LinuxQuestions.org
Download your favorite Linux distribution at LQ ISO.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware
User Name
Password
Linux - Hardware This forum is for Hardware issues.
Having trouble installing a piece of hardware? Want to know if that peripheral is compatible with Linux?

Notices


Reply
  Search this Thread
Old 07-16-2012, 05:54 PM   #1
nec207
Member
 
Registered: Apr 2011
Posts: 105

Rep: Reputation: 0
Unhappy Is it true that CPU from 2012 is 3 times faster than CPU from 2008?


Some was saying to me in e-mail that CPU from 2012 is 3 times faster than CPU from 2008?

If so that seems very low.I thought it would be 5 to 8 times faster.

If only 3 times faster from 2008 to 2012 that seems very very very slow and CPU are hitting a brick wall and in 5 to 10 years may come to stop with not being any faster at all.
 
Old 07-16-2012, 06:04 PM   #2
guyonearth
Member
 
Registered: Jun 2012
Location: USA
Distribution: Ubuntu
Posts: 419

Rep: Reputation: 82
Well, it would depend on the CPU. CPUs running at the same clock speeds certainly are not 5-8 times faster with the same number of cores, nor do I think they are even 3 times "faster". CPUs now are not running any "faster" than they were 10 years ago, in raw gigahertz, they are just much more parrallel, have faster buses, and more cores. There is an absolute limit to how much you can scale down and speed up silicon technology. The focus in recent years has been scaling down die size and reducing power requirements.
 
Old 07-16-2012, 06:36 PM   #3
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853
Didn't we had this speed issue in one of your previous threads? Without specifying the used CPU models and which benchmark was used this discussion doesn't make any sense, and this was already explained to you. Having said that, if you choose a 2012 CPU like the Atom N2600 and a 2012 CPU like the Core i7-3960X you will find many benchmarks that will show that the Core i7 is more than 3 times faster. If you compare the same Atom N2600 with the 2008 Phenom II X4 or the Core2Quad 9550 you will find that those older CPUs also are faster.

Re-read your old thread about CPU speed and benchmarks.
 
Old 07-16-2012, 06:50 PM   #4
nec207
Member
 
Registered: Apr 2011
Posts: 105

Original Poster
Rep: Reputation: 0
Quote:
Originally Posted by TobiSGD View Post
Didn't we had this speed issue in one of your previous threads? Without specifying the used CPU models and which benchmark was used this discussion doesn't make any sense, and this was already explained to you. Having said that, if you choose a 2012 CPU like the Atom N2600 and a 2012 CPU like the Core i7-3960X you will find many benchmarks that will show that the Core i7 is more than 3 times faster. If you compare the same Atom N2600 with the 2008 Phenom II X4 or the Core2Quad 9550 you will find that those older CPUs also are faster.

Re-read your old thread about CPU speed and benchmarks.

So that mesee if I understand what you saying. A 2012 CPU Atom N2600 and 2012 CPU Core i7-3960X it be 3 times faster. And Atom N2600 2012 to 2008 Phenom II X4 or the Core2Quad 9550 are also 3 times faster?

So some CPU from 2012 will be 3 times faster than other CPU from 2008 and some even more than 3 times faster? So some CPU from 2012 can be 5 to 7 times faster than other CPU from 2008.

Last edited by nec207; 07-16-2012 at 06:53 PM.
 
Old 07-16-2012, 07:19 PM   #5
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853
You didn't understand at all and that makes me wonder if you have understood the last thread you opened regarding a similar topic.

So let us recapitulate what we told you in the last thread first:

1. You can't say: This CPU is 3 times faster than that CPU because it totally depends on the Benchmark you use.
2. You can't say: A 2008 CPU is 3 times slower than a 2012 CPU without stating which CPUs exactly you compare and which benchmarks you use (see point one).
3. You can't say that there has to be a certain amount per year that CPUs have to be faster. This does not correspond with Intel's or AMD's development model.

So now let us recapitulate what I have said in my previous post:
1. If you compare the 2012 Atom N2600 ( a low end mobile CPU where low power consumption is a key factor) with the 2012 Core i7-3960X (a high-end CPU where power consumption plays no role at all) you will find that in some benchmarks the Core i7 will be more than 4 times faster than the Atom. This says nothing about past or future CPUs and their performance in general.
2. If you compare the 2012 Atom from above with a 2008 high-end CPU like the Phenom II 940 or the Core2Quad Q9550 you will find that the 2008 CPUs are faster than the 2012 CPU. That says nothing about past or future CPUs in general.
3. Let us add one, compare a 2008 Atom N270 with the 2012 Core i7-3960X and you will find that in some benchmarks the Core i7 will be about 10-12 times faster. Again, this says nothing about past and future CPUs in general.

