Intel Broadwell-E, Apollo Lake, and Kaby Lake Details Emerge In Leaked Roadmap

bigwophh writes: In Q4 2016, Intel will release a follow up to its Skylake processors named Kaby Lake, which will mark yet another 14nm release that's a bit odd, for a couple of reasons. The big one is the fact that this chip may not have appeared had Intel's schedule kept on track. Originally, Cannonlake was set to succeed Skylake, but Cannonlake will instead launch in 2017. That makes Kaby Lake neither a tick nor tock in Intel's release cadence. When released, Kaby Lake will add native USB 3.1 and HDCP 2.2 support. It's uncertain whether these chips will fit into current Z170-based motherboards, but considering the fact that there's also a brand-new chipset on the way, we're not too confident of it. However, the so-called Intel 200 series chipsets will be backwards-compatible with Skylake. It also appears that Intel will be releasing Apollo Lake as early as the late spring, which will replace Braswell, the lowest-powered chips Intel's lineup destined for smartphones.

Comments

native USB 3.1 is not that big of a thing

[Joe_Dragon]

native USB 3.1 is not that big of a thing as on most board be it native or add on chip it's still over the same DMI bus.

Now intel needs to add more cpi-e to the cpu. At least 20 lanes + DMI. 16 for video and 4 for other stuff like TB 3.0 PCI-e SSD's.

Re:native USB 3.1 is not that big of a thing

[viperidaenz]

Since USB 3.1 Gen 2 is faster than a PCI Express 3.0 lane, perhaps it's better to implement it closer to the CPU and memory controller?

Re:native USB 3.1 is not that big of a thing

[Anonymous Coward]

Now intel needs to add more cpi-e to the cpu. At least 20 lanes + DMI

Already existed on sandy and ivy bridge, if the IGP was disabled they had an extra PCIe 3.0 x4. Used on a whole bunch of C200-series server boards (e.g. Supermicro X9SCL/SCM).
Haswell lost that ability, guess it ate too much into the low end E5 xeons...

Re:native USB 3.1 is not that big of a thing

[blankinthefill]

One of the issues that I've been running into for a long while, and expect to be running into even more with the expansion of the M.2 and related slots, has been the serious lack of PCI-E lanes that Intel supports. It's very easy, running SLI and one of two other things that use PCI-E, to run out of PCI-E lanes on today's boards, especially if you're a power user. And with new expansion slots for SSDs and other applications starting to enter the market, using multiple PCI-E lanes (up to 4 for a single M.2 slot), it's going to be even easier to suck all those lanes up and still need more. Honestly, for some power users, Intel could probably double the number of PCI-E lanes natively supported, and still not provide enough.

Re:native USB 3.1 is not that big of a thing

[blankinthefill]

*One or two. Maybe I should start using that preview instead of ignoring it and going straight to posting...

Re:native USB 3.1 is not that big of a thing

[AHuxley]

+1 Re "t's going to be even easier to suck all those lanes up and still need more"
The news about lanes, gpu, M.2 needs and todays lack of lanes is getting interesting. Once a user starts to add up the lane options and the ability to run the M.2, gpu, USB as expected it becomes an issue at the consumer, entry level.
Lets hope the lane count is much better and actually not an issue next gen.

Re:native USB 3.1 is not that big of a thing

[viperidaenz]

There's quite a few PCI-E lanes.

16 directly from the CPU
20 from the chipset via the DMI link (in the Z170, it was 8 2.0 lanes prior). The new chipset for these new CPU's ups that to 24 lanes.
That's a total of 40 PCI-E 3.0 lanes.

Re:native USB 3.1 is not that big of a thing

[WarJolt]

There are a few markets where fitting extra USB chips on a board is actually a big deal.
Power consumption might also be improved.
Also an external chip adds cost and many times pennies matter.

Re:native USB 3.1 is not that big of a thing

[sexconker]

2 video cards will take 32 of them, a high end SSD will take up 4, if you've got a wireless card, a sound card, or some other shit you're eating a couple more. And then you've got all the legacy SATA ports and whatnot that may eat up some of those lanes opportunistically.

40 is by no means future-proof. I'd like to see 48 or 64 for a pro/enthusiast rig,

Re:native USB 3.1 is not that big of a thing

[Joe_Dragon]

The DMI link from the CPU is only pci-e X4

Re:native USB 3.1 is not that big of a thing

[thegarbz]

The GP does have a point though. The number of PCI-E lanes is being actively addressed by Intel already. 40 may not be much but it's a step up from the status-quo.

Re:native USB 3.1 is not that big of a thing

[KingMotley]

I may be confused, but the older Haswell line had 40 PCI lines directly off the CPU and either 4 or 8 off the chipset. The newer architecture drops the CPU to having only 16 lanes directly off of it, and the chipset now has up to 24. 40+8 in the old, and 16+24 in the new is a downgrade, right?

With most motherboards having a SATA controller, USB 3.1 controller, network card (where are the 10GB network ports???), and sound. Then drop in a couple video cards (32 lanes), a M.2 SSD (4 lanes) and you are either out of lanes or real close. I also need 4 lanes for my RAID controller so I'll likely need to find a motherboard that has a PCI-E switch because the CPU/chipset have run out.

Re:native USB 3.1 is not that big of a thing

[Agripa]

With an external MAC and PHY, there is no reason for a USB 3.1 Gen 2 controller to use just one PCI Express 3.0 lane. PCIe is convenient that way.

Upcoming names

[Anonymous Coward]

when does Intel "Cornf Lake" come along?

3+GHz speeds, extra cores, more lanes.

[AHuxley]

6~10 cores, some with 15~25MB of L3 cache.
A generation of experts will have to work to ensure computer math, science and games can often be spread over the many cores.

Re:3+GHz speeds, extra cores, more lanes.

[Anonymous Coward]

what's the L1 and L2 size?

Re:3+GHz speeds, extra cores, more lanes.

[JoeMerchant]

Do problems really have to scale up to consume the available compute power?

Big CPU suckers like Monte Carlo and HiDef video processing are near trivial to parallelize, while most "normal" compute tasks are sub-millisecond on a single 2GHz thread, especially with FPU and other specialized instructions.

Granted, as camera prices fall, I want to have real-time intelligent video processing on an array of 20 cameras, but, can you spot the parallel opportunity there?

Re:3+GHz speeds, extra cores, more lanes.

