AMD's 3GHz K10 to break 30,000 3DMark06

[Nagasama]

hey, intel might be topping the charts right now, but think of it like an argument between cars. two guys, one has a classic mustang, and one has a classic corvette. both are gonna be upgrading and improving until theirs is better then they will hold off and gloat over the victory until the other takes first, then the battle starts over......

its been this way for a while...intel, then amd, then intel, then amd, etc....back and fourth tug of war, its just a freaking battle that leaves us, the enthusiast consumer begging the companies to fight more violently and create some serious performance, like the c2d when then k8 was out, now we are going to have the k10, then they will have the c3d, then k12, etc....down the line untill we have a k64 and a c29o (a 8 core cpu, maybe? or how about a 64bit 64 core 64ghz cpu...oohhhh the possibilities!) only the future will tell. lets just hold off until the k10 rolls of the line and we can try it ourselves, i would buy one just to settle the argument even if i had to pull 3 jobs in one day, plus night shifts all while figuring out how to take care of my schooling and stuff.

awesome! great analogy bmxfreakrider, wo0t!
p.s. i hate math.

 

[Apokalipse]

Yes the the difference should be somewhere 10 to 20%

I just gave you two examples in my previous post

Have you read them


A + B + C + D + F = X


If I increase A, B, C, D and F by 12%, 15%, 18%, 9% and 20% respectively. Then the percentage increase on X would be a number between 9 and 20

Faulty assumptions. Simple as that.
you can't just assume what you want to prove in the first place.

No it is not

You need to learn more math

I took Mathematical Methods and Specialist Maths, which were the two highest level classes for high school here.

Now, back up your assumptions.

It's a really simple concept. If I increase the horsepower of the engine by 20%, and put in a gearbox that's 10% more efficient, then my performance increase is going to be more than 20%

When increasing clock speeds, you don't just get linear increases in performance.
Video card performance is dependent on CPU performance.
Just like the power the gearbox transfers to the wheels is dependent on engine power.

 

[maroon1]

It's a really simple concept. If I increase the horsepower of the engine by 20%, and put in a gearbox that's 10% more efficient, then my performance increase is going to be more than 20%

Can you show how ??

Faulty assumptions. Simple as that.
you can't just assume what you want to prove in the first place.

I just did an example in my previous, right ?

Can show what is wrong with that ?

I will give you an example

100 + 200 = 300

You increase the number 100 and 200 by 10% then you get 110 and 220, right ?

110 + 220 = 330

330 is only 10% more than 300


........


Another example:

100 + 200 + 300 = 600

Let us increase the numbers 100, 200 and 300 by 10%, 5% and 20% respectively ...... you will get 110, 210 and 360

110 + 210 + 360 = 680

680 is only 13.3% more than 600

 

[Apokalipse]

Can you show how ??



I just did an example in my previous, right ?

Can show what is wrong with that ?

yes, they are assuming linear transformations of performance, and is assuming that the video card and CPU performance are independent from each other.
Not only that, but they're not taking into account scalability (which varies with clock speeds, memory speeds, bus speeds, types of applications running etc)

In fact, video card performance is dependant on CPU performance.
Secondly, increases in frequency don't give linear increases in performance. Some products scale better than others.

I'd imagine that PCI-E 2.0 and HTT 3.0 allow much better scaling on the K10 platform that was tested, than on K8 or Core 2 platforms.

Now, you're ignoring my car analogy.
A CPU is like an engine, and a video card is like a gearbox.

The gearbox (video card) can't do anything without the engine (CPU) giving it data/power

 

[Nagasama]

really doesnt it matter how much horsepower influences performance...
does 20% more horsepower give a 20% performance increase (ie will top speed go from 100 mph to 120 mph)? not neccessarily, i dont think.
i can see both sides of this argument...
from apokalipse's side, yes...it would seem that a 20 + 10 % increase would total out to more than 20% overall.
from maroon's side, his math figures look correct (from a math idiot's (me) viewpoint).

if you take a car that has 200 horsepower and give it 240 horsepower (a 20% increase in horsepower)...i dont know if it would actually increase the PERFORMANCE by 20%, it would only increase the actual horsepower by that much. would depend on several factors..wieght, fuel, driver, tires etc...

 

[Apokalipse]

from maroon's side, his math figures look correct (from a math idiot's (me) viewpoint).

If his assumptions were correct, they would be.

 

[mrfluffles]

lol the intel boys who are suh-wettin.
dude you are fools for thinking that amd has no answer to c2d.
foolish i say!

You are fools for thinking any benchmark where AMD wins must be real. Phenom could be good, but this article is fake, and that's fact.

 

[Nagasama]

really, with cars lol...i think that a 20% increase in horsepower at the top end would result in MORE than a 20% increase (if we equate performance with speed and or torque/starting line power).
i dont know math, but i know cars lol.

 

[Nagasama]

You are fools for thinking any benchmark where AMD wins must be real. Phenom could be good, but this article is fake, and that's fact.

lol @ u fluffles.
amd wins cinebech ALL THE TIME. it wins memory bandwidth bechmarks ALL THE TIME. and THATS a fact, nothing personal. Lmao.
mark my words...if phenom is that good ya'll will cry like little girls (right before you go buy one). it wont beat c2d in ALL benchmarks i dont think. but it is going to put AMD in the faster OVERALL bracket, which is where intel sits now...its not faster in everything, but it is faster for the most part.

 

[Apokalipse]

really, with cars lol...i think that a 20% increase in horsepower at the top end would result in MORE than a 20% increase (if we equate performance with speed and or torque/starting line power).
i dont know math, but i know cars lol.

I wouldn't say it would necessarily get 20% increase in performance with 20% extra horsepower.
With more torque, there's more friction, so some of the extra horsepower is lost.

However, if you add a more efficient gearbox on top of the extra 20% horsepower - let's say the gearbox is 10% more efficient. You might get close to 30% extra performance.

It probably won't be 30% exactly, but it'll be more than 20%. That's the point.