Intel Says It Will Move Away From 'Tick-Tock' Development Cycle

An anonymous reader writes: In its latest annual report, Intel says that it will be moving away from its decade-old "tick-tock" strategy (PDF) for developing new chips. From the company's 10-K filing, "We expect to lengthen the amount of time we will utilize our 14nm and our next generation 10nm process technologies, further optimizing our products and process technologies while meeting the yearly market cadence for product introductions." Anand Tech's Ian Cutress explains, "Intel's Tick-Tock strategy has been the bedrock of their microprocessor dominance of the last decade. Throughout the tenure, every other year Intel would upgrade their fabrication plants to be able to produce processors with a smaller feature set, improving die area, power consumption, and slight optimizations of the microarchitecture, and in the years between the upgrades would launch a new set of processors based on a wholly new (sometimes paradigm shifting) microarchitecture for large performance upgrades. However, due to the difficulty of implementing a 'tick', the ever decreasing process node size and complexity therein, as reported previously with 14nm and the introduction of Kaby Lake, Intel's latest filing would suggest that 10nm will follow a similar pattern as 14nm by introducing a third stage to the cadence."

Comments

R.I.P. Andy Grove

[Thud457]

My grandfather's clock was too large for the shelf,
So it stood ninety years on the floor;
It was taller by half than the old man himself,
Though it weighed not a pennyweight more.
It was bought on the morn of the day that he was born,
And was always his treasure and pride;
But it stopped short â" never to go again â"
When the old man died.

Ninety years without slumbering
(tick, tock, tick, tock),
His life's seconds numbering,
(tick, tock, tick, tock),
It stopped short â" never to go again â"
When the old man died.

Re:R.I.P. Andy Grove

[Penguinisto]

I was thinking the same thing...

(for those who haven't heard: http://arstechnica.com/informa... )

Maybe the board was waiting for him to be safely on The Other Side before doing this?

Re:R.I.P. Andy Grove

[thegarbz]

Maybe the board was waiting for him to be safely on The Other Side before doing this?

Then they got it a year too early. Articles were floating around already last year about this, and really why would they? Is it an offence to someone to change a strategy that has been in place for a very very long time because of a changing market, reaching limitations, and wasn't there a law of physics or 2 involved in there somewhere?

I would have liked to believe a man of Andy Grove's genius would have been the first to recognise that doing the same thing over and over again will eventually lead to problems.

AMD

[Anonymous Coward]

In other words, AMD finally catches up with Intel and ARM has a big up over Atom because of Intel's lost fab advantage. In good news for us, though, computer chips will get cheaper because it will finally make sense to build more fabs. If we are going to be stuck at 10nm for an indeterminate period of time, the process gets cheaper and it makes sense to build more foundries.

Arythmia model

[goombah99]

Then the flatline model

Re:Arythmia model

[blueshift_1]

Then I guess the government bail out would be the defibrillator model.

Re: Arythmia model

[goombah99]

Good one!

Moore's law, say hello to the law of economics

[JoeyRox]

At $5B+ for a single fab and the market for computers continuing its backward slide it's no surprise that Intel is putting the brakes on its capital expenditures.

Re:Moore's law, say hello to the law of economics

[Anonymous Coward]

I also wonder what they call paradigm shifting micro-architectures. Basically, apart from the PIV/Netbu(r)st series, all Intel processors are descendants of the Pentium-Pro, their first OOO processor. The changes have been in the area of completely different FPU units (good riddance for the x87 stack), switching to 64 bit (you have to thank AMD for that), and a few other improvements. But putting the memory controller or the GPU on the chip, using faster/wider I/O busses, or multiplying the number of cores is _not_ a paradigm shift, it's only using the transistor budget wisely because you can't design a core which needs 1 billion transistors (even if most of these are in the cache).
Bottom line, the core (no pun intended) of their current processors is closer to the 20 year old Pentium-Pro than to the Pentium IV or the original Pentium (which itself was largely a dual 486). Now I don't criticize the decision: if it isn't broke, don't fix it (Netbust was an ill-fated attempt). What I object to is calling it a paradigm shift when there have been only incremental improvements (in the micro-architecture) between PPro, PII, PIII, and the whole series of Core processors.

Re:Moore's law, say hello to the law of economics

[theendlessnow]

And there simply is no competition anymore. Tick-tock was designed to hammer, hammer and keep on hammering against AMD until they were dead, deAD, DEAD! (for those that don't know AMD used to compete against Intel)

Re:Moore's law, say hello to the law of economics

[known_coward_69]

what about all the new instructions they have added outside of x64. MMX, SSE and i forgot what else

Re:Moore's law, say hello to the law of economics

[castionsosa]

You hit the nail on the head. "Good enough" has knocked Moore's Law off the rails. Since there isn't that much demand, other than adding cores for virtualization [1], it isn't surprising that Intel is backing off the gas pedal with CPU development.