[gweihir]

That is unlikely to happen. Parallelizing most things is orders of magnitude more complex than writing them single-task, and for quite a few things it is either impossible or gives poor results.

Re:3+GHz speeds, extra cores, more lanes.

[serviscope_minor]

Big CPU suckers like Monte Carlo ... are near trivial to parallelize

MCMC isn't. The first MC part of MCMC means each calculation depends on the previous one, more or less the definition of not parallelisable. Of course, you can run several in parallel which is fine, but they still have to burn in. If the burn in is a significant part of the time it takes for the computation, then parallelisation doesn't buy you all that much. I've seen problems when the burn is amazingly the main cost.

If it's hard to estimate an approximation of the state, and bad states are quite sticky then it can happen. Problems where you have to change the dimensionality as part of the MCMC process seem to me emperically prone to long burn in times.

Re: 3+GHz speeds, extra cores, more lanes.

[Anonymous Coward]

Why is there a myth perpetuated that parallelising tasks is hard? Ever made coffee and toast at the same time? Then congrats, your brain can grasp it.

Re:3+GHz speeds, extra cores, more lanes.

[Anonymous Coward]

I find parallelizing most things easy, except latency sensitive workloads, like small audio buffers. But if all you care about is throughput, parallelizing is simple. My first non-homework program was a production program and was threaded, scaled near linearly, and I taught myself multi-threading in 3 days. Watching a 32 core server hit 100% load was fun. That program went into production and hasn't been touched in almost 10 years now. There was one bug found, and it was a race condition in the logging. I easily found it and fixed it, without even using the debugger. A bit of whiteboard and thinking is all I needed to fix a race condition in a program that I hadn't touched in almost a year.

When you program, make sure you have a single point of responsibility and your logic does not result in inconsistent states. Suddenly threading is easy. Threading is hard when you have spaghetti code.

Re:3+GHz speeds, extra cores, more lanes.

[JoeMerchant]

Most MonteCarlos I've seen do benefit from multiple runs to improve accuracy - not to insult a very important area of computational methods, but the whole idea of MC simulation seems an extravagant use of compute resources just to get a statistical prediction of an unknown quantity. In nuclear medicine, ok, fine, you are actually simulating physical particles that have reliable statistically modeled behaviors, but Blackman-Scholes pricing? That's sociology, and I have a hard time believing that the market can't flip over and start obeying a completely different (and unknown) set of models if the players change, whether by world events (war, natural disaster, religious upheaval), or the personal life circumstances of large players (Buffet, Gates, Ellison, Koch and Walton attend a party and all get dosed with LSD...)

Re: 3+GHz speeds, extra cores, more lanes.

[gweihir]

Why do idiots like you feel they can blatantly ignore some 30 years of CS research into the subject?

Re:3+GHz speeds, extra cores, more lanes.

[gweihir]

There are _very_ few workloads that scale close to linear. Most workloads scale massively worse or not at all. Have a look at the relevant research some time. Parallel algorithm and multi-CPU systems are not a new thing. At all.

Re: 3+GHz speeds, extra cores, more lanes.

[lucien86]

Parallelising tasks is not hard? Here's a test I came up with to test human parallelism. - Try reading four books at once.. I mean lay them out in front of you and read all four simultaneously.. Its impossible, humans basically cant parallel process. What we can do semi-well is multi-task, which is the brain switching between one task and another - and even then if the tasks require intellectual effort we tend to get muddled and do things wrong..

Nor is HDCP 2.2

[Anonymous Coward]

More anti-consumer crap, that we pay for,
I'll pass, thanks.

Re:Nor is HDCP 2.2

[Anonymous Coward]

What are you talking about? Everything Joe_Dragon described is targeted toward speeding consumer performance - not s single server or enterprise feature in that list.

Re:Nor is HDCP 2.2

[Anonymous Coward]

What are you talking about?

He's talking about HDCP 2.2 which is in the title of his reply. It's anti-consumer crap, that we pay for, in more ways than one.

Re:Nor is HDCP 2.2

[Anonymous Coward]

HDCP "anti consumer?" You do realize that consumers WANT to play protected media, right?

This kind of reality-disconnect is exactly why Linux has 1% of the consumer desktop. Failure to understand what people want to do with their machines.

Re:Nor is HDCP 2.2

[mrchaotica]

No. People want to play media. They have no desire whatsoever to have it "protected" against them.

Re:Nor is HDCP 2.2

[Anonymous Coward]

No, they want to play media. The fact that it's protected is part of the problem.

Pretty soon people will be bragging that their shackles come in blue instead of black.

Re:Nor is HDCP 2.2

[Gadget_Guy]

No. People want to play media. They have no desire whatsoever to have it "protected" against them.

People also would rather not pay for their media, so if they have to choose between protected content and no content at all (because the content providers think that it is not economically viable enough for them to release it DRM-free) then the consumer will choose the former option. And if the protection is implemented well so that it doesn't adversely affect the consumer then they probably wouldn't give a damn.

Re:Nor is HDCP 2.2

[Kjella]

No. People want to play media. They have no desire whatsoever to have it "protected" against them.

People also would rather not pay for their media, so if they have to choose between protected content and no content at all (because the content providers think that it is not economically viable enough for them to release it DRM-free) then the consumer will choose the former option. And if the protection is implemented well so that it doesn't adversely affect the consumer then they probably wouldn't give a damn.

I think you confused "not economically viable" with "profit maximizing". You think that famous artists, movie stars and authors that make tens of millions of dollars would say "Nah, I'd rather go work at McDonald's" if you cut their wage in half? And I'm sure you noticed how the music industry imploded after iTunes gave up the DRM. Oh wait, it didn't. And there's a whole lot of countries I'd live in if North Korea was the other option, we don't have to allow unreasonable terms if we don't want to. Just because it would be economically profitable to weld shut the hood of the car and control how you drive it after you've sold it, doesn't make it right. The doomsday scenarios are false. We could easily drop the DRM-protection, ban DRM and go back to plain old copyright infringement without the world coming to an end.

Re:Nor is HDCP 2.2

[Gadget_Guy]

I think you confused "not economically viable" with "profit maximizing".

I'm not confusing any terms, because it is not my decision to make. It is the publishers who make that decision.

And I'm sure you noticed how the music industry imploded after iTunes gave up the DRM. Oh wait, it didn't.

