Intel 4-5 year tick/tock cycles when AMD is done?

[Hugo Drax]

With AMD dying. I see already the smoke, delayed Broadwell, 2014 gets only a Haswell refresh.

I wonder if Intel will go back to 90s style cycles of 4-5 years between new cpus. Milk the most out of a process before jumping to the next one.

It would make sense since it seems each new cpu cycle lately seems to provide small gains of 10-15%

also a price increase of 10-15% would probably happen as well to help increase margins.

I see this happening once AMD is gone.

 

[NTMBK]

With AMD dying. I see already the smoke, delayed Broadwell, 2014 gets only a Haswell refresh.

I wonder if Intel will go back to 90s style cycles of 4-5 years between new cpus. Milk the most out of a process before jumping to the next one.

It would make sense since it seems each new cpu cycle lately seems to provide small gains of 10-15%

also a price increase of 10-15% would probably happen as well to help increase margins.

I see this happening once AMD is gone.

If Intel stalls CPU development, there is a danger that some ARM developer would get a foothold in their strongholds (server, desktop, laptops) and their dreams of taking mobile from ARM would be dead. It's not going to happen.

 

[VirtualLarry]

Yes, with ARM nipping at their heels, and the necessity to maintain their process node advantage, they need as much volume for those process nodes as possible, so it seems likely that they will still produce desktop and server CPUs on advanced process nodes as well as their Atoms and Quarks,

 

[sm625]

That's what intel has been doing. SB is basically nehalem. We havent had any true innovations in their architecture since then. A processor from 5 years ago is in the neighborhood of the latest and greatest haswell in terms of raw performance. But in the smartphone space, a processor from 5 years ago can barely even run a modern smartphone OS or the typical app suite.

But even without AMD the market is still punishing them for it. Intel's lack of innovation has allowed smartphones to pop up like weeds, sort of like how weeds grow around a boulder. Billions of em, 99% powered by competitors.

In 5 years the majority of desktop PC users will have entirely replaced their desktop with a smartphone. Expect there to be no more than one actual pc and several smartphones connectable to several external displays (dumb terminals). In other words, expect no more than one intel x86 microprocessor in each home. I'd say the market is delivering intel the justice it deserves.

 

[Hugo Drax]

Well Intel will probably focus on designing more advanced Atom cpus and focus their efforts in that area. But I think the Desktop market is pretty saturated when it comes to idle MIPS. No need to rush.

They can focus the design teams on making better Atom chips to compete against ARM and milk the desktop market with extended cycles and higher prices.

 

[witeken]

With AMD dying. I see already the smoke, delayed Broadwell, 2014 gets only a Haswell refresh.

I wonder if Intel will go back to 90s style cycles of 4-5 years between new cpus. Milk the most out of a process before jumping to the next one.

It would make sense since it seems each new cpu cycle lately seems to provide small gains of 10-15%

also a price increase of 10-15% would probably happen as well to help increase margins.

I see this happening once AMD is gone.

AMD hasn't been a threat to Intel for years. Intel isn't going to slow down their tick-tock cycle, that doesn't make any sense. Intel's whole advantage is built around their 2-4 year process lead, which gives Intel an enormous advantages in all possible ways: smaller transistors are faster, more energy efficient and ~2 times cheaper. So even if Intel didn't see the slightest need to make their processors faster, they would still have a reason to move to lower process nodes: higher margins. But if Intel would stop now, TSMC, Samsung and Global Foundries will catch up, removing Intel's biggest weapon to quickly move into the smartphone and tablet market.

It's very simple in the technological industry: if you stop innovating, you will fall behind and your income will collapse.

I don't know Intel's motivations for Haswell refresh, but I assume Atom has something to do with it. Certainly after this 1 quarter delay, they probably want the most important products to be released as fast as possible. Broadwell will give a nice power consumption improvement, but that's least important in desktops, which aren't bottlenecked by power consumption.

So I don't think your statements make a lot of sense, and we'll see Skylake in H2 Q15, Cannonlake at the end of 2016, and so on.