There are other things as well to add to a CPU. Disk I/O hasn't kept up with capacity gains, and there is always working on better power management which is something I'm sure Intel's enterprise customers are heavily damanding for PR reasons.

[1]: The ideal would be faster cores, since Microsoft has hopped on the Oracle and Sybase bandwagon and started licensing by core, and not CPU socket, but more cores is better than nothing.

Re:Moore's law, say hello to the law of economics

[avandesande]

The competition now are the chips that are already out there and for most people they work just fine.

Re:Moore's law, say hello to the law of economics

[Anonymous Coward]

New instructions don't change the fundamental flow of data inside the processor. In the PPro presentation, they said that the FPU needed 86 bit wide busses (80 bits data + status) and that this was " a lot of bits". Now they have AVX256 and 512 is right around the corner. Using parallelism to implement vector instructions is great for some tasks, but a compiler, for example or an interpreter (Python, Ruby, Perl) still executes mostly basic i386 instructions (or their 64 bit extensions).
Making the instruction set byzantinely complex does not help programmers and compiler writers, and the encoding is so complex that there two instruction caches on some processors: one in the native, inscrutable encoding, and one in another encoding easier on the decoders and schedulers. A saner, simpler, encoding (ARM and MIPS are quite good in this respect) would avoid this dual cache (which in the end means that the first level instruction cache is smaller because of transistor budgets).

Re:Moore's law, say hello to the law of economics

[Kjella]

Actually it's just Moore's law breaking down, the difficulty is in producing smaller transistors, the technology can't keep up. We know Intel had to delay the 14nm launch because of bad yields, now on 10nm it's probably a lot worse. And to go beyond that you need extreme ultraviolet lithography (EUV) which is still heavily in the R&D phase. I'm guessing that what Intel really knows at this point is that with a lot of tweaking they can probably do 10nm with acceptable yields using mostly known technology. What the world really looks like after three generations of 10nm? I don't think anybody knows. Intel once had a roadmap where they were tick-tocking all the way down to 5nm. This new roadmap is also just wishful thinking on where they'd like to be.

Re:Moore's law, say hello to the law of economics

[armanox]

It is my understanding that the PPro line ended with the Core 1, and the Core 2 was a redesign, and then Sandy Bridge was another redesign.

Re:Moore's law, say hello to the law of economics

[ChrisMaple]

In a given technology, going faster means more power dissipated, and there's a limit on the temperature that silicon semiconductors can tolerate. Power management is one way to gain a little on the speed-heat tradeoff. It's more than PR.

Re:Moore's law, say hello to the law of economics

[tlhIngan]

And there simply is no competition anymore. Tick-tock was designed to hammer, hammer and keep on hammering against AMD until they were dead, deAD, DEAD! (for those that don't know AMD used to compete against Intel)

AMD is still around, and I'm sure Intel is keeping them alive because they serve as "competition".

Should AMD disappear, Intel would be in a world of hurt from government regulators (the EU has found Intel to be in violation of anti-monopoly laws). So AMD right now is right where Intel wants them - strong enough to provide competition in the markets, weak enough to not really challenge them.

Re:Moore's law, say hello to the law of economics

[oh_my_080980980]

Actually it's an admission they can't do 10nm. Their current process will not support it. They'll need to follow IBM's lead in order to produce 10nm.

Re: Moore's law, say hello to the law of economics

[WarJolt]

I'm waiting for the single instruction that renders all my html5, runs all my Javascript and decodes all my video in 80 cycles. My poor old computer grinds to a halt when I open my 35 tabs.

Re:Moore's law, say hello to the law of economics

[theendlessnow]

This is true. I meant serious competition vs. competition for anti-trust sake.

Re:Moore's law, say hello to the law of economics

[trparky]

I'd have to disagree with the disk I/O part of your comment.

Disk I/O speed and bandwidth has been growing by leaps and bounds in the last three years due SSDs. SSDs have made huge improvements in computer performance lately, so much so that if you were to take even a four year old computer and put an SSD into it it would figuratively take off like a rocket. That just goes to show you, if you can't get the data and instructions to the CPU fast enough you're going to be staring a screen wondering why your application hasn't started yet. It's not the CPU and RAM that's the bottleneck in today's PCs, it's the data storage systems that we've all be using since the lowly hard drive showed up on the scene. Replace that hard drive with an SSD and suddenly everything is so much faster.