I also noticed that for the majority of people, the removal of DRM made little to no difference at all. That is because they made the protection as unobtrusive as possible. Yes, the protection did prevent you from moving your digital files around, but it didn't stop playing on the Apple devices or burning the tracks to an audio CD (up to 7 times).

We could easily drop the DRM-protection, ban DRM and go back to plain old copyright infringement without the world coming to an end.

We are not talking about the world coming to an end, we are talking about whether consumers are willing to accept DRM-encumbered media. Many people here may be opposed to using any protected video and music, and I'm one of those people (I still buy my music on CD & rip them, and I still hate iTunes because it prevents you from copying non-iTunes-store music from an iPhone if you don't have access to the original computer that copied them to the phone).

But we are in the minority. The majority of people in the world either don't notice DRM or they are accepting of it.

We could easily drop the DRM-protection, ban DRM and go back to plain old copyright infringement without the world coming to an end.

And DRM could stay as it is and the world won't come to an end.

Re:Nor is HDCP 2.2

[ickleberry]

No Intel processor release is complete these days without the latest installment of media industry-approved DRM shite

Re:Nor is HDCP 2.2

[Kjella]

I'm not confusing any terms, because it is not my decision to make. It is the publishers who make that decision.

Between money or more money. Not making money or losing money.

I also noticed that for the majority of people, the removal of DRM made little to no difference at all. That is because they made the protection as unobtrusive as possible. Yes, the protection did prevent you from moving your digital files around, but it didn't stop playing on the Apple devices or burning the tracks to an audio CD (up to 7 times).

But it made it impossible (or at least extremely inconvenient) to move away from an Apple device. The market effects were obvious and was a huge part of the iPod's success and cost the consumers millions through lack of competition. The consumer might not have really understood, but they knew it worked on Apple and didn't work anywhere else.

But we are in the minority. The majority of people in the world either don't notice DRM or they are accepting of it.

They don't notice it because what millions and millions of people download have DRM removed.

And DRM could stay as it is and the world won't come to an end.

True, but you were the one claiming that publishers wouldn't publish without DRM.

Re:Nor is HDCP 2.2

[serviscope_minor]

People also would rather not pay for their media, so if they have to choose between protected content and no content at all (because the content providers think that it is not economically viable enough for them to release it DRM-free) then the consumer will choose the former option.

Ecept that's crap. Just about everything is freely available on the Pirate Bay. Everything is released on DVD still which while technically hsa DRM, it's so thoroughly hacked that it may as well not have.

And guess what? People still pay.

Likewise, Amazon music store has DRM free, perfectly standard MP3s. (I gather Apple is DRM free, too, don't know if they're MP3 or AAC though). And people still go there. there are several resons:

1. Most people are actually happy to pay a fair price for a fair product in a timely manner.
2. The price is fair.
3. The product is fair.
4. Stuff is up in a timely manner.
5. The stores are good (everything in one place, good UI, easily searchable).

And if the protection is implemented well so that it doesn't adversely affect the consumer then they probably wouldn't give a damn.

Sooner or later it always affects someone in an adverse manner. Every DVD or BluRay with unskippable bullshit (and there are plenty of those) is affecting the consumer in an adverse manner.

This is what DRM gets you:

http://farm5.static.flickr.com...

Personally I don't know anyone who pirates music any more. It's just not worth it to fuck aronud when you can spend a quid and get the track you want. Also, you don't have that vaguely unpleasant feeling of screwing someone over.

Many of those same people use TPB. I bet if the video stores offered a better product, they'd get more business. In almost all cases TPB is not only free, but better in every regard.

I ended up torrenting stuff available on 4OD and iPlayer (note, both already free for me), because it was just flat out easier and the software I can use is much better.

Re:Nor is HDCP 2.2

[thegarbz]

I also noticed that for the majority of people, the removal of DRM made little to no difference at all. That is because they made the protection as unobtrusive as possible.

No for the majority of people it made no difference at all because at the first error they just jumped online and pirated the content. Serious, I see luddites every month who can't figure out why they can't move their e-book, or song onto their media player, or don't understand why their "ultraviolet bundled download" doesn't seem to work on their device simply fire up their malware infested copy of uTorrent and obtain another copy of something they've already paid for.

We are not talking about the world coming to an end, we are talking about whether consumers are willing to accept DRM-encumbered media.

By-n-large they're not. You just think they are because the majority of consumers do the "approved thing" with their copy. The second they are in any way inconvenienced they pull out the pitchforks. (See shutdown of major services, games that log out when internet connections drop, and the political discussions that every government around the world went through en mass as they tried to figure out if they were legally allowed to copy their store bought CD to their iPods).

And DRM could stay as it is and the world won't come to an end.

If DRM stayed as is, no one would complain. But it's not. It's an endless game where publishers try and push a little further to erode the rights of the consumer and extract more money for something they've already sold to us. (See the number of times over the past few years where publishers have been taken to court, released updates that weakened or disabled DRM, and apologised to purchasers. It's a non-trivial amount)

Re:Nor is HDCP 2.2

[Applehu Akbar]

"No for the majority of people it made no difference at all because at the first error they just jumped online and pirated the content. Serious, I see luddites every month who can't figure out why they can't move their e-book, or song onto their media player"

Luddite is not the right term for these people. They are simply consumers who expect a consistent interface and know that there is no technological problem with implementing the operations they seek. They just resent that lawyers won't let them do it.

Re:Nor is HDCP 2.2

[Anonymous Coward]

" Just because it would be economically profitable to weld shut the hood of the car and control how you drive it after you've sold it, doesn't make it right."

Google would like to have some words with you.

Re: Nor is HDCP 2.2

[hackwrench]

That makes it sound like all existing media would stop existing. If all existing media would be freely available, I wouldn't have a problem with major players closing up shop.

Re:Nor is HDCP 2.2

[hairyfeet]

Except if you look at the numbers the artists ain't getting shit and in fact are making less of a percentage now than they did in the 50s.

This is why I have zero fucks to give about "the industry", because I know that Meatloaf had to file bankruptcy because Bat Out Of Hell I, which just FYI holds the fricking record for the longest run on the top 200 in history BTW, according to the record company didn't make a cent therefor they didn't owe him a penny. I know that Cheap Trick is currently suing their former record company because you know all those iTunes sales? Yeah the record company says that "since that didn't exist at the time you were with us in the 70s and 80s we don't owe you a cent from those sales" and I know that Keith Richards says all those iconic Stones albums from the 60s? Yeah they haven't seen a single dime from those since 75. Want something newer? How about Lyle Lovett sells millions, earns nothing.