So let us now go to the statements you made (or have gotten from an email):
1. The general statement that 2012 CPUs are 3 times faster than 2008 CPUs makes no sense without stating which CPUs you compare and which benchmarks you use for the comparison.
2. The general statement that CPUs are hitting a brickwall is not valid, because point 1 is a non-valid statement.
3. The general statement that CPUs may stop becoming faster in 5-10 years is illogical. Assumed CPUs would become faster by factor 3 in 4 years this would mean that we would have CPUs that are 3 times faster in 2016, 9 times faster in 2020, 27 times faster in 2024 and so on. You would have to compare with more and older CPUs to get knowledge about a general trend. If you would do that you would see that there are times where the CPU speed increases slower over time and there are times when the CPU speed increases faster over time (with sometimes even making more or less jumps, for example with the first and second generation of the Core2 technology).

Last edited by TobiSGD; 07-16-2012 at 07:22 PM.
 
Old 07-16-2012, 07:38 PM   #6
Roken
Member
 
Registered: Oct 2011
Location: Bolton, UK
Distribution: Arch local, Debian on VPS
Posts: 258

Rep: Reputation: 41
Moore's law is, by and large, still holding up, so generally speaking, A CPU created now using the equivalent technology and the equivalent resources that were available 3 years ago would, by and large, be 2x2 times faster and/or more efficient. It is, however, impossible to make a "like for like" comparison, since what is new and available now wasn't possible three years ago.

EDIT: By "equivalent" I mean suitably enhanced to allow for progression.
 
Old 07-16-2012, 07:44 PM   #7
nec207
Member
 
Registered: Apr 2011
Posts: 105

Original Poster
Rep: Reputation: 0
Quote:
Originally Posted by TobiSGD View Post
You didn't understand at all and that makes me wonder if you have understood the last thread you opened regarding a similar topic.

So let us recapitulate what we told you in the last thread first:

1. You can't say: This CPU is 3 times faster than that CPU because it totally depends on the Benchmark you use.
2. You can't say: A 2008 CPU is 3 times slower than a 2012 CPU without stating which CPUs exactly you compare and which benchmarks you use (see point one).
3. You can't say that there has to be a certain amount per year that CPUs have to be faster. This does not correspond with Intel's or AMD's development model.

So now let us recapitulate what I have said in my previous post:
1. If you compare the 2012 Atom N2600 ( a low end mobile CPU where low power consumption is a key factor) with the 2012 Core i7-3960X (a high-end CPU where power consumption plays no role at all) you will find that in some benchmarks the Core i7 will be more than 4 times faster than the Atom. This says nothing about past or future CPUs and their performance in general.
2. If you compare the 2012 Atom from above with a 2008 high-end CPU like the Phenom II 940 or the Core2Quad Q9550 you will find that the 2008 CPUs are faster than the 2012 CPU. That says nothing about past or future CPUs in general.
3. Let us add one, compare a 2008 Atom N270 with the 2012 Core i7-3960X and you will find that in some benchmarks the Core i7 will be about 10-12 times faster. Again, this says nothing about past and future CPUs in general.

So let us now go to the statements you made (or have gotten from an email):
1. The general statement that 2012 CPUs are 3 times faster than 2008 CPUs makes no sense without stating which CPUs you compare and which benchmarks you use for the comparison.
2. The general statement that CPUs are hitting a brickwall is not valid, because point 1 is a non-valid statement.
3. The general statement that CPUs may stop becoming faster in 5-10 years is illogical. Assumed CPUs would become faster by factor 3 in 4 years this would mean that we would have CPUs that are 3 times faster in 2016, 9 times faster in 2020, 27 times faster in 2024 and so on. You would have to compare with more and older CPUs to get knowledge about a general trend. If you would do that you would see that there are times where the CPU speed increases slower over time and there are times when the CPU speed increases faster over time (with sometimes even making more or less jumps, for example with the first and second generation of the Core2 technology).

So that me understand the reason I cannot get proper reply is there is too much at play with CPU . You saying some 2008 CPU are faster than some CPU today 2012 and some 2012 CPU are only 10% to 12% times faster and well other CPU may be 3 or 4 times faster .

And the reason people are saying this CPU is fast than this CPU is this fast is there is too much at play .

So some 2008 and 2009 CPU may be 2 or 3 times faster than some 2012 CPU and some 2012 CPU may be 10% faster to 2 times faster than other 2008 and 2009 CPU ?

If so why does intel and AMD do this? why is CPU not linear growth like 20% faster every year or 2 times faster every 4 or 5 years? Why does intel and AMD make it so mix match and so hard to understand.

Why is not linear growth with making better CPU every year?

Last edited by nec207; 07-16-2012 at 07:49 PM.
 