Digital computers are reaching the end

[110010001000]

This will make a LOT of people here mad, but the exponential growth computational power of digital computers is ending. We can no longer count of the computers of tomorrow to be significantly faster or have more memory than today. If you have been following the industry closely you can already see start to happen 10 years ago. So we can forget about projections that used the argument of exponential growth creating the "Singularity" or "AI". There just simply won't be enough processor power available with classical digital computers. The computer you use 10 years from now will look and perform a lot like the one you have today.

Re:Digital computers are reaching the end

[Anonymous Coward]

Your comment is bad and you should feel bad for making it

Re:Digital computers are reaching the end

[basscomm]

This will make a LOT of people here mad, but the exponential growth computational power of digital computers is ending. We can no longer count of the computers of tomorrow to be significantly faster or have more memory than today. If you have been following the industry closely you can already see start to happen 10 years ago. So we can forget about projections that used the argument of exponential growth creating the "Singularity" or "AI". There just simply won't be enough processor power available with classical digital computers. The computer you use 10 years from now will look and perform a lot like the one you have today.

Heck, the computer I use 10 years from now might very well be the same computer that I'm using today.

Re:Digital computers are reaching the end

[timholman]

This will make a LOT of people here mad, but the exponential growth computational power of digital computers is ending. We can no longer count of the computers of tomorrow to be significantly faster or have more memory than today. If you have been following the industry closely you can already see start to happen 10 years ago. So we can forget about projections that used the argument of exponential growth creating the "Singularity" or "AI". There just simply won't be enough processor power available with classical digital computers. The computer you use 10 years from now will look and perform a lot like the one you have today.

And the key phrase here is "classical digital computers", i.e. von Neumann architecture microprocessors, with conventional process shrinks. Yes, there's no doubt that the traditional computing path of the past few decades is coming to an end. You can't fight the laws of physics when you're dealing with devices and layers only a few atoms thick operating at gigahertz switching frequencies.

Now read about "IBM TrueNorth" and "DARPA SyNAPSE". A colleague of mine went to a demo of TrueNorth last week and said it was absolutely jaw-dropping. Neuromorphic computers operating asynchronously with power dissipations in the tens of milliwatts are going to change the processor landscape in the next decade. Next, look up "3D Xpoint". At that same conference, my colleague talked to someone working in Xpoint R&D who told him, "If you need a solid-state drive right now, buy the cheapest Samsung model you can get by with, because in the next two years we're going to blow the competition completely away."

Moore's Law (as in exponential computing and data storage growth) isn't going to come to a screeching halt. It's going to find different paths with different technologies, but it will definitely continue.

Re:Digital computers are reaching the end

[peragrin]

What you are forgetting is that to keep processor fabs paid for they keep shrinking everything else. While CPUs are 14nm most gpus are not. Ram is not.

In time expect to see ram, gpus, and the other components shrink as well. In 10 years you will buy a computer where all transistors inside it are at 14nm or less and it is Using a fraction of the power.

Re:Digital computers are reaching the end

[Maritz]

The cost per calculation per second capabilty of humanity has been following that exponential moore-like curve since at least the late 1800s. When 2-D silicon bounces off physical limitations then a new paradigm will likely keep the curve going. I guess we'll see.

Re: Digital computers are reaching the end

[armanox]

They already confirmed they were screwing us by making future processors lower power rather than faster. They're shoving their green agenda down our throats.

Oh yes, how dare they focus on things that effect users, like how long the battery in their laptop lasts, or their electric bills. The majority of users aren't seeing issues with CPU speed, so it is becoming less of a focus then other factors. Heaven forbid they focus on the consumer's needs.

Re:Digital computers are reaching the end

[Psiren]

At that same conference, my colleague talked to someone working in Xpoint R&D who told him, "If you need a solid-state drive right now, buy the cheapest Samsung model you can get by with, because in the next two years we're going to blow the competition completely away."

Oddly enough, I was at a conference last week where we had a keynote by HP. He was saying pretty much the same thing about memristors. He held up a roughtly credit card sized model that would apparently hold 1.5PB of data. It all sounds cool, but I'll believe it when I see it. These "just around the corner" technologies sometimes take a lot longer than expected to reach market.

Re:Digital computers are reaching the end

[petermgreen]

In my experiance ram is a big issue with older systems, especially laptops. I have a mid 2007 macbook (so about 8 and a half years old ) that mostly runs linux and I find myself often struggling for ram. I put 4GB of ram in it recently but linux only sees 3GB, presumablly due to some combination of firmware and chipset limitations. It's very easy to bring the machine to the point of grinding swap (running a 64-bit browser probablly doesn't help as it means the browser can keep chewing up memory rather than crashing and being restarted).

