What will cause a $100 core i5 to happen?

[john5220]

What would it take to bring down a i5 to $100?

What kind of processor would AMD need? Would K12 allow AMD to once again compete with intel?

Is it possible for intel to sell a i5 for $100 and still make a profit?

 

[postmortemIA]

of course they can sell it that that price. But jut like everything else, they charge what they can get away with.

The cost of whole ARM SoCs is pretty cheap. $100 a piece means they can't fund (read sell at loss) their tablet business

case in point
http://www.microsoftstore.com/store...-Signature-Edition-Tablet/productID.308781500

 

[Yuriman]

My best guesses:

For Intel to price their i5 at $100, AMD would need to build an FM2 chip with around 60% better IPC and sell it for 2/3 what their current top chips are going for. That would put pressure on Intel to drop prices. Intel will continue to ignore AMD's iGPU gaming advantage regardless, because the vast majority of consumers don't game. Intel's iGPU already features just about everything it needs (4K support, hardware accelerated encoding/decoding, etc.) and is compatible with the latest directx.

It's difficult to answer the profit question, because if you sell a chip at the cost needed to build it, you're not making any money to put back into designing the next chip.

Die size and transistor count comparison:

Kaveri 4 core - 245mm²- 2.41 billion
Piledriver 8 core - 315mm² - 1.2 billion
Ivy Bridge 6 cores - 256mm² - 1.86 billion
Ivy Bridge 4 core -160mm² - 1.2 billion
Ivy Bridge 2 core - 94mm²
Haswell 4 core - 177mm² - 1.4 billion

^ Please correct me if I have any of these wrong.

Note how the AMD's Piledriver is twice as big as an Ivy Bridge quad with the same transistor count, and Kaveri is around the same size as Intel's 6 core chips. Piledriver doesn't have an iGPU, and Kaveri's is in another league, but the iGPU that AMD adds uses a tremendous amount of die space and makes them very expensive to manufacture. AMD is severely disadvantaged by not having access to the latest process nodes.

Development costs aside, I'm certain Intel could build an i7 for less than AMD could make a 2 module chip, but it wouldn't be too long before they became irrelevant because they wouldn't have any money to put back into designing their next node and CPU design.

 

[ShintaiDK]

Its not about making profit, its about paying for R&D.

For an i5 to cost 100$, it would be because its only worth 100$ performance wise.

But its a pee in the pants solution. You get cheaper CPUs now, you get slower CPUs in the future. (Less R&D and the situation AMD is in.)

 

[frozentundra123456]

Its not about making profit, its about paying for R&D.

For an i5 to cost 100$, it would be because its only worth 100$ performance wise.

But its a pee in the pants solution. You get cheaper CPUs now, you get slower CPUs in the future. (Less R&D.)

Well it for sure is about making profit. That is the whole point of a business isn't it? I am not saying there is anything wrong with that though. Being a old timer and having lived through the days of 500 MHz chips costing hundreds of dollars, I think cpus from both makers are a great bargain right now. As for a hundred dollar i5 it is just not going to happen. If amd magically comes out with a more competitive line-up, they would undoubtedly raise prices to similar levels as Intel and actually make money in significant amounts for a change.

 

[john5220]

Interesting.. I think we should be seeing cheaper intel quad core CPUs

I am not too impressed with the current offerings from sandy bridge to haswell.

I was far more impressed with
Nehalem to Sandy Bridge performance Jump in CPU performance

Ever since then it has never been the same anymore.

 

[john5220]

I wonder if AMD K12 would be able to give intel competition?

AMD said they are coming back to high end CPU competition in 2016 or so if I remember correct.

 

[Yuriman]

I'm not holding my breath, but it would be nice. The cards are stacked against AMD. Intel's process node advantage alone makes me extremely skeptical.

I don't think we should be seeing cheaper quads unless Intel decided to build some without the iGPU component, and shed 60% of the die space used. Intel is not competing against AMD, really, but against their older chips, ARM (Qualcomm / Samsung), IBM, Apple, etc., and they need that revenue to stay ahead of the other giants.

Maybe the way for us to see $100 i5's would be for IBM, Apple, Qualcomm, and Samsung go out of business, and for Intel to cut back their R&D budget tremendously and settle for slower improvement of their CPUs. As it is, each generation is providing huge benefits to performance per watt, and is vastly improving performance AND battery life in mobile devices.

 

[Rezist]

It would straight up never happen, Intel is already slowing down obviously, and even if AMD could catch up in IPC/core, they'd lose badly at thermals and costs with the node disadvantage. I do hope AMD does step it up, because currently there is a $300 barrier on all PC gaming for a intel quad core k and Z series mobo. If you don't spend that your going to be making huge tradeoffs.

 

[john5220]

^ yeah I notice mobile has huge improvements.

200k, 3670K and 4790K or whatever you can nit

has little performance in crease. Back in the days you could get 100% performance increase. today its like 15%

I wonder if its because humans have reached all they can go in this kind of complex technology? BTW if AMD went out of business would this be good for consumers?
Some people claim it will but I doubt it.

 

[Yuriman]

I don't think it would be good, no. AMD is still providing alternatives that make sense in some contexts, and I appreciate a variety of choices.

It's probably worth mentioning that Intel's desktop CPUs are probably largely a result of their designing a core that will work well in both mobile devices and servers, where performance per watt is important. Those areas are where the real money is.

 

[postmortemIA]

ARM is the competition, not AMD.