 

[azazel1024]

Unless I badly missed something, Broadcom is still coming this year, just Summer/Fall. I still haven't seen anything concrete on this Haswell refresh. The best I am getting is Broadwell's chipset will debut with it, and we'll get minor clock speed bumps on Haswell processors, which happens almost always on Intel's product lines. I think Ivy was the last time we didn't see it (or maybe we did and I missed it?).

That happens sometime this spring with Broadwell still expected in at least limited numbers this summer.

I know Sandy, Nahalem and others had that tick, minor tick, tock, minor tock going on.

i5-2500...and then 9 months later, i5-2600 and so on.

Generally Intel improves their process and yeilds and the end result is 8-10 months on we get a new release of, generally, 100MHz clock bumped processors at the same or slightly higher price points to what Intel initially released, often times with slight real world power savings (but the same TDP).

Intel has acknowledged issues with 14nm, delaying the release by roughly 6 months according to them.

Try TSMC and GloFo who have mostly conceeded that 20nm is/has been delayed a minimum of 12 months off their originally projected time line (of 2 years after 28nm). It sounds like they'll be likely if there are ANY shipping 20nm chips from any of the fabs before the end of the year, looking more like early/mid next year is much more likely, putting them 2 full nodes behind Intel for at least 6-12 months.

I think these smaller nodes are proving much more difficult for everyone. If the newest EUV lithography gear proves effective, it might change that, maybe, but things don't sound promising even with it. Intel has talked about playing with 10nm in the lab and I know some universities are working with 6-10nm too, but I am not terribly confident that 10nm is going to be able to be ready 2 years after 14nm for Intel and below that...well, it may not be workable (Intel has prototype 10nm according to them, so I assume that means they will be able to eventually take it to a shipping product. Eventually).

Below 8-10nm we might well be talking needing something other than Silicon to handle the smaller sizes, or a hybrid silicon doped with something. Or maybe nothing will work as a transistor effectively below those sizes and we'll have to develop effective photonics if we want to increase performance per watt.

We can increase processing power density by chip stacking at some point, but that doesn't solve power consumption/performance per watt, which is the overriding issue for most computing.

 

[ShintaiDK]

With AMD dying. I see already the smoke, delayed Broadwell, 2014 gets only a Haswell refresh.

I wonder if Intel will go back to 90s style cycles of 4-5 years between new cpus. Milk the most out of a process before jumping to the next one.

It would make sense since it seems each new cpu cycle lately seems to provide small gains of 10-15%

also a price increase of 10-15% would probably happen as well to help increase margins.

I see this happening once AMD is gone.

First of all, Intel havent had any competition the last 7-8 years.

Secondly, ask yourself. Would you upgrade your CPU if there wasnt something new? The answer is obviously no.

Hence leading to the next part. Intel needs a high volume to keep the fabs running. High cashflow and so on. To do what you purpose would actually bankrupt the company. ARM is completely irrelevant in that case.

Remember, the first CPU cost billions, the next a few $. The more of the next you can sell, the better. Intel would lose profit by raising prices from the current level due to decrease in volume. No innovation would mean lower volume and you are back to lower profit, closed fabs, higher R&D cost per chip and the cycle continues.

Intels CAPEX and R&D spending havent been higher since AMD ceased to be competition. And CPUs havent been cheaper either.

Remember 10 years back? You would have to pay 500$+ for something like a 4670K. And thats without adding inflation that would run up another 20-25% or so on the 500$+.

Now look at the place with most (duopoly) competition, GPUs. Price is very high there.

 

[ElFenix]

With AMD dying. I see already the smoke, delayed Broadwell, 2014 gets only a Haswell refresh.

I wonder if Intel will go back to 90s style cycles of 4-5 years between new cpus. Milk the most out of a process before jumping to the next one.

It would make sense since it seems each new cpu cycle lately seems to provide small gains of 10-15%

also a price increase of 10-15% would probably happen as well to help increase margins.

I see this happening once AMD is gone.

no, ARM will drink their milkshake if the do that.

 

[witeken]

That's what intel has been doing. SB is basically nehalem. We havent had any true innovations in their architecture since then. A processor from 5 years ago is in the neighborhood of the latest and greatest haswell in terms of raw performance. But in the smartphone space, a processor from 5 years ago can barely even run a modern smartphone OS or the typical app suite.