So if you want to "steal" from the industry? Go right ahead, at least when it comes to music you are NOT hurting the artist as they ain't getting a thin dime from those songs.

Re:Nor is HDCP 2.2

[Gadget_Guy]

Between money or more money. Not making money or losing money.

Irrelevant to this discussion.

But it made it impossible (or at least extremely inconvenient) to move away from an Apple device. The market effects were obvious and was a huge part of the iPod's success and cost the consumers millions through lack of competition. The consumer might not have really understood, but they knew it worked on Apple and didn't work anywhere else.

That is correct, but as you say the consumer didn't understand and in most cases didn't care because they simply didn't ever try to move away from the Appleverse.

They don't notice it because what millions and millions of people download have DRM removed.

No, they didn't notice the removal of DRM because they dutifully installed iTunes and never tried anything that would trigger the rights management. People were far more likely want to write their music to CD format than copy it to a non-Apple brand of player and that was still supported.

True, but you were the one claiming that publishers wouldn't publish without DRM.

No, I never claimed that. It only requires one publisher to decide not to release one song/movie in a DRM-free digital format for my statement to be true. My point has always been about the public's willingness to accept DRM that isn't onerous. If they have to connect to the Internet for the sole purpose of playing a local file then they will get annoyed. But if they package the DRM in a way that seems like a benefit to the consumer (or if it is effectively invisible like the iTunes DRM) then the consumer doesn't care. No amounts of arguments about the pros and cons (however correct they are and however much I personally agree with them) will change the my point about consumer attitudes. If the consumer can't play their media because of onerous DRM them they will be pissed off. But if we pulled all support for DRM technologies and their media files stopped working then that is also an anti-consumer practice and will piss off the very consumers that you seek to protect.

Re:Nor is HDCP 2.2

[Gadget_Guy]

Ecept that's crap. Just about everything is freely available on the Pirate Bay. Everything is released on DVD still which while technically hsa DRM, it's so thoroughly hacked that it may as well not have.

The people who download stuff from the pirate bay are not consumers, they are pirates. The argument about DRM does not apply to them because they don't ever use DRM-encumbered media. DRM is not designed to stop the people who identify themselves as pirates; it is to prevent those people who would balk at being called that (and wouldn't dream of loading torrent software) but who see no problems with copying an album or a movie to give to a friend. Morals are not absolute. While there is no difference in downloading something from the Pirate Bay and being handed a CD from a friend to rip onto your computer, there are people who see this as two different moral levels. The former is seen as stealing, while the latter is legitimate because it's just their friend's CD.

My point has always been that people will accept DRM if it is not onerous (that is, as long as they don't have to do anything for their media to play like connecting to the Internet when there is no other reason for them to have to do this). Equally, the general public will not copy something to give to a friend if it seems onerous or if there is the vague idea that somehow there duplication might be detectable. I think that the old days of handing around physical media made casual copying much more acceptable because it felt like you were lending something physical to someone even though they got to keep the contents of that physical media.

With digital files it's a bit different, especially if you have client software like iTunes to manage it. I asked my mother-in-law (who regularly buys music from the iTunes store) whether she thought she could give a copy of an album to someone and she had no idea. The concept of ownership is blurred when the music is attached to accounts and accessed with special software. It didn't occur to her that the music had a physical form (as a file) so she has never once thought about being able to duplicate any music, despite her and my wife regularly sharing CDs with each other.

Like a lot of people, she doesn't care if her music and TV shows have DRM on them; ll she wants is to be able to play them. And that was the entire point of my original post.

Re:Nor is HDCP 2.2

[Gadget_Guy]

No for the majority of people it made no difference at all because at the first error they just jumped online and pirated the content.

No they didn't. The majority of people in the world do not pirate stuff. They do not have torrent software loaded. If they did then all forms of DRM would have died out years ago. DRM works because, as you said:

the majority of consumers do the "approved thing" with their copy

You never hear about the people who come up against the limits of DRM and simply accept it because they don't jump online to complain. If you only see the people who complain (which by definition you do) then you are seeing a skewed picture of the situation.

So I stand by my original statement that for the majority of people, the removal of DRM made little to no difference at all.

Mozilla will solve the many-core problem!

[Anonymous Coward]

A generation of experts will have to work to ensure computer math, science and games can often be spread over the many cores.

I think that Mozilla will take care of that problem.

They're working on Rust, which according to its home page is "a systems programming language that runs blazingly fast, prevents nearly all segfaults, and guarantees thread safety."

They're also working on Servo, which according to its home page is "a modern, performant browser engine designed to be appropriate for applications including embedded use. Written in Mozilla's new systems programming language, Rust, the Servo project aims to achieve better parallelism, security, modularity, and performance."

Wait, who the fuck am I kidding?! After witnessing the destruction of Firefox over the last few years, I have my doubts about the eventual success of both Rust and Servo. My only experience with Rust so far has been horrible. It's like C++, but done wrong, and C++ isn't done all that right, either! I tried Servo, too. Browsing modern web sites using Servo was like browsing them using Netscape Navigator 3: more seemed broken than was working!

Re:Mozilla will solve the many-core problem!

[Bruce Perens]

This is assuming that Mozilla will survive long enough to do any of this.

Re:Mozilla will solve the many-core problem!

[Anonymous Coward]

Bruce, Bruce, Bruce! Did you read to the last paragraph?

We're almost at the end with current tech

[tkrotchko]

14nm for these chips puts us close to the end of currently deployed technologies for transistor densities.

"The path beyond 14nm is treacherous, and by no means a sure thing, but with roadmaps from Intel and Applied Materials both hinting that 5nm is being research, we remain hopeful. Perhaps the better question to ask, though, is whether itâ(TM)s worth scaling to such tiny geometries. With each step down, the process becomes ever more complex, and thus more expensive and more likely to be plagued by low yields. There may be better gains to be had from moving sideways, to materials and architectures that can operate at faster frequencies and with more parallelism, rather than brute-forcing the continuation of Mooreâ(TM)s law."

http://www.extremetech.com/com...

Re:We're almost at the end with current tech

[Anonymous Coward]

We need to transition away from old fashioned digital computing and more more towards quantum computing. This is where the future lies. A quantum computer can perform calculations that would take old fashioned computers longer than the age of the current universe.