Re:Digital computers are reaching the end

[tlhIngan]

What you are forgetting is that to keep processor fabs paid for they keep shrinking everything else. While CPUs are 14nm most gpus are not. Ram is not.

In time expect to see ram, gpus, and the other components shrink as well. In 10 years you will buy a computer where all transistors inside it are at 14nm or less and it is Using a fraction of the power.

Memory almost certainly is at 14nm, if not smaller (they're usually a half-node ahead).

Memory is the most transistor-dense device you could make - of the billions of transistors on a chip, 90% of them would be memory related, and all together, they take up less area than the 10% "random logic" used in a CPU/GPU/chipset.

Modern VLSI logic has it that in random logic parts (i.e., everything that's not memory), you can stash in lots of additional unconnected transistors, so if there's an issue, they're there so you only need to remake a few masks instead of the entire set. (With each mask costing $100K, and a 10 metal process requiring easily 15-20 masks, that's $1.5-2M in startup costs.).

If you look at a revision label (aka stepping), they go like A0/A1/B0/etc. The number represents metal revisions - the first chip would be A0. They find bugs, fix them by re-laying out the metal traces using those spare transistors and increment the value to A1 )and this may only require producing 4-5 new masks). But then they find a critical bug that can't be fixed with the existing diffusion set (either requires too many transistors, or they missed some block they need, or there are too many revisions), so they remake the diffusion set of masks as well and bump it up to B0.

Anyhow - memory is very dense. Moore's law is basically targeting memory - and this is all memory - SRAMs, DRAMs, flash, etc. If it holds a bit, the arrays are made as dense as possible. Whole legions of logic went into the production - the NAND flash array is by design way more dense than a NOR flash array purely because of wiring. (This is why random logic is very not-dense - the wiring is what takes up all the space).

Memory arrays are very regular, and technologies like DRAM and NAND flash have optimized transistor layouts ensuring that the spacing between transistors is minimal by keeping the wiring required to a minimum by sharing the wires with adjacent cells so wiring is a minimum. (NAND cells share a common gate line - the page select, and the drain of one is connected the source of another so you have 32-128 transistors connected in series, making for a very tight 2D array. NOR flash requires every transistor have access to power, row select and the data lines, and routing the power lines means the density drops because transistors are now further apart).

In fact, random logic parts, because of their irregularity, use transistors that are larger than minimum size because there is space to spare, and the fanout is unknown, so you want larger transistors to drive potentially higher loads or large fanouts. A memory array uses smallest transistors because you want density, and the fanout and loads are predictable.

Re:Digital computers are reaching the end

[timholman]

Oddly enough, I was at a conference last week where we had a keynote by HP. He was saying pretty much the same thing about memristors. He held up a roughtly credit card sized model that would apparently hold 1.5PB of data. It all sounds cool, but I'll believe it when I see it. These "just around the corner" technologies sometimes take a lot longer than expected to reach market.

Even more oddly enough, I actually did some consulting with HP back in 2001 concerning their prototype memristor chips. They were having significant issues with 1/f noise making it difficult to read the difference between a "1" and "0" in the memory array.

They've been developing memristor technology more than 15 years, so hopefully they've finally licked the problems. But my impression is that Intel and Micron are a lot more confident about Xpoint, given the types of demos they're already doing with it. We shall see.

Re:Digital computers are reaching the end

[thegarbz]

We can no longer count of the computers of tomorrow to be significantly faster or have more memory than today.

Sure we can. We just can't count on them being in the same identical form factor. I fully expect the computer I use 10 years from now to have 4x the RAM, and 4x the number of processing cores. There's still plenty of space in my case for it. I don't expect my laptop to achieve the same thing.

Re:Digital computers are reaching the end

[edxwelch]

Sorry, but that's simply not true. Look at the case of Nvidia they had hoped for 16nm for Maxwell (their Kepler successor), but it simply wasn't ready on time. So, they redesigned it and made it more efficient and faster, and that was despite it being on the same 28nm process as Kepler.

Re:Digital computers are reaching the end

[K. S. Kyosuke]

We still have async CPUs.

And improved languages that will allow us to use CPU packages with larger numbers of smaller cores (this could actually work well with asynchronous designs, which seem to be easier for smaller units, see GreenArrays for example). And HSA-like architectures with specialized units for certain tasks, perhaps including FPGA-like computations in the same address space as CPU tasks (and definitely including GPU units in the same address space). Or even neural processing accelerators.