 

[TheELF]

What would it take to bring down a i5 to $100?

What kind of processor would the competition need?

Is it possible for intel to sell a i5 for $100 and still make a profit?

Two ways for something like that to happen,just like intel switched from single core being the smallest cpu to dual core being the smallest,they could at some point switch to either option 1 all the lineup with Hyperthread or option 2 all the lineup being quads.
Option 1 seems quite doable and very cheap, I mean look at the pentium g3450 at 87$ and the i-1450 at 100mhz higher is just 13$ more,people would be jumping with joy for cellerons with hyperthreading at 10$ more.
They would cream any current amd quad at gaming,well even more so than now.

But 4 real cores at current i5 speeds for 100$ will never happen,at least not for a very very long time.

 

[frozentundra123456]

Undoubtedly, the pace of performance improvement in desktop cpus has slowed tremendously, but that is in part due to the focus on mobile and also just an unavoidable effect of the technology becoming mature. Performance improvements become increasingly difficult to achieve as the easily achieved improvements are exhausted.

Personally, I would love to see Intel come out with a cpu with the igp replaced with a couple more cores, but economically, I dont think it makes sense, because it would not be feasible to make a separate die for what would undoubtedly be a small segment of the market that adds a dgpu.

 

[Ketchup]

AMD is only providing alternatives in the budget sector for the most part. System builders like us can find AMD alternatives a little higher up, but that percentage doesn't amount to much, especially considering they don't have a whole lot to offer in our group, IMO.

Intel makes huge jumps when they need to. If AMD never gets up to that level, Intel does not have any incentive to cut prices on these chips.

 

[mrmt]

Undoubtedly, the pace of performance improvement in desktop cpus has slowed tremendously, but that is in part due to the focus on mobile and also just an unavoidable effect of the technology becoming mature. Performance improvements become increasingly difficult to achieve as the easily achieved improvements are exhausted.

With both notebooks and servers growing and desktop shrinking, why would Intel (or AMD for that matter) invest in better desktop processors?

 

[frozentundra123456]

With both notebooks and servers growing and desktop shrinking, why would Intel (or AMD for that matter) invest in better desktop processors?

Well, if they could achieve 20 to 40 percent increase per generation easily why would they not?

My point was that it probably is not that easy anymore, especially since die shrinks are becoming increasingly difficult (as for 14nm), and dont seem to automatically bring the performance increase they used to, and the architecture is mature as well.

 

[jhu]

It's already happened. Go check ebay.

 

[mrmt]

Well, if they could achieve 20 to 40 percent increase per generation easily why would they not?

Because they might not get their investment returned? Because there is no point in focusing in an stagnant market bracket when there are other that are more profitable and are growing? Because the overall PC market is paying big time for performance/watt, a feature that isn't make or break for the desktop market?

 

[john5220]

Do you guys think that if intel wanted, they could bring a 100% performance increase in Skylake? that is if they focused strictly on desktop market.

BTW in all honesty I will admit if I could get something like those MAC mini or mini ITX systems with the power to run all my current computing and gaming needs at max settings I would probably never upgrade for the next 20 years.

There are so many people who just play games like DOTA 2 and Starcraft series that never requires any sort of upgrade. And no matter how hard you try they will never play any game except those, they are after gameplay that is.

Like me I was gifted Battlefield 4 and Titanfall. I never touched it for more than 30 minutes.

My little time I have available to spare will just be sunk into some Black Ops 2, battlefield 3, and DayZ stand alone which is coming along nicely btw.

I was gifted so many games including all the metro that I never even cared to install.

I think I will keep my haswell system for the next 20 years. Maybe drop in a devils canyon used off ebay in about 10 years from now. I might get it for $80 by then. lol
All games being console ports also there is such little need to upgrade CPU anymore.

I think personally speaking games have evolved to the point where its always online nature allows for so much updates and longevity that many see no need to play anything else. I have sunk about 500 hours or so in counter strike GO so far and recently got black ops 2 how could I possibly upgrade?

sure I bought a i3 but that was for DayZ. And its about the same FPS as a i5 anyways. Then DayZ will come with more performance patches as it moves to a new Engine and goes into beta soon. Even less reason to change CPU.

 

[SteveGrabowski]

If the US went into a full-on 1929 style great depression again you'd probably see i5's selling for $100. Otherwise, a high end CPU for $100? Good luck with that.

 

[john5220]

interesting....

I wonder if an intel scientist / electrical engineer gets paid what they deserve for designing all these chips that intel makes billions off?

 

[SunburstLP]

John, you have to realize that servers and mobile are why Intel's innovating. We get desktop chips as a fruit of their labors. If Intel could double performance, they surely would. The market would bare the prices for the big iron and they would scale down that performance for mobile($) and desktop.

There's no conspiracy to be found on this front.

 

[skipsneeky2]

I think when the 6 cores become mainstream we will see some $100 quad cores but i bet like the quads will be gimped where you can't overclock them or they have a standard top clockspeed like a i3 that is non adjustable.

I think we could be a good 3 years out from a mainstream 6 core if more games are efficiently built around the newer 8 threaded consoles.

 

[Enigmoid]

As pressure from AMD? Never. AMD if they ever get their act together will try to price their CPUs higher than currently and make greater profits. Remember 8150 at $249? a10-7850k at a launch price of ~$180. The 9370 and 9590?

Intel would also never dilute their brand by pricing their mid range so low. You would see a radical renaming of their chips.