Have you ever heard of the dead of Dennard scaling? A good microarchitecture can only get you so far, the biggest performance improvements come from higher clock speeds, which we haven't seen in about 10 years.

It's logical that smartphone processors have improved much faster: smartphones only exist for so long, and in a volume that can give you nice income. So CPUs were worse than what they could have been and there's been a lot competition in that market. Also, these TDPs didn't suffer as much from the dead of Dennard scaling yet as desktop TDPs did. At the same time, peak power consumption did increase quite a lot.

But even without AMD the market is still punishing them for it. Intel's lack of innovation has allowed smartphones to pop up like weeds, sort of like how weeds grow around a boulder. Billions of em, 99% powered by competitors.

I don't see what Intel has to do with it. Certainly their 'lack of innovation'. As far as I know, Intel is by far the most innovative company out there (for example, pushing Moore's law at an impressive pace).

Smartphones popped up because they're convenient: you have a basically a small computer in your pocket, and they're also cheaper than regular computers and more personal so smartphones aren't shared between family members. Intel has explained their absence in the smartphone market already quite some times, here's Brian Krzanich's recent statement on Reddit:

we wanted the world of computing to stop at PC's...and the world.. as it never does... didn't stop innovating.. The new CEO of Microsoft Satya said it well the other day... our industry does not respect tradition, it respects innovation.... i think he was 100% right.. and it's why we missed the mobile move..

In 5 years the majority of desktop PC users will have entirely replaced their desktop with a smartphone. Expect there to be no more than one actual pc and several smartphones connectable to several external displays (dumb terminals). In other words, expect no more than one intel x86 microprocessor in each home. I'd say the market is delivering intel the justice it deserves.

Maybe, but certainly all those smartphones are going to powered by Intel inside .

 

[spat55]

If Intel did that and let themselves get that far behind then someone else could just swoop in with their ideas and make them have to fight again to get themselves to where they are now.

 

[escrow4]

AMD (on the CPU side) is already gone. AM3+ is a dead man walking, FX, well let's not go there, and APU's are niche to put it mildly. Constantly refining Atom will likely stave off ARM, especially if people want tablets for work (i.e. Win 8.1+), not pretty toys like iPad/ iOS, (slightly) more CPU grunt will still be required there - I wouldn't be surprised if we see P4 -> Core 2 level jumps in Atom performance as tablets/phones increase in power.

The desktop will be there in the corner with tiny jumps - I have 2 4770 builds now and upgrading to Broadwell/Haswell refresh for either of my boxes makes little sense. On the gaming side, maybe Haswell E if games really need the threads, on the office side, no need.

 

[NTMBK]

First of all, Intel havent had any competition the last 7-8 years.

A certain Anand Lal Shimpi would disagree with you.

Phenom II is AMD's return to competition, and unlike the best the original could do, the sequel is actually worthwhile.

http://www.anandtech.com/show/2702/20

It was Sandy Bridge and Bulldozer when the wheels really fell off for AMD, not Conroe.

 

[el etro]

Many people may interpret I wrong, but this title

Intel 4-5 year tick/tock cycles when AMD is done?

Is flamebait, no?



Three days ago we spot a boy opening a thread with the same flamebait in title, and now it happens.

BTW, to OP:

With AMD dying. I see already the smoke, delayed Broadwell, 2014 gets only a Haswell refresh.

I wonder if Intel will go back to 90s style cycles of 4-5 years between new cpus. Milk the most out of a process before jumping to the next one.

In 90's, processors IPC don't evolved so much as you're thinking over the years, but frequency had(10 times from 2000 to 1994). Processors were becoming faster with more and more growth in in power consumption/heat dissipation. In over 2002 we're already had a processor with ~10 times Higher TDP and ~25 times higher Frequency.
Nowdays processors manufacturers effort is to, while still increasing IPC and not so much frequencies, make processors fit into lower TDPs in order to reduce the PC forms that the processors will go. Letting high TDP processors to compete in the High-End.