Re:Digital computers are reaching the end

[bored]

I heard the same thing nearly two years ago in the form "we are going to produce a 10TB 2.5" drive this year that will kill flash based devices".

I was skeptical then, because I could see someone making the device, I just couldn't see them making it for a price where all the flash vendors up and gave up.

As no one has actually seen a xpoint device, I suspect its still a couple years out for high end applications. Once intel/micron/etc milk that market for a couple years you might see one for your PC, maybe... Thats assuming they don't decide to be like 10Gbit ethernet and keep the price/markup outrageously high because they don't have anything to replace it with.. Then when a replacement comes out, and the price starts to fall no one in the consumer space is interested because they have moved on to other things (aka Wifi).

Re:Digital computers are reaching the end

[PCM2]

They've been developing memristor technology more than 15 years, so hopefully they've finally licked the problems.

Doesn't sound like it. Their world-changing God device known as The Machine is supposed to be based entirely on non-volatile memristor storage. But the first demo units are going to ship based on DRAM.

The HP mouthpiece's excuse? "DRAM essentially is non-volatile as long as the power doesn't go out."

Re: Digital computers are reaching the end

[Mal-2]

Oh yes, how dare they focus on things that effect users, like how long the battery in their laptop lasts, or their electric bills. The majority of users aren't seeing issues with CPU speed, so it is becoming less of a focus then other factors. Heaven forbid they focus on the consumer's needs.

If they could keep whipping the "faster, faster, faster" horse, they would. When the primary advantage of a new computer over an old one is that it takes less power and generates less heat, people don't see much pressure to upgrade. The old one still works just fine, even if the cost of operation is higher. This is not to say that pushing "smaller, cooler, quieter" is a bad thing for the world at large. It's obviously good. But it's not as good for Intel as pushing speed at all costs used to be. Therefore the conclusion has to be that they're doing it this way because as successful as the old way was for them, they can't make it work any longer.

Re: Digital computers are reaching the end

[armanox]

Absolutely - and I'm more then okay with pushing different aspects of advancement - much like when cars started selling on fuel efficiency without losing power. And, from a profit standpoint, Intel knows that they can't market on faster for too much longer, so they found a new exciting point that they can make big strides with. And it is a small point for some people, but there is a measurable difference in the power bill replacing my Quad G5, Core 2 Quads and Phenoms with i5s (I run BOINC all winter long)

Re: Digital computers are reaching the end

[toddestan]

If it uses some variant of the Intel 945 chipset, which was pretty common back then, then 3GB is the maximum amount of ram it will be able to use.

My almost 10 year old laptop is still mostly fine with the modern web. Though I've maxed it out with 3GB, SSD, and it sports a Core 2 Duo as opposed to a Core/Pentium M and it's got a Mobile Radeon for graphics. It's still more powerful than most tablets (not tablet PCs). It's usually not Javascript but shitty Flash sites that cause the most problem. Part of it could be graphics - I have a Mobility Radeon, which while dated, is much better than Intel's integrated graphics with the 945 which was terrible.

If you have the slightly newer 965 chipset, then you can (probably) use all 4GB.

Most people wouldn't need the power anyway

[jones_supa]

For an ordinary joe it won't matter that much. Most of the services he uses (Facebook, Twitter, Spotify, Netflix, Skype, lightweight gaming) could be implemented even on a Pentium II with a little bit of optimization. Even Microsoft does not bother artificially bloating their operating system anymore.

Re:Most people wouldn't need the power anyway

[Anonymous Coward]

That's the most stupid commentary ever. Software has become so bloated that it's hard to find any machine that can handle it gracefully, let alone in a 'snappy' way. The fact that software is rewritten for Browsers behind 10 additional layers of abstractions doesn't help either.

Re:Most people wouldn't need the power anyway

[K. S. Kyosuke]

It's funny that you first say it's "the most stupid commentary ever" and then you follow on with rephrasing his very complaint about accidental complexity of software in your own words.

Should have mentioned diminishing returns

[avandesande]

Differences in performance (speed, power consumption etc.) are now almost imperceptible between process changes.

Re:Should have mentioned diminishing returns

[avandesande]

What do you mean by puts out? They do that by shrinking the die and creating a new fab, which is precisely what the article is talking about. Back in the day you could see a double or triple in clock speed with each iteration. You really think that Intel is sitting on a 10ghz i8 for business reasons?

