Why aren't processors getting faster?

[Poledancerz]

I made a quick Google on this question, read some articles, but I'm interested in starting something here. 

 

I've been looking for a new processor, to upgrade from this i7-2600 @ 3.4GHz. I've had this processor for around 5 years now (... or something), I think it's time for an upgrade. 

I go check for some newer intel processors, and it seems like they really haven't improved that much (for the $300 range at least). There are a few processors with more cores that cost around a buttload of money, and slightly higher clock speeds. 

I didn't spend too much time comparing (it's bedtime =p), but I didn't really notice many differences between newer processors. 

Can someone explain to me why a newer processor, for example the i7-4790k. It's at 4.0GHz, which is one advantage I guess. 

 

But is it even worth upgrading my processor? 

What are the perks of getting a newer processor? (in the reasonable price range ($300))

... I'm going to sleep, I'll read the replies in the morning. 

 

Thanks for your replies, I'm still new learning this stuff. 

[TheKDub]

Processors have gotten alot faster even though the clock speed has not significantly increased.

 

What is it you use your computer for? unless you are doing stuff like large amounts of video editing or multi-tasking then you would be fine with a i5 processor.

[gundershot]

Intel is so far ahead of AMD they can get away with marginal improvements. Also clock speed is no longer a major factor in how effective a CPU is.

[eLucid]

CPU architecture is far more important than clock speed. Processors are faster these days.

[LurzKesh1138]

I think it's because after awhile clock speed delivers diminishing returns for the power required. That and I believe manufacturers had to wait for RAM speeds to catch up and become stable. 

[Builder]

They are getting faster. As evidenced by Broadwell, we're coming up to the limits of how small silicon parts can get. 

[Glenwing]

Intel has not been focusing on performance lately, just reduced power. Mobile devices are an ever-growing market, and Intel does not have a strong presence there because their platform is not as power efficient as ARM chips used in most phones and tablets currently. Intel is trying to change that.

You are correct there is not much better than the i7-2600 and that generation, unless you are looking at motherboards and features and things like that outside of pure performance. I'm still running a 2500K and an i5-2520M laptop myself. Nothing has really improved since Sandy Bridge generation except power consumption, what you've got now is right up there with the best of them. There's no reason to upgrade unless you want a more modern motherboard.

As other posters have noted, core frequency by itself is not an indicator of performance between generation, the design of the cores (micro-achitecture) is what really matters, but in either case performance has not been improving noticeably, neither in clock frequencies or architecture. Not sure if you were aware of that, or just meant that clock frequencies haven't improved and were accidentally right that performance hasn't gone anywhere.

I don't know what everyone else is talking about when they say current chips are much faster than 2nd generation chips. No, they aren't. Anyone can do a Google image search of "Sandy Bridge vs. Haswell" and browse through benchmark after benchmark of nearly identical performance, keeping in mind that the clock frequencies do get set 100MHz higher each generation too, which makes the true difference even slimmer than what the benchmarks show.

[rmac52]

IPC is just as important as clock speed.  The new Intel chips are a lot faster that the sandy bridge era chips.  Probably still not worth upgrading your cpu though.  If you are gaming then the GPU is far more important to gaming.

[StingBull]

Do not upgrade. There is no point for most tasks. Especially for gaming CPU is largely irrelevant. Upgrade your GPU instead or get SSDs if you don't have them.

[dlf]

Also ain't it something to do with how hot the CPU (rather the silicon of the CPU) can get. What we currently use - silicon can only stand so much heat, while something like graphene can hypo (or theoretically) stand far far far more gigahertz then silicon while maintaining a cool (no pun).

[TheRagingGamer]

Intel is so far ahead of AMD they can get away with marginal improvements. Also clock speed is no longer a major factor in how effective a CPU is.

that's the problem, amd has to do something. It kills me that they are making apus when that time could be spend on beating intel like they have dont for the past 10 years 

[Lord Pantaloons]

Intel's newer chips are faster, clockspeed is not comparable between generations nor is it always a true representation of performance. 4.0Ghz on Sandy is slower than 4.0Ghz Ivy, 4.0Ghz on Ivy is slower than 4.0Ghz on Haswell/DC. Whether to upgrade or not is more of a personal opinion, some will tell you to upgrade solely based on the fact that your chip is a few years old despite the fact that it's completely fine performance wise even when compared to the newer chips that are out while others would tell you that it's no necessary. First gen i5/i7's are still strong so there's no real reason to upgrade solely based on the performance of the newer chips.

[Guest]

They are. Dramatically.

 

Clockspeeds have hit a wall, because chip manufacturers can't just add more power (and therefore crank up clockspeeds) with new chips and call it a day. Cores, cache, and other factors are what make it faster- first-generation Pentium chips could go up to 3.8GHz, but a new 1.7GHz MacBook Air would crush it in every respect.

[rodion.zissou]

What material will we move to after Silicon? Graphene? Hemp?

[niofalpha]

Because their's no need too. One of the reasons its not continuing further is that Silicon is nearing its end of being a Processor material. It would just cost too much to move on too a new Material (Carbon Nanotubes, Graphene (I know its a super conductor, not really a Processing chip.),  E.t.c.). 