 

[Durp]

Many people may interpret I wrong, but this title

Is flamebait, no?



Three days ago we spot a boy opening a thread with the same flamebait in title, and now it happens.

This is not even remotely the same situation. This would only be thought of as flamebait if the person reading it was emotionally or financially attached to AMD.

This is a user looking at facts from the current and previous years of AMD declining and asking how Intel will react when AMD ducks out of the CPU market entirely which looks inevitable.

And to stay on topic: I think it's clear Intel has already been doing this since Sandy was released. They "parked the bus" since then.

 

[senseamp]

If AMD is gone, Intel will fall under intense anti-trust pressure to open up x86 licensing. So competition will only increase.

 

[ninaholic37]

Works for me. I have no problem with using my Pentium M CPU from 2004 to browse the net for another 10 years as the high-end keeps seeing minimal gains.

 

[jmartin357]

did I miss a memo or something? amd is dead?

 

[ShintaiDK]

A certain Anand Lal Shimpi would disagree with you.



http://www.anandtech.com/show/2702/20

It was Sandy Bridge and Bulldozer when the wheels really fell off for AMD, not Conroe.

Competition?

 

[ShintaiDK]

If AMD is gone, Intel will fall under intense anti-trust pressure to open up x86 licensing. So competition will only increase.

There is nothing stopping anyone from making x86 CPUs besides economics. Those patents expired long ago.

And no CEO in his right mind would want to compete with Intel using x86. Just see how it goes for AMD or the other 20 or so companies that had to raise the white flag.

Its questionable if there is even room for competition due to the increased R&D and IC design cost. x86 volume would have to increase dramaticly for any economic worthwhile competition to happen. Not to mention the forever behind foundries for fabless companies. Qualcomm and Apple already uses more on R&D for ARM chips than AMD does on ARM, x86 and GPUs combined.

Its not competition to have 5 companies making phenom/bulldozer uarch chips for example with 28nm chips into 2015.

 

[NTMBK]

Competition?

Yeah, competition. Not beating, but competing. The conclusion of that article makes it pretty clear, and cherry picking a single benchmark from it isn't going to alter that. Another snippet from the conclusion:

Compared to the Core 2 Quad Q9400, the Phenom II X4 940 is clearly the better pick. While it's not faster across the board, more often than not the 940 is equal to or faster than the Q9400.

That cherry picked benchmark you posted would be an example of "not faster across the board".

 

[NTMBK]

There is nothing stopping anyone from making x86 CPUs besides economics. Those patents expired long ago.

If you want to make a Pentium compatible CPU, sure. But good look making anything with modern, competitive instruction sets. There's a reason why NVidia weren't able to make Denver x86 like they wanted to.

 

[jpiniero]

Intel has explained their absence in the smartphone market

I wouldn't really buy any excuse Intel says about their absence. It really boils down to their attachment to Windows.

Remember, the first CPU cost billions, the next a few $. The more of the next you can sell, the better. Intel would lose profit by raising prices from the current level due to decrease in volume. No innovation would mean lower volume and you are back to lower profit, closed fabs, higher R&D cost per chip and the cycle continues.

The problem is that Intel's business model isn't really built to be mainly 20,30 dollar chips because of the massive fab costs. Basically it boils down to getting the processors as small/cheap/power thrifty as possible while increasing prices on people who need performance.

Competition?

IIRC, AMD had close to 50% share of PC desktops for the longest time, even long after Conroe's release.

 

[ShintaiDK]

Yeah, competition. Not beating, but competing. The conclusion of that article makes it pretty clear, and cherry picking a single benchmark from it isn't going to alter that. Another snippet from the conclusion:



That cherry picked benchmark you posted would be an example of "not faster across the board".

The Q9400 was 75-100$ cheaper, 30W less TDP.

 

[ShintaiDK]

If you want to make a Pentium compatible CPU, sure. But good look making anything with modern, competitive instruction sets. There's a reason why NVidia weren't able to make Denver x86 like they wanted to.

They could release a SSE2/x64 CPU today. But it sounds more like a bad excuse. Because nVidia is simply too small to be competitive in that segment. They simply dont have the R&D budget or design team for it.