Single-thread performance is stuck

[PeeAitchPee]

It's pretty telling when the CPU single-thread desktop performance leader, the i7-4790K, is almost two years old. That used to be an eternity in silicon fab. Intel is busy on the server side cramming ever-more cores into their Xeons for high-density server rooms and reducing power consumption on the mobile side. The market (and Intel, who in part sets the market) has decided the Devil's Canyon is apparently fast enough for any single-threaded work you'll ever do. That doesn't help those of us who count on software whose vendors haven't yet implemented / optimized multi-core support in their apps.

Tick-Tock-Tock

[zhen.sydow]

Sooo... Tick-Tock-Tock then?

Re:Tick-Tock-Tock

[SeaFox]

That's how Intel will waltz into the next decade.

Not even close to fast enough

[Bruinwar]

My dream is to have a computer waiting for my input. Today, even with the fastest machine, I am contunually waiting on the machine. Blame it on crappy software, networks, whatever, but CPUs really need to be a LOT faster IMO.

Rockstar engineers

[jones_supa]

Who are these engineers who design the new, smaller manufacturing processes? I'm quite sure Intel or TSMC will reward you quite gratuitously if you are an engineer in a research team that makes 10nm feasible. Can you imagine, those guys change the world.

Re:Rockstar engineers

[oh_my_080980980]

IBM already has. Problem is Intel would have to switch fabrication process.

Moore's law is dead; physics killed it

[dlenmn]

I'm not picking on you in particular, but I'm seeing a lot of posts implying that Moore's law could keep going but it's too expensive, there's not enough competition to warrant it, etc. The fact is that physics is the nail in the coffin for Moore's law. Making small fab processes is getting more and more difficult because these size scales are super tiny, and the difficulty means that Moore's law simply cannot keep going because we have to develop fundamentally new technology -- not just scaled down current technology.

There's a reason Intel is planning to stop using Silicon at 7 nm (not clear what they'll move to -- maybe indium gallium arsenide), and getting up to production quality with a new material is a huge task that is fundamentally incompatible with Moore's law. (InGaAs is not "new" per se, but InGaAs has never seen real commercial use; it has been confined to research labs.)

There's also a reason that research in classical (not only quantum) computing with superconducting circuits is again being seriously researched by commercial enterprises -- including companies like Northrup Grumman which are not traditionally associated with designing computer chips. (IBM poured a lot of money into superconducting computers in the 1980s but ultimately gave up because Si computing was marching along just fine. I think that IBM is back in the superconducting game too.) Again, getting superconducting circuits up and running is _hard_ and fundamentally incompatible with Moore's law.

Moore's law is intrinsically dead. End of story. Even if/when the non-Si chips get up and running, I don't expect that Moore's law will be revived. 7 nm equates to about 14 silicon atoms. The end of the road is in sight. It's trying to march through quicksand from here on out.

PS. I don't get the "lack of competition" hypothesis for why Intel is slowing down; there are a number of manufacturers matching or closing in on Intel's fab process. E.g. Samsung and Globalfoundries are already at 14 nm. TSMC is at 16 nm. These aren't in direct competition with Intel at the moment, but they will be if Intel ever gets serious about putting their chips in things other than desktops/laptops/servers. Intel isn't stupid; they see these other companies as competitors, and Intel really wants a leg up on them. If Intel could keep up with Moore's law, they would.

Re:Moore's law is dead; physics killed it

[JoeyRox]

The premature death of Moore's law due to physics has been falsely predicted for about three generations of fabs. Those predictions were wrong then and they're wrong now. The profit motive and billions of dollars always found a way to find a solution to intractable technical problems - that profit motive is now disappearing (for Intel specifically) due to conditions in their markets.

Re:Moore's law is dead; physics killed it

[DrJimbo]

Your overall point may (or may not) be valid but this passage in particular is either incorrect or grossly misleading:

Making small fab processes is getting more and more difficult because these size scales are super tiny, and the difficulty means that Moore's law simply cannot keep going because we have to develop fundamentally new technology -- not just scaled down current technology.

We have had to develop new technology after new technology for decades to keep pace with Moore's Law. This is one of the things that makes Moore's Law so fascinating -- it has already spanned over five orders of magnitude (powers of ten). Take a look at the section on enabling factors and future trends on the Wikipedia page. It is possible we have finally reached the end of Moore's Law but to me it seems equally possible that we have not.

Re:Moore's law is dead; physics killed it

[Alomex]

Three? Try 15... I remember reading thirty years ago how fabs would fail once we reached dust size particles/visible light circuit size. For the first problem we got the humans outside the fab, for the second we moved to x-ray frequencies.

Re:Moore's law is dead; physics killed it

[MetricT]

Moore's Law isn't completely dead, it's just metamorphosing into a new form.

