How much of a difference would there be in a "true" quad core compared to quad core?

[dario03]

I hear a lot of people keep saying that the q6600 isn't a true quad core but does that really matter? I know Penryn(sp?) is suppose to be true quad core but it will probably have other differences to. But if everything else was the same how much of a difference would a true quad make?

EDIT: For comparison with apps that can use 4 cores.

 

[meat_helmet]

Re: How much of a difference would there be in a "true" quad core compared to quad co

I hear a lot of people keep saying that the q6600 isn't a true quad core but does that really matter? I know Penryn(sp?) is suppose to be true quad core but it will probably have other differences to. But if everything else was the same how much of a difference would a true quad make?

Im guessing not actually that much. The problem is that there are few applications and no games that make use of 4 cores. I think even in games like crysis (which apparently has support for quads) we will only see maybe a 5% gain from using "true" quad cores.

 

[dario03]

Re: How much of a difference would there be in a "true" quad core compared to quad co

For this question assume that the apps can use 4 cores.

 

[peterhuang913]

Re: How much of a difference would there be in a "true" quad core compared to quad co

Im guessing not actually that much. The problem is that there are few applications and no games that make use of 4 cores. I think even in games like crysis (which apparently has support for quads) we will only see maybe a 5% gain from using "true" quad cores.

crysis will support four cores

 

[AP4LIFETN]

Re: How much of a difference would there be in a "true" quad core compared to quad co

probably temperature differences? and an increase in price?

 

[peterhuang913]

Re: How much of a difference would there be in a "true" quad core compared to quad co

btw dario, that phrase in your sig is weird

 

[catfacem3n]

Re: How much of a difference would there be in a "true" quad core compared to quad co

I was reading about this:

This situation is analygous to the comparison between the Pentium D and the Pentium core 2 duo: The Pentium D is like two Pentium 4's slapped together, while the core 2 duo is two cores on the same die.

The q6600 is like two core 2 duos slapped together, while a true quad core would be four cores on the same die.

How does this relate to perforamce? No empirical data, but I have read that on the q6600, stricly speaking, the four cores can not communicate with one another: a single core must commit its duo to communicate information to the other duo.

Just my analysis, I hope it sparks further conversation or debate.

 

[meat_helmet]

Re: How much of a difference would there be in a "true" quad core compared to quad co

I was reading about this:

This situation is analygous to the comparison between the Pentium D and the Pentium core 2 duo: The Pentium D is like two Pentium 4's slapped together, while the core 2 duo is two cores on the same die.

The q6600 is like two core 2 duos slapped together, while a true quad core would be four cores on the same die.

How does this relate to perforamce? No empirical data, but I have read that on the q6600, stricly speaking, the four cores can not communicate with one another: a single core must commit its duo to communicate information to the other duo.

Just my analysis, I hope it sparks further conversation or debate.

Just remember that the leap from the pentium D to the core 2 duo is huge whereas the leap to from the core 2 duo to penryn is not as big. The core 2 duos were a completely new architecture which got rid of the ancient netburst architecture.

Im starting to wonder how much performance increase there will really be with the penryns - only 18months after the core 2 duos, and even though they will be 45nm this might only translate into lower power usage and cooler processors.

 

[peterhuang913]

Re: How much of a difference would there be in a "true" quad core compared to quad co

Just remember that the leap from the pentium D to the core 2 duo is huge whereas the leap to from the core 2 duo to penryn is not as big. The core 2 duos were a completely new architecture which got rid of the ancient netburst architecture.

Im starting to wonder how much performance increase there will really be with the penryns - only 18months after the core 2 duos, and even though they will be 45nm this might only translate into lower power usage and cooler processors.

wrong, the c2d's are based off an existing architecture: the pentium 3 one.

 

[meat_helmet]

Re: How much of a difference would there be in a "true" quad core compared to quad co

wrong, the c2d's are based off an existing architecture: the pentium 3 one.

Seriously? I didnt know that... I guess my comment still sorta holds though.