Re:We're almost at the end with current tech

[Anonymous Coward]

We need to transition away from old fashioned digital computing and more more towards quantum computing.

That's like walking over an unfinished bridge.

Re:We're almost at the end with current tech

[Anonymous Coward]

Quantum computers are slow at what classical computers are good at.

Re:We're almost at the end with current tech

[JoeMerchant]

We've been moving sideways for 10 years. In the 20 years before that, clock speeds were doubling every year or two. For the last 10, we've moved from a norm of single cores to a norm of 4 (or 2 + "Hyperthreads"), rotating hard drives to SSD, and specialized architectures to support HD video, but clock speed has been basically stagnant while the processors are getting fatter, more parallel, and not just in core count.

10 years ago, Intel was hinting at a massively parallel future (80 core processor rumored in development at the time), they've been slow to deliver on that in terms of core count, but are making progress on other fronts - especially helping single cores perform faster without a faster clock.

Re:We're almost at the end with current tech

[phantomfive]

I'm just hoping we make it to 20ghz before the party ends. Imagine all the processing you can do between frames at that speed!
Of course, faster would be better, but I don't have much hope.

Re:We're almost at the end with current tech

[AHuxley]

Just add more cores, threads to the consumer end :)

Re:We're almost at the end with current tech

[Gadget_Guy]

That's like walking over an unfinished bridge.

No problem. If you close your eyes the quantum bridge will be both finished and unfinished.

Re:We're almost at the end with current tech

[Anonymous Coward]

Modern transistors can run at 100ghz, but synchronizing all of the transistors is hard and we get left with 3.5ghz CPUs. Async CPUs could help a bit with this by reducing the amount of synchronization required.

Re:We're almost at the end with current tech

[gweihir]

There is a reason 5HGz seems to be a hard "wall" and around 4GHz commercial viability starts to end. It is interconnect. That is unlikely to go away anytime soon, if ever.

Re:We're almost at the end with current tech

[Zero__Kelvin]

All CPUs have both synchronous and asynchronous circuitry, so it makes no sense to say " Async CPUs could help a bit with this by reducing the amount of synchronization required."

Re: We're almost at the end with current tech

[Anonymous Coward]

High parallelism didn't come quite as Intel predicted, but I think you're not giving it enough credit. IBM tried ramping up the cores with the Cell, which turned out to be notoriously hard to work with. What we did end up with highly parallel GPUs.

Re:We're almost at the end with current tech

[phantomfive]

It is interconnect.

What does that mean? I'm not familiar with this.

Re:We're almost at the end with current tech

[Blaskowicz]

Can they even do stuff like addition, multiplication and conditional branching?

Re:We're almost at the end with current tech

[gman003]

We've been moving sideways for 10 years. In the 20 years before that, clock speeds were doubling every year or two. For the last 10, we've moved from a norm of single cores to a norm of 4 (or 2 + "Hyperthreads"), rotating hard drives to SSD, and specialized architectures to support HD video, but clock speed has been basically stagnant while the processors are getting fatter, more parallel, and not just in core count.

We hit a wall on MOSFET clock speeds way before we expected. Turns out that power consumption is quadratic, not linear, to clock speed. Once you get over 4GHz or so, it becomes a substantial problem, and getting over 5GHz is a real ordeal. There are ideas for non-FET transistors, but so far none has worked out.

10 years ago, Intel was hinting at a massively parallel future (80 core processor rumored in development at the time), they've been slow to deliver on that in terms of core count, but are making progress on other fronts - especially helping single cores perform faster without a faster clock.

Well, Intel was right. They just aren't CPUs, but GPUs. Even a bottom-end GPU will have 80 cores, the price/performance is pretty good all the way up to 1500 cores, and if you really want, you can get 4000-core cards. Those "cores" mean "ALUs", but even if you demand your cores have discrete schedulers, an R9 Fury has 64 compute units (scheduler + 64 ALUs), so 64 separate threads at once, each of which has massive SIMD power.

Re:We're almost at the end with current tech

[Anonymous Coward]

Intel did release (and continue to iterate on) Xeon Phi which is what they call their 80 or so core processors. The problem is the software, not the hardware. Outside of scientific computing, there's not much you can do with 80+ cores, especially among tasks that you can't do better on a GPU.

Re:We're almost at the end with current tech

[Kjella]

The real problem is that we're mostly redistributing the watts.

4 core @ 4GHz (i7-4790K) = 91W, 4*4/91 = 0.175 GHz/W
4 core @ 3.2GHz (i7-4790S) = 65W, 4*3.2/65 = 0.197 GHz/W
4 core @ 2.2GHz (i7-4790T) = 35W, 4.*2.2/35 = 0.251 GHz/W

So from top to bottom we're seeing 40% better perf/W with perfect linear scaling. Neat, buit not exactly revolutionary when you subtract overhead. We've already got so much scale out capability that power is clearly the limiting factor:

8 core @ 4GHz (doesn't exist) = ~185W
8 core @ 3.2GHz (1680v3) = 140W
8 core @ 2.2GHz (2618Lv3) = 75W
16 core @ 4GHz (doesn't exist) = ~370W
16 core @ 3.2GHz (doesn't exist) = ~280W
16 core @ 2.2GHz (E7-8860v3) = 165W

We can't go faster or wider unless we find a way to do it more efficiently, either that or we need extremely beefy PSUs and water cooling.

Re:We're almost at the end with current tech

[Anonymous Coward]

The amount of computation per clock is not fixed. At this point, increasing clock speeds is effectively impossible (with the exception of some unforeseen future breakthrough, of course), while performing more work on each clock cycle is easier. Which is why Intel keeps adding new instructions (and presumably is reducing the clock count for existing commonly used instructions).

Re:We're almost at the end with current tech

[gweihir]

Chips basically have components (transistors, diodes, capacitors, resistors, and recently inductors) and interconnect ('wires').

Interconnect has been the primary speed-limiter for about 20 years. At 5GHz or so, it starts to become exceptionally difficult to get signals from one component to the next, and in particular distributing clocks becomes a limiting issue as clocks need long wires in order to reach everything. Making transistors smaller helps a bit because the wires get shorter and signal-strength (voltage) can be reduced. But that effect is limited and seems to have mostly reached its end.