If Intel is moving into spintronics (as rumor suggests), then next-generation chips should use millions of times less power (like, run your CPU for a month on a AAA).

If so, it becomes possible to start stacking CPU layers like memory/flash is today. Imagine a next-generation Moore's law stating the number of transistors in a 3D stack doubling every X months.

Re:Moore's law is dead; physics killed it

[Ramze]

Intel is at 14nm and working hard to push out 10nm after many setbacks, but not long after that, they're done. Did you not read the whole explanation about why Intel is stopping at 7nm on silicon and/or 5nm with other materials? 7nm is 14 silicon atoms wide. Any smaller, and quantum tunneling becomes such a serious issue, they need new materials. With other materials, they MIGHT be able to go as small as 5nm. That's it, though. Any smaller and you basically either need an optical computer or a quantum computer. Even with those, you'll be limited to the size of an atom per transistor at best which is 0.2nm for silicon.

Moore's Law is dead.

Re:new cycle

[ChrisMaple]

Hickory Dickory Dock.

Moving to a three stage process?

[Chas]

I vote that we call it "Boom Shaka Laka"!

Intel not really ahead on 14nm

[edxwelch]

The article says: " Intel introduced 14nm back in August 2014, and has since released parts upwards of 400mm2, whereas Samsung 14nm / TSMC 16nm had to wait until the launch of the iPhone to see 100mm2 parts on the shelves"
This is not really a fair statement, as Intel's 14nm process began with very poor yields, while TSMC began from the startoff with very good yields. It was only mid - 2015 that Intel fixed their yield problems.

Re:Actually there is competition.

[spire3661]

Idiots keep saying this and we keep telling them they are idiots but you just wont listen. ARM cannot replace x86, will never happen.

Translation

[tom229]

"We're going to artificially slow our release cycle to squeeze as much money out of the consumer as possible."

Of course they've already been doing this all along. As we rapidly approach the size of a molecule the new frontier will just be power consumption.

ARM only replacing x86 for servers and consumers

[raymorris]

Yeah, ARM can never replace x86. There will never be a day when consumers mostly buy ARM devices rather than x86. Well not until 2014 anyway. The fact is, most CPUs purchased in the last two years were ARM.

In the datacenter, power (and it's associated cooling) is expensive, so we're already starting to see ARM replacing x86 in the datacenter too.

  Business desktops still mostly run x86, because they mostly run Windows and Windows is currently x86-centric. Microsoft has already released an ARM version of Windows, though, and they are currently making a big push toward "apps" that cpu-architecture independent. That is to say, any application written according to Microsoft's recommendations will be ARM compatible.

ARM hasn't completely replaced x86, but ARM does now have most of the market, and the most significant hurdle"for ARM, Windows compatibility, appears to be going away.

It's like hard drive density

[DidgetMaster]

Remember the good old days when hard drive capacities doubled or tripled every single year? There was a time when if a 20 GB drive was the biggest thing this year; you could expect 60 GB or 80 GB drives next year. Those days are over. We see higher capacities, but they take a lot longer between cycles. The same thing is happening with CPUs. Once the dies got below about 50 nm, it became increasingly hard to keep shrinking it further. I'm not saying that 1 nm is impossible, but it's going to be very difficult to get there. It's also kind of like clock speeds. Once they reached the GHz range, it started hitting physical barriers (e.g power and heat) and stopped doubling every generation.

Re:Actually there is competition.

[K. S. Kyosuke]

ARM cannot replace x86, will never happen.

And Intel couldn't replace S/360 and POWER and MIPS and SPARC, never happened.

Tick/Tock/Wow, new capacitors repost

[justthinkit]

Moore's rule of thumb expired two years ago.

It can't continue forever. The nature of exponentials is that you push them out and eventually disaster happens.
- Gordon Moore, in 2005

recently, we've just scaled down existing methods

[dlenmn]

Back in the day, moving from bipolar to MOSFET transistors was a fundamentally new technology, but we haven't done anything like that any time recently. Almost all of the examples on that list are old or speculative. All the chips in recent memory have been silicon MOSFETS made using ultraviolet photolithography. Moving from planar transistors to FinFETs is the closet thing to a new technology, but that really seems like a refinement. Moreover, banking on a fundamentally new technology won't save Moore's law because the technology needs to be ready now, and it's not.

We have reached the end of Moore's law. That isn't up for debate. Computer performance has been slipping from Moore's schedule for a while now. The question is why. I maintain that it's due to physical limitations. The OP says it's due to economics. I guess that you could argue that we'll get back on pace to follow Moore's law at some point in the future, but that's a much harder argument to make.

