CPU Benchmarks on Anandtech.

[iwodo]

I brought my Pentium M Dothan Laptop nearly 4 years ago. Looking at that graph.

Buying a TOP of the Line DESKTOP CPU today will only give me less then 4 times the performance of a Mid Range Laptop CPU.

This is like buying a $1000 System today that is only 4x faster then a $500 System 4 years ago.

2x the price, 4x the performance, 4 years later.

That doesn't sound impressive to me at all.

 

[Andrew1990]

Well performance per core hasnt really gone up much. As of now it seems Intel and AMD are just glueing more and more cores together to get more "performance"....

Also, those Pentium M Dothans were nice CPUs. I had a desktop board with one in it and it ran all my games fine with a 9600GSO.

 

[Flipped Gazelle]

Originally posted by: iwodo
I brought my Pentium M Dothan Laptop nearly 4 years ago. Looking at that graph.

Buying a TOP of the Line DESKTOP CPU today will only give me less then 4 times the performance of a Mid Range Laptop CPU.

This is like buying a $1000 System today that is only 4x faster then a $500 System 4 years ago.

2x the price, 4x the performance, 4 years later.

That doesn't sound impressive to me at all.

I guess you're not familiar with Moore's Law...

 

[Idontcare]

Originally posted by: iwodo
2x the price, 4x the performance, 4 years later.

That doesn't sound impressive to me at all.

Yikes, talk about having high expectations.

What kind of performance increase or cost decrease would it take in order for you to be impressed?

 

[Zenoth]

Also we could ask if we, general public and gamers actually need a 6.0Ghz CPU for instance, do we? I don't think so. We should get better "performance" out of multi-Cores working together especially if the application is built with that in mind. Just working with Photoshop CS3 (for me) on my previous AMD X2 4400 brought an immense difference than when I was using my single-Core AMD 64 3500+, although I had the 3500+ OC'ed to 2.9Ghz. If I'd do the same work I still do today on a Core i7 I'm sure I'd see very tangible differences than with my current C2D E8400.

To me it is clear that in the future we will benefit more from multi-Cores clocked at or around 3.0Ghz than if we were still all stuck with dual-Cores even if clocked at or around 5.0Ghz. And it's not only because both AMD and Intel decided to take that road and that we pretty much have no choice in following them, but because it actually makes sense, at least in my opinion. In my book the "only up to 2x performance increase" we're seeing each years or two is completely fine, and the more Cores we have operating more efficiently together and the better it will be, and we always have over-clocking to make things slightly or tangibly faster than they should be.

 

[Jabbernyx]

Originally posted by: Idontcare
Yikes, talk about having high expectations.
What kind of performance increase or cost decrease would it take in order for you to be impressed?

As an engineer, I'd say all the various achievements that went into making the current CPUs (fab tech, u-architecture, CAD tool improvements (timing closure, floorplanning, routing), etc.) already impresses the hell out of me.
But I guess there's no pleasing the consumer until you can get an i7 that clocks 300% for $50 😛

 

[nyker96]

well when you say 4x performance, you have to specify how to measure it. in game? in app? composite benchmarks? i don;t know but it seems quite vague here to compare directly. pentium m is also made for a different purpose so you can't make direct comparisons of apples and oranges. anyways i think moore's law is prediction of trasnsitor count, so let's see how many in each:
pentium m has 140 million, 731 million for i7, so I'd say roughly double every 2 years here. about right.

 

[iwodo]

May be it is just those marketing that keep makes me think CPU has gone a lot faster.

On Benchmarks - Core i7 Quad Core 2.93 Ghz is about 4 times faster then Pentium M 1.8Ghz
That is 3 more core, 1 Ghz faster, all other tech improvment.... etc

Now i know this is totally unscientific. But a general user would expect, 3 more core and 1 Ghz faster and 4 years of technology advancement bah bah would give you much more then that. Oh not to mention Core i7 cost twice as much........ so in reality, if you scale down we are paying same price for double the performance after 4 years.

And that is on benchmarks....... even worst is the user asking, why is everything the same.....
With more memory, super fast CPU, faster HDD, we are.......
Still waiting 30+ sec for bootup and shutdown.
Flash still consume most of our CPU usage when we surf.
She was expecting everything to zip through...... and she asked a question....... we uses DOS in our old office, why do 10 - 20 years old DOS program open instantly while today we still have to wait for apps to load? She thought her office computer was too old.... She brought a new $1000+ USD PC and expect everything to be faster....

 

[Zstream]

Originally posted by: iwodo
May be it is just those marketing that keep makes me think CPU has gone a lot faster.

