't you think? ;)