The overall effect is that extreme over-clocking for the 15 year old Pentium 4 could reach 8GHz, but current chips do not do much better with both AMD and Intel going up to about 9...10GHz as absolute maximum.

Re:We're almost at the end with current tech

[Anonymous Coward]

Well, Intel was right. They just aren't CPUs, but GPUs.

Too bad isn't really in the business of making GPUs: (compare Skylake integrated to GTX 780 Ti and it's no real comparison.

Even a bottom-end GPU will have 80 cores, ...

Yea, only up to 72 cores for Intel.

... the price/performance is pretty good all the way up to 1500 cores, and if you really want, you can get 4000-core cards.

A Titan Z apparently has 5760 CUDA cores.

Those "cores" mean "ALUs", but even if you demand your cores have discrete schedulers, an R9 Fury has 64 compute units (scheduler + 64 ALUs), so 64 separate threads at once, each of which has massive SIMD power.

Yet what does that really mean? The general issue is that having massive SIMD power is pretty much useless most the time for most things. So, it's used for graphics, media decoding, physics, and various super-computing related needs. Meanwhile, 99% of code still runs on the CPU. But all those GPU cores are still sucking up tons of power doing basically nothing. That doesn't seem like much of a future to me.

Re:We're almost at the end with current tech

[Anonymous Coward]

That this pesky guy named Albert Einstein (with the help of James Clerk Maxwell) limited the velocity at which data can be transmitted at 300000km/s. At 5GHz (200pS cycle time), this is about 6cm, but actually less because of the dielectric constant of the environment in which conductors work in integrated circuits. If Einstein had never existed we could have faster circuits, blame him for these speed limits.
Joke apart, small geometries also means that you have to charge and discharge gate capacitance through fairly long and thin wires as soon as a structure is of decent size (a cache array). Signals coming from a memory (i.e., cache) array are weak (large capacitance on bit lines) and must be converted to logic through sense amplifiers (signal to noise problem, noise proportional to square root of bandwidth).
Previous generation of IBM mainframes ran at 5.5GHz (as sold, no overclocking), new ones dropped to 5GHz. These are still great machines, thanks to the memory controller chips which include (each one, typically 2) 480MB of L4 cache to hide DDR3 latency and reorganize memory access queues.

Re:We're almost at the end with current tech

[nateman1352]

10 years ago, Intel was hinting at a massively parallel future (80 core processor rumored in development at the time)

I think the 80 core processor Intel was developing at the time eventually turned in to the Knights Corner aka Xeon Phi chip. Originally Intel developed this tech for the Larrabee project, which was intended to be a discrete GPU built out of a huge number of X86 cores. The thought was if you threw enough X86 cores at the problem, even software rendering on all those cores would be fast. As projects like llvmpipe and OpenSWR have shown, given a huge number of X86 cores this isn't as crazy of an idea as it initially sounds... but still a little crazy :) Ultimately Intel cancelled that project and decided to use that tech for super computing instead of graphics. A result of this is Intel retained the "Gen" design for their graphics core, which is a more traditional GPU design.

Re:We're almost at the end with current tech

[phantomfive]

But that effect is limited and seems to have mostly reached its end.

Does this mean that as features get smaller, the interconnects have not?

Re:We're almost at the end with current tech

[tkrotchko]

I don't have mod points today, but this is funny joke. Well played.

Re:We're almost at the end with current tech

[Anonymous Coward]

They both can and can not

Re:We're almost at the end with current tech

[Alomex]

10 years ago, Intel was hinting at a massively parallel future (80 core processor rumored in development at the time),

An Intel higher up told me a while back that they could ship them today if they wanted. The problem is that users in the field report having a hard time using more than 6 cores outside host virtualization. Since then Intel has been dedicating the extra real estate to more cache, which programs can easily take advantage of, and less to cores, which no one knows quite how to use beyond 6 to 8 cores.

Re:We're almost at the end with current tech

[Anonymous Coward]

Turns out that power consumption is quadratic, not linear, to clock speed

Yes and not. Power is linear to clock speed, but quadratic to voltage, and the voltage required increases quadratically with clock speed. Directly linear, indirectly quadratic. And CPU clocking is complex. You don't always have to increase voltage to increase clock speed. On average it's quadratic, but there are plateaus of stability where the CPU just works well for a wide range of frequencies within a very small voltage range.

Re:We're almost at the end with current tech

[thegarbz]

We can't go faster or wider unless we find a way to do it more efficiently

Isn't that exactly what Intel has been doing for the past decade anyway?
1 core @ 3GHz (Pentium 4) = 89W 1*3/89 = 0.033GHz/W
4 core @ 2.4GHz (Core2Duo Q6600) = 105W 4*2.4/105 = 0.091GHz/W

(Both previous processors to my current i7)

Re:We're almost at the end with current tech

[JoeMerchant]

Agreed - cooling is the issue, and moving to smaller feature sizes (22nm, 14nm, 5!?!nm) is improving thermal efficiency, while simultaneously shrinking packages, making things like the Cedar Trail Compute Stick a possibility. People who really need 1000 core machines are getting them today, smaller, cheaper, and lower power than ever - if there were a market, you could shoehorn about 50 of your 4GHz cores into a "Full Size Tower" case that wasn't at all unusual (size-wise) 20 years ago - dissipating ~1000W out of a single case would be "extreme" but not exotic yet.

Re:We're almost at the end with current tech

[Anonymous Coward]

Millions are running AMD's 8-core 4.7GHz FX-9590 with a 220W TDP, so the market for cooling at those TDP levels is already there.

Re:We're almost at the end with current tech

[gweihir]

Interconnect gets smaller if you reduce speed as well when you reduce size. If you keep speed constant, interconnect stays the same size and it will consume the same amount of power. Well, roughly. The problem is that at these speeds you are dealing with RF laws, not ordinary electric ones and RF laws are pretty bizarre.

Re:We're almost at the end with current tech

[JoeMerchant]

All depends on the app. In 2008 I was doing some signal processing work that would have easily parallelized out to 22 cores, and probably get partial benefit up to 80+ cores - nature of the source data (22 time series signals going through similar processing chains, the chains themselves might not get use out of more than 4-8 cores, but there are 22 of these things, so....)

Lots of video processing work can be trivially split up by frame, so if you don't mind a couple seconds of processing delay, you can grab 80 frames and throw them one to a core... Other video processing work wants to go sequentially frame by frame, and would have to gather these results to chew on them.