 

There is really no need on making them "Faster" In aspects of GHz. Look at the current CPU market too see evidence.  Intel CPU's that go over a GHz slower than an AMD CPU Outperform them. The speed means nothing, as long as it's above a decent amount.

[mr moose]

Intel is so far ahead of AMD they can get away with marginal improvements. Also clock speed is no longer a major factor in how effective a CPU is.

 

 

that's the problem, amd has to do something. It kills me that they are making apus when that time could be spend on beating intel like they have dont for the past 10 years 

 

I tend to feel that Intel might have actually hit a wall.  It is getting increasingly harder for them with the limitations of manufacturing lithography.   Either way though, it is sad AMD haven't been able to keep up.  Competition always helps.

[TheRagingGamer]

I tend to feel that Intel might have actually hit a wall.  It is getting increasingly harder for them with the limitations of manufacturing lithography.   Either way though, it is sad AMD haven't been able to keep up.  Competition always helps.

amd has not made a new cpu in 2 years, they are not even trying. They are just overclocking old ones and selling it as new stuff 

[mr moose]

amd has not made a new cpu in 2 years, they are not even trying. They are just overclocking old ones and selling it as new stuff 

 

Oh they are trying, they just don't have the money to invest. They have been severely hamstrung over the last decade with no profits.  On top of that even if they did have the cash to inject into R+D how are they going to go one better than intel?  especially if intel's improvements are hamstrung by manufacturing and silicon limitations?  I just don't think it was complacency that landed AMD where they are,  In retrospect they have made a few bad choices that didn't help, but at the time I am sure they would have been the best option.

[TheRagingGamer]

Oh they are trying, they just don't have the money to invest. They have been severely hamstrung over the last decade with no profits.  On top of that even if they did have the cash to inject into R+D how are they going to go one better than intel?  especially if intel's improvements are hamstrung by manufacturing and silicon limitations?  I just don't think it was complacency that landed AMD where they are,  In retrospect they have made a few bad choices that didn't help, but at the time I am sure they would have been the best option.

well its not like that can barley survive google: amd 

their stocks are rising, and i don't see why can't they beat intel, they beat nvidia with 295x2 and the made the first 2 core cpu as well as the first one to hit 1ghz. If AMD gets their shit together they can probably make a good cpu. 

And they are also making apu's for potatoes and also making ssd's so i don't see why can't they make some money 

But knowing amd their new cpu will probably be something like 16 cores or something stupid like that 

[mr moose]

well its not like that can barley survive google: amd 

their stocks are rising, and i don't see why can't they beat intel, they beat nvidia with 295x2 and the made the first 2 core cpu as well as the first one to hit 1ghz. If AMD gets their shit together they can probably make a good cpu. 

And they are also making apu's for potatoes and also making ssd's so i don't see why can't they make some money 

But knowing amd their new cpu will probably be something like 16 cores or something stupid like that 

 

Don't get me wrong, AMD have been really putting out some good stuff on GPU front. Given their financials they have managed stellar results.  But the issue is they have had a decade with virtually no profits.  Intel have been investing Billions into R+D while AMD is in the hundreds of millions.  This discrepancy is hard to match and presents a very large challenge for AMD.    I think they are turning out the ssd's and APUs as a way of maintaining revenue and thus in return it gives them a better R+D budget.

 

http://ycharts.com/companies/AMD/r_and_d_expense

https://ycharts.com/companies/INTC/r_and_d_expense

[TheRagingGamer]

Don't get me wrong, AMD have been really putting out some good stuff on GPU front. Given their financials they have managed stellar results.  But the issue is they have had a decade with virtually no profits.  Intel have been investing Billions into R+D while AMD is in the hundreds of millions.  This discrepancy is hard to match and presents a very large challenge for AMD.    I think they are turning out the ssd's and APUs as a way of maintaining revenue and thus in return it gives them a better R+D budget.

 

http://ycharts.com/companies/AMD/r_and_d_expense

https://ycharts.com/companies/INTC/r_and_d_expense

Well if i were someone at amd id start selling gpu's that are made by amd and spend more money on advertising and focus on either gpu's or cpu's. you can't have everything 

[mr moose]

Well if i were someone at amd id start selling gpu's that are made by amd and spend more money on advertising and focus on either gpu's or cpu's. you can't have everything 

 

To be honest I don't know what they need to do to get back in front,  I can't help but wonder sometimes if the divestiture of global Foundries was their best move.  I know they were up against a wall however one of intels strong points is it owns all it's own manufacturing. 

[TheRagingGamer]

To be honest I don't know what they need to do to get back in front,  I can't help but wonder sometimes if the divestiture of global Foundries was their best move.  I know they were up against a wall however one of intels strong points is it owns all it's own manufacturing. 

i wonder how would a fx 9xxx cpu perform but with multithreading, provided that is even possible 

[Legion]

http://en.wikipedia.org/wiki/Moore%27s_law

 

All you need to know

[mr moose]

http://en.wikipedia.org/wiki/Moore%27s_law

 

All you need to know

 

But it's been used as a guide for roadmaps as much as observation of technological advance.  In other words it won't hold true forever (in fact we may be very close to the end unless an alternative to silicon can take over) and it may only be true because the industry has elected to advance at that speed.