Old 07-16-2012, 07:59 PM   #8
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853
Quote:
Originally Posted by nec207 View Post
So that me understand the reason I cannot get proper reply is there is too much at play with CPU . You saying some 2008 CPU are faster than some CPU today 2012 and some 2012 CPU are only 10% to 12% times faster and well other CPU may be 3 or 4 times faster .
Not one time in my posts I mentioned a percentage. When I speak of 10-12 times faster then I mean 10-12 times, not 10-12%.

Quote:
If so why does intel and AMD do this? why is CPU not linear growth like 20% faster every year or 2 times faster every 4 or 5 years? Why does intel and AMD make it so mix match and so hard to understand.

Why is not linear growth with making better CPU every year?
Do you really think they do it intentionally? That they say: Hey, this new CPU would be 30% faster than the one from last year, we have to make it 10% slower? Or that they say: Hey, we are only up to 15% speed increase, we can't release this until we are at 20%?
This is how technology goes, they have to find ways to make their CPUs faster, they have to invent things and they have have to have a look at other factors as well, like power-consumption (the major drawback with the Netburst architecture of the Pentium 4, that was planned to go up to 10GHz clockspeed but hit the power-consumption wall at 4GHz and therefor had to be replaced with a different architecture). A big part of this is not only to create new CPUs, they also have to come up with ways to actually build those CPUs. The structures on the CPU dies have to get smaller to be able to become more complex within the same size (to hold the prices), this is also not an easy thing. Producing new CPUs is much more complex as you seem to think.

Last edited by TobiSGD; 07-16-2012 at 08:01 PM.
 
Old 07-16-2012, 08:01 PM   #9
Roken
Member
 
Registered: Oct 2011
Location: Bolton, UK
Distribution: Arch local, Debian on VPS
Posts: 258

Rep: Reputation: 41
All CPUs from the major players are built using an unreliable manufacturing method. It's probable that two CPUs, built from the same silicon, on the same production line within seconds of each other will exhibit different characteristics and tolerances.

The manufacturer goes on to test the CPUs, and they are graded according to those tests. One of them may be top grade, passing all tests at the highest frequencies in which case they will be sold as such, performing better than lesser models. Some may fail tests at the highest frequencies, and as such will be sold only at the level of reliability that the tests prove them to have.

It is highly probable in such a scenario that even of 2nd and 3rd generation models, some will test with inferior specs to the preceding model, and therefore be outperformed by the best of the earlier generation.

The manufacturer is in the business of making money, and as such they aren't simply going to junk the lower graded components. Instead, they will sell then, at a reduced specification and price as such.

It is possible, owing to the laws of supply and demand, for high grade products to be rebranded as lower grade, which is why you can buy a (relatively) inexpensive model that will perform as well as the highest grade. It's simply because it IS the highest grade, but not enough lower grade appeared to fulfil demand.
 
Old 07-16-2012, 08:19 PM   #10
nec207
Member
 
Registered: Apr 2011
Posts: 105

Original Poster
Rep: Reputation: 0
Quote:
Originally Posted by TobiSGD View Post
Not one time in my posts I mentioned a percentage. When I speak of 10-12 times faster then I mean 10-12 times, not 10-12%.

Do you really think they do it intentionally? That they say: Hey, this new CPU would be 30% faster than the one from last year, we have to make it 10% slower? Or that they say: Hey, we are only up to 15% speed increase, we can't release this until we are at 20%?
This is how technology goes, they have to find ways to make their CPUs faster, they have to invent things and they have have to have a look at other factors as well, like power-consumption (the major drawback with the Netburst architecture of the Pentium 4, that was planned to go up to 10GHz clockspeed but hit the power-consumption wall at 4GHz and therefor had to be replaced with a different architecture). A big part of this is not only to create new CPUs, they also have to come up with ways to actually build those CPUs. The structures on the CPU dies have to get smaller to be able to become more complex within the same size (to hold the prices), this is also not an easy thing. Producing new CPUs is much more complex as you seem to think.
Yes but you said Quote You can't say that there has to be a certain amount per year that CPUs have to be faster. This does not correspond with Intel's or AMD's development model.
Quote


So that means you cannot say CPU is 20% faster every year or 2 times faster every 4 to 5 years or I would got simple answer with same answer so in other words there too much at play that CPU from 2008 to 2012 may be anywhere from 10% faster to 5 times faster to other times 2012 CPU being slower.In other words intel or AMD does not make linear growth it is mix match some being slower and some faster and all over the place some 10% faster and some 4 times faster so avarage person on the street are going to understand .

So some 2008 and 2009 CPU may be 2 to 3 times faster than some 2012 CPU and some 2012 CPU may be 10% faster or 3 times faster thus reply Quote You can't say that there has to be a certain amount per year that CPUs have to be faster. This does not correspond with Intel's or AMD's development model.
Quote

If so I understand than this thread is meaningless than with out comparing the two CPU models has it does work that way that every year CPU is 20% faster or 2 or 3 times faster every 4 years has it is mix match of speed at all levels some new CPU faster and some new CPU slower

Last edited by nec207; 07-16-2012 at 08:31 PM.
 