Re:We're almost at the end with current tech

[Alomex]

...and all of those are ideal for a GPU not extra cores, which brings us back to the Intel problem. Either is embarrassingly parallelizable (hence GPU) or you have a hard time using more than a handful of cores via multi-threading (hence 6-8 cores in most upper-end, non-virtualization CPUs).

Re:We're almost at the end with current tech

[slew]

Interconnect gets smaller if you reduce speed as well when you reduce size. If you keep speed constant, interconnect stays the same size and it will consume the same amount of power. Well, roughly. The problem is that at these speeds you are dealing with RF laws, not ordinary electric ones and RF laws are pretty bizarre.

The problem can easily be described to first order "electrically". No bizarre RF laws necessary.

Interconnect is dominated by "resistive" issue (a good approximation of RF-impedance) and capactive coupling (a good approximation to RF field effects)... Since the interconnect is relatively getting thinner and longer, the resistance of that wire is going up (R ~ L/w/h) and it capacitively couples more with nearby lines (Cild = W*L/X or Cimd = H*L/Ls) and makes it take longer to move charge to and from the gate.

Second order effects are mostly "noise" and edge-rate coupling, but even then aggressor/victim and crosstalk issues can be thought of mostly as just distributed "lumped" approximation (e.g., capacitance per um, and mutual inductance per um) where the result is coupling being different at higher frequencies and spacing. No bizarre RF need to get the gist (well, no more than the basic concept of a wall-wart transformer)...

Re:We're almost at the end with current tech

[Anonymous Coward]

you can have 8 cores @ 5GHz:
https://en.wikipedia.org/wiki/IBM_z13_%28microprocessor%29
it is water cooled.

Re:We're almost at the end with current tech

[JoeMerchant]

Back in 2008, CUDA and friends were too bleeding edge for the applications I was working on, plus - a standard desktop PC had acceptable performance, so why kill yourself with exotica? Since then, I haven't had any applications where CUDA would have been practical, well, o.k., I did work with a group that did video processing who _should_ have been using CUDA, but they were having enough trouble keeping their stuff stable on ordinary servers.

And, that 22 signal application, probably would be a major pain to port into CUDA, even today - it was a port out of Fortran into C++, and some of the Fortran code did some fairly exotic stuff - not found in your normal signal processing toolkit (written, validated, and used by statisticians...)

Re:We're almost at the end with current tech

[Anonymous Coward]

Of course they have both, rarely is a complex object in the Universe exactly all one thing. Modern CPUs have clock domains, but compared to the frequencies they run and the number of transistors in those domains, they're huge. "Async" CPUs focus on extremely small domains, allows much faster clocks. The point is a fully async CPU could run about 100ghz, assuming none of the transistors needed to be synced with another.

Re:We're almost at the end with current tech

[martinfb]

...or... quantum computing?! ... or... Getting software developers to cut the bloat-ware BS and get serious about EFFICIENT, bug-free, secure apps, that allow faster performance?!

Re:We're almost at the end with current tech

[lucien86]

With modern chips, if you went say 4-bit and cut every corner you could I would think a single chip with 10 million plus cores would be possible. Not useful but possible.

Re:We're almost at the end with current tech

[gweihir]

You know, it can possibly be described by fairies and dragons as well. That would just be a fantasy as much as your "description" is.

Re:We're almost at the end with current tech

[Agripa]

It is actually worse or better than that depending on your viewpoint.

Over the last several generations the limit has been power density. If you make a plot of total power versus chip area going back through at least the beginning of the Core2 line of processors, the power density is roughly constant. In addition, total chip area has decreased because process density has increased faster than area needed to implement the processor. The result is that power has decreased roughly following the decreasing chip area and newer chips are less expensive based on area and draw less power because they have no other choice do to physics.

So increasing the area used for a given number of transistors could offset higher clock rates however that will result in more expensive chips both because of fewer chips per wafer and lower yields do to defects. If you are someone like IBM that does not matter and they make the largest physical chip they can with power ratings to match but Intel only does that for server CPUs and check out how much they cost.

Re:Donald Trump

[Anonymous Coward]

That's to keep vampires away. He has a non-stop line of people kissing his butt, and he doesn't trust the TSA (tushie security agency) to keep the vampires off his ass.

Looks Like My i7-920 @3.8 Ghz

[zenlessyank]

Just needs to last 5 more years. Hopefully a real reason to upgrade will happen around 2020.

Re:Looks Like My i7-920 @3.8 Ghz

[ArylAkamov]

Agreed. I'm perfectly happy on my 920, though it does tend to run hot it can overclock like a beast. I'm starting to think my motherboard won't support modern video cards before I have an actual need to upgrade my processor.

It seems like most applications still benefit far more from increased clock speed than core count.

Re:Looks Like My i7-920 @3.8 Ghz

[ArylAkamov]

Either I can't find the edit post button or it doesn't exist. I'm curious what cooler you're using to get to 3.8 Ghz, I've been having temperature problems with mine.

Re:Looks Like My i7-920 @3.8 Ghz

[Blaskowicz]

Do you change thermal paste every two years or so

Re:Looks Like My i7-920 @3.8 Ghz

[ArylAkamov]

Yep, arctic silver 5. Core temp still sends me into shutdown at 95c a few times a week though.

Re:Looks Like My i7-920 @3.8 Ghz

[zenlessyank]

I'm using a Thermax Eclipse 2 with an Antec 120 and Noctua 120 doing push-pull work since the fans that came with that cooler were cheap and died after 2 years. Doubt you will find one anymore but a Noctua D15 should work just as well.

Re:Looks Like My i7-920 @3.8 Ghz

[bemymonkey]

And that keeps you happy? Sounds extremely unreliable...

Don't get me wrong, I'm running a 2500k @ 4.5GHz myself, but I've never had a crash since I set the system up a few years ago.

Re:Looks Like My i7-920 @3.8 Ghz

[hairyfeet]

Sounds like your cooling sucks, maybe you should try something like this. this is the one I use on my FX-8320E and while that normally isn't a hot chip but when I'm doing a ton of A/V work I'll use the Asus OCing tool that came with my board and even with the chip running full bore with 8 cores at 4Ghz it still keeps temps reasonable,IIRC the highest I ever got was 130F and without OC it stays a nice cool 80F according to the board sensors. To see how it performs with an Intel here is a review and as you can see it scored just 8 degrees higher than the huge Corsair H100.