Re:Actually there is competition.

[LWATCDR]

A better example would be Intel can never replace PDP-11s, VAXs, Data General Novas, Wang 2200, or Prime 50s.
IBM still is making mainframes.

The predictions have been right all along!

[dlenmn]

Guess what: Moore's law has been failing for several generations of fabs! The divergence from Moore's law has been gradual. No one is saying that progress will suddenly stop, but we've been slowly falling behind the "doubling every two years" schedule for a while now (arguably since at least 2012).

Now, you can argue about why that is. However, the problem is not a lack of effort or funding. I have a bunch of friends who work at Intel, and they're not taking it easy. They're working their asses off but making progress slower than they used to. Making transistors smaller and smaller is proving to be a very difficult task. Now, if you want to call that an engineering problem rather than a physics problem, go right ahead. But the fact remains: the difficulties stem from the physical size of the transistors, not from managerial issues.

Re:recently, we've just scaled down existing metho

[DrJimbo]

I fully agree with you that if we are at the end of Moore's Law then it is because of physical limitations and not economics. As for no preceding tech breakthroughs, Intel's first CTO said (in 2008):

I compare Moore's Law to driving down the road on a foggy night, how far can you see? Does the road stop after 100 metres? How far can you go?

[...] That's what it's been like with Moore's Law. We thought there were physical limits and [now] we casually speak about going to 10 nanometres. We have work going on different transistor structures. Silicon has become scaffolding for the rest of the periodic table. We're putting these other structures into the materials. We see no end in sight and we've had 10 years of visibility for the last 30 years.

I think it is quite possible he is wrong about Moore's Law extending out to 2028 but I find it very hard to believe he is wrong about the history of Moore's Law leading up to 2008. He was in a position to see the tech breakthroughs first-hand. I don't see why he would lie about it.

Re:Called the Tick-Tock-Cock model elsewhere.

[Anonymous Coward]

Oh, for pete's sake. The CEO's name is spelt Cook, Tim Cook.

Must you haters turn every doggamned story into a homophobic rant about how evil Apple is?

Re:Actually there is competition.

[ChunderDownunder]

A rumour this week is that Apple were haggling over chip maker Imagination.

So it's more a case of MIPS replacing ARM.

Re:Actually there is competition.

[K. S. Kyosuke]

Well, a lot of those users who used to use S/360s are using Wintel boxes today. Universities, for example. And a lot of companies. Especially those that would have been using the lower end mainframe models back then.

Re:Actually there is competition.

[K. S. Kyosuke]

users who used to use S/360s are using

Please excuse me while I'm going to shoot myself.

Bedrock?

[Tough Love]

Oh, tick-tock was the bedrock of Intel's success? Silly me, I thought it was more about monopoly control and cutting off AMD's air supply.

Re:new cycle

[Anonymous Coward]

Your mom was sucking my cock.

Re:Actually there is competition.

[PCM2]

Nah, Apple wants 'em for their mobile graphics chips, mainly. You might see some MIPS in some gadget down the road if the deal does go down, but I can't see Apple throwing away all the investment it's already made in its A* line of ARM chips.

Re:Actually there is competition.

[ChunderDownunder]

true, PowerVR is the main reason.

But I wouldn't expect them to chuck the MIPS architecture in the bin entirely. You'd think they'd at least keep the IP alive and make an effort to port an internal build of iOS/OS X to the architecture for fun. Understandably, productization would be some way away if at all.

Re:Actually there is competition.

[LWATCDR]

They are but we still have mainframes but minicomputers with the exception of the IBM System/38,AS400,iSeries have all but become extinct.

AMD

[DarthVain]

Intel and "Tick-Tock" basically ground AMD into dust. With AMD unable to keep up R&D development, they are no longer really competitive in many of the CPU segments. Meaning that Intel doesn't need to bother anymore (or at least for awhile), as they are really only competing against themselves. Not only are most of Intel CPU offerings "good enough" they are also "better than anything else" so why bother...

water and phase change cooling finally make sense

[0111 1110]

As CPUs slow to a crawl and soon come screeching to a...pause I'm thinking that water cooling and phase change cooling is going to get a boost. People can finally justify spending money--real money on sophisticated cooling systems. I already have a high end water cooling setup that I haven't used for years, but I've never seriously considered making the jump to phase change..until now.

a new golden age of assembly language programming

[0111 1110]

Time for programmers to once again enthusiastically embrace assembly language. The age of depending on ever faster hardware as an excuse for fast/lazy/elegant programming is about to end.