On Benchmarks - Core i7 Quad Core 2.93 Ghz is about 4 times faster then Pentium M 1.8Ghz
That is 3 more core, 1 Ghz faster, all other tech improvment.... etc

Now i know this is totally unscientific. But a general user would expect, 3 more core and 1 Ghz faster and 4 years of technology advancement bah bah would give you much more then that. Oh not to mention Core i7 cost twice as much........ so in reality, if you scale down we are paying same price for double the performance after 4 years.

And that is on benchmarks....... even worst is the user asking, why is everything the same.....
With more memory, super fast CPU, faster HDD, we are.......
Still waiting 30+ sec for bootup and shutdown.
Flash still consume most of our CPU usage when we surf.
She was expecting everything to zip through...... and she asked a question....... we uses DOS in our old office, why do 10 - 20 years old DOS program open instantly while today we still have to wait for apps to load? She thought her office computer was too old.... She brought a new $1000+ USD PC and expect everything to be faster....

I'll take two!

 

[Flipped Gazelle]

Hard drives have not improved in terms of speed at anywhere near the rate that CPU's have. 4 years ago, 7200 rpm drives were standard. Now, 7200 rpm drives are standard.

Windows does a bunch more than DOS. A DOS program is much smaller than it's Windows counterpart.

I rarely see Flash use more than 35% CPU time on any of my PC's.

Of course, when you buy a new PC it is usually so loaded with crap that everything gets bogged down.

Look at it this way: a CPU is like a car's engine. The engine is putting out more horsepower than it did 10-20 years ago, but the body is much heavier. All the background processes that Windows has to run, DOS did not.

 

[Idontcare]

Originally posted by: Flipped Gazelle
Look at it this way: a CPU is like a car's engine. The engine is putting out more horsepower than it did 10-20 years ago, but the body is much heavier. All the background processes that Windows has to run, DOS did not.

Great analogy.

And costs have increased in part because of the addition of secondary items such as safety features (standard airbags, anti-lock brakes, air emissions, etc) which have little to do with improving the performance of a one-dimensional metric of assessment such as torque or HP or mpg.

 

[cusideabelincoln]

Glass half full, half empty.

For me, there is a distinct difference in just everyday normal computing (web browsing) between a single core Athlon XP and a dual core Athlon X2. So frankly, I am impressed with the performance improvements over the time span of four years.

Oh and thanks for letting us know what benchmarks you are even referring to.

 

[blackpanther26]

i think CPU's came a long way sence Athlon XP and Pentuim III.

 

[dflynchimp]

take into account inflation, and you should be more than happy with what you're getting on the dollar. Besides, performance scaling and price gauging versus time has never been in the same graph...

 

[Idontcare]

Originally posted by: dflynchimp
take into account inflation, and you should be more than happy with what you're getting on the dollar. Besides, performance scaling and price gauging versus time has never been in the same graph...

You may not realize it but there is some irony to what you posted, see the first graph (on second page) of Moore's original paper from 1965.

<a target=_blank class=ftalternatingbarlinklarge href="ftp://download.intel.com/research/silicon/moorespaper.pdf">ftp://download.intel.com/re......on/moorespaper.pdf</a>

Moore's law is based on the number of components per integrated circuit which minimizes manufacturing costs. The minimum found in lines from the first graph on page two is the value that goes into the second graph on page three.

Performance is actually not a metric of Moore's law, but the implication that performance increases commensurately with increasing components per integrated circuit is what drives the association of performance doubling every 2 yrs (more or less) to Moore's law.

But the truth of it is, if you read Moore's paper, that we see Moore's law already predicts the end of Moore's law. Again referring to the first graph, what we can expect these lines to look like as we move forward in time is that the spacing between the lines for each successive year will get less and less (the rate of cost reduction is decreasing because production costs are rising faster with each node) and the curvature of the line will get more and more flatter as the fixed costs (R&D for the process tech, mask sets for each device, fab setup costs) of producing an IC in an advanced node are beginning to dwarf the total sales volume generated by the IC itself.

The effect this has on the data point that goes into the second graph on page three (the graph we typically think of when we think of Moore's Law) is that the error bars on the y-axis for the data points get really large to the downside (because we don't have good curvature to easily define the number of components per IC that bring about a minimum in the relative manufacturing cost in the first graph on the second page.

And as you can see the shallow tail on the first graph favors the low-component count side, meaning the y-axis error bars on the second graph will be longer to the downside and shorter to the upside, meaning the line on Moore's graph is expected to bend-over and flatten out horizontally as the x-axis marches onwards from left to right.

It's all there, since 1965, cost basis and all.