I personally love it, its quiet, easy to install, and works well.

Re:Looks Like My i7-920 @3.8 Ghz

[Anonymous Coward]

Just needs to last 5 more years. Hopefully a real reason to upgrade will happen around 2020.

Nope. They're going to milk this for all it's worth, so probably 2027 would be a nice time to upgrade.

Mozilla Foundation now gets money from Microsoft.

[Futurepower(R)]

Good point. Mozilla Foundation now gets most of its money from Microsoft. Microsoft pays Yahoo. Yahoo pays Mozilla Foundation to make "Yahoo search" (actually Microsoft Bing search) the default search engine in Firefox. Most people don't have the technical knowledge to know how they've been manipulated, or how to restore the default search engine to Google search.

Thunderbird and SeaMonkey Composer GUIs: Damaged, apparently deliberately. Every time you do a file save, the newer versions of both ask for a new file name, and don't suggest the last one chosen. The damage was reported several months ago, but has not been fixed.

It makes for faster backup & restores too

[Anonymous Coward]

See subject: Says it all - it was 'left over' from my 920 & still worked, so - there ya are (the caching controller & 10k disk = faster than the SATA III 1TB WD storage disk I have, by far (it's noticeable)).

APK

P.S.=> To each his own though - however, it doesn't take a brain to realize that your PURE unidentifiable AC post means you downmodded my post http://hardware.slashdot.org/c... you replied to also (is your favorite color 'transparent'? I say that since I see RIGHT THRU YOU...)

... apk

What has Dice done to you, Mr. Perens?

[Anonymous Coward]

You are one of the handful of special folks who have a default +2 score on all your comments

Without the many wonderful things true pioneers like you, Mr. Perens (and others) there would be no ecosystem for Dice to survive in the first place!

Dice - stop being so disrespectul !!

Re:Donald Trump

[Anonymous Coward]

Specifically, the dread Count Soros.

Re:Donald Trump

[Zero__Kelvin]

I thought I smelled garlic on your breath!

III-V semiconductors

[blind biker]

As a materials scientist, I think they squeezed the last bit of potential out of silicon. Well, they could perhaps go for isotopycally pure silicon, but the gain would be relatively modest for a high price. III-V semiconductors such as GaAs, InGaAs etc. are expensive mostly because it's hard to grow large crystals, but it is worth it due to the far higher mobilities of electrons in them.

Re:III-V semiconductors

[Anonymous Coward]

As a materials scientist, I think they squeezed the last bit of potential out of silicon. Well, they could perhaps go for isotopycally pure silicon, but the gain would be relatively modest for a high price. III-V semiconductors such as GaAs, InGaAs etc. are expensive mostly because it's hard to grow large crystals, but it is worth it due to the far higher mobilities of electrons in them.

Mobile electrons help, but that's not the limiting factor these days, it's leakage. The problem with leakage is that small feature sizes mean lots of leakage and small feature sizes are needed to cram billions of transistors into an economical die size.

Having big-fast transistors won't really save the industry, we've been relying on more transistors for the same $$$ to drive the industry forward and got a free ride on performance increases per transistor for a while and more mobile electrons will help with switching speed, but not with leakage an that's one of the biggest part of the thermal budget these days. Fin-Fet just makes everything bigger to reduce leakage, so although the minimum feature size went down between 20nm and 16nm (for TSMC or 32nm and 22nm for Intel), the transistor density stayed about the same to keep the leakage under control. By not really shrinking, they killed the economics of moving to the new feature size that generation (because the more radical technology costs more, so the only way to pay for it is to get more die per wafer to offset this premium.

Simply moving to brittle GaAs (or other similar semiconductors), are unlikely to result in significant increases in transistors for a given price and thus won't be able to save the industry...

Roadmaps are for cows.

[Anonymous Coward]

You are all cows. Cows say moo. MOOOO! MOOOO! Moo cows MOOOO! Moo say the cows. YOU LOST COWS!!

Tick? Tock?

[hackwrench]

Tick you're alive; tock it was nice knowing you.

Re:Tick? Tock?

[Anonymous Coward]

Tick you're alive; tock it was nice knowing you.

This one is basically a tock-tick / tick-tock (Nice knowing you're alive) .

I had one of those (good stuff)... apk

[Anonymous Coward]

See subject: It lasted me from 2010 to earlier this year (& still runs, a ramchip went 'out' was all) - was a great system for it's time (one of the best I ever owned in fact).

HOWEVER - I stepped up to this setup (& it's literally 35% faster on tasks I run typically that are heavy string processing off file on disk into memory - hosts file data via -> APK Hosts File Engine 9.0++ SR-2 32/64-bit http://start64.com/index.php?o... )

ASUS B85-E Motherboard
Intel Core I7 4790k CPU (vs. my last CPU Core I7 920 -> http://www.anandtech.com/bench... )
EVGA/NVidia GeForce 970 GTX video OC stock-oem (+140mhz) 4gb GDDR5 RAM (vs. my last vidcard 470 GTX -> http://www.anandtech.com/bench... )
Intel 530 240gb Flash SSD (SATA 6) - strictly OS & Program disk - latest 3.0 firmware & trim tools (vs. my WD Velociraptor -> http://www.anandtech.com/bench... )
Western Digital 10,000 rpm 8mb buffer Velociraptor 150gb (SATA II) - strictly for backup & programming data
Promise Ex-8350 128mb ECC ram caching raid sata 1/2 controller (SATA 1/2) - for backup WD Velociraptor
GigaByte IRAM 4gb DDR2-Ram based SSD (SATA I) - strictly for PageFile placement
Western Digital 7,200 rpm 8mb buffer 1tb (SATA 6) - strictly for downloads
HP DVD+-RW Dvd 1265i Burner (SATA 3)
8gb Kingston DDR-3 RAM (1gb for 64-bit NTFS Compressed Software RamDrive = webbrowser cache, hosts file location, print spooler, %TEMP% ops, + %COMSPEC% location)

APK

P.S.=> It all yields FAR BETTER performance locally vs. that 920 (great machine though it was) - just something to consider between now & then (especially as prices drop): Still, I see YOUR point though - "wait it out" for the REAL boosts in speed (especially regarding SSD's I see coming that are going to blow away what we have now - & speeding up the slowest parts of systems are where the TRUE speed gains really show themselves)

... apk