Old 07-16-2012, 08:25 PM   #11
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853
Quote:
Originally Posted by nec207 View Post
If so I understand than this thread is meaningless
Now you got it.
 
Old 07-16-2012, 08:40 PM   #12
273
LQ Addict
 
Registered: Dec 2011
Location: UK
Distribution: Debian Sid AMD64, Raspbian Wheezy, various VMs
Posts: 6,910

Rep: Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999
Core speed, yes CPUs aren't much faster now.
It does seem that the maximum speed has been approached for the processors we own, though maybe lower voltage will out.
I recall your posts though, nec207, and I'm not sure I ever understood your question.
I know the CPU and GPU I bought last week both outshine their predecessors in both efficiency and results.
 
Old 07-16-2012, 08:42 PM   #13
nec207
Member
 
Registered: Apr 2011
Posts: 105

Original Poster
Rep: Reputation: 0
Quote:
Originally Posted by TobiSGD View Post
Now you got it.
So why does intel and AMD bring out CPU every year at different levels and not linear growth ? Even if there is engineering problems that they cannot be 40% faster every year than some years if there is engineering problems it may be 20% faster than .

Why does intel and AMD bring out CPU every year at different levels . Tell a person on street better do you home work warning this CPU 2012 may be 20% faster than 2008 CPU and other 2012 CPU 3 times faster than 2008 CPU and if that not enough do your home work some 2008 CPU are faster than some 2012 CPU.

Than there is a misconception with public that CPU speed is linear growth.

Last edited by nec207; 07-16-2012 at 08:46 PM.
 
Old 07-16-2012, 08:54 PM   #14
273
LQ Addict
 
Registered: Dec 2011
Location: UK
Distribution: Debian Sid AMD64, Raspbian Wheezy, various VMs
Posts: 6,910

Rep: Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999Reputation: 1999
I bought a CPU, motherboard, RAM, and GPU the other day. This machine is faster, cooler (temperature), more power efficient and quieter.
Newer CPUs tailor the instruction sets and make more efficient use of what they have.
The GPU and GPU in this thing are just ordinary parts but such that I could run my old system four times on this system and not slow down.
Quote:
Originally Posted by 273 View Post
I recall your posts though, nec207, and I'm not sure I ever understood your question.
 
Old 07-16-2012, 08:56 PM   #15
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Germany
Distribution: Whatever fits the task best
Posts: 17,148
Blog Entries: 2

Rep: Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853Reputation: 4853
Quote:
Originally Posted by nec207 View Post
So why does intel and AMD bring out CPU every year at different levels and not linear growth ?
Because they can't do it with linear growth.

Quote:
Even if there is engineering problems that they cannot be 40% faster every year than some years if there is engineering problems it may be 20% faster than .
And that is the exact opposite of linear, so actually you answered your question yourself.

Quote:
Why does intel and AMD bring out CPU every year at different levels
As said above, because it is impossible to do it in a linear way.

Quote:
Tell a person on street better do you home work warning this CPU 2012 may be 20% faster than 2008 CPU and other 2012 CPU 3 times faster than 2008 and if that not enough do your home work some 2008 are faster than some 2012 CPU.
Even the person from the street would not expect a 2012 CPU for 30$ to be 3 times faster than a 2008 CPU for 300$. And they don't do it. People that are not interested to inform themselves about CPU speed have a different approach to buy computers. They just expect the 500$ PC they buy now to be significant faster than the 500$ PC they bought in 2008. They are not interested in absolute numbers.

Quote:
Than there is a misconception with public that CPU speed is linear growth.
I would rather think that this is not a public misconception, it is your misconception. People are not interested in absolute numbers, they don't care if there is a linear growth, they expect computers to be fast enough to do there every day task in a reasonable way. Not one "average computer buyer" goes to a store and expects the new computer to be exactly 3 times faster than the old one. They just expect that it is fast enough for the tasks they want to do with their 2012 software, in the same way the 2008 computer was fast enough to do that with 2008 software.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
true cpu clockspeed shell script anomic Linux - Software 14 07-13-2012 01:14 PM
Differentiating a true CPU core vs a Hyper-Threaded one? koyi Programming 2 12-03-2011 11:13 AM
[SOLVED] Slower CPU with Faster HDD vs. Faster CPU with Slower HDD ? for a web server WhisperiN Linux - Server 4 11-03-2011 07:42 PM
Slow SMB vs. much faster single CPU kstarks Linux - Hardware 0 02-03-2005 05:06 PM
Detect my true CPU rating (NOT the current speed) qwijibow Linux - Hardware 5 11-16-2004 09:53 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware

All times are GMT -5. The time now is 01:43 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration