AppleInsider › Forums › Mac Hardware › Future Apple Hardware › Intel Core 2 Quad on MBP and iMac.
New Posts  All Forums:Forum Nav:

Intel Core 2 Quad on MBP and iMac. - Page 3

post #81 of 86
6/25/09 at 10:25am
Quote:
Originally Posted by mcarling View Post

What some people here seem to overlook is that a 32nm Nehalem processor with any number of cores will necessarily and dramatically outperform a 45nm Nehalem processor with any number of cores if the clock rates are set such that the TDP is the same.

So are you saying that a Microprocessor from the same family running at the same frequency will yield a performance increase strictly from a processes shrink?

I've never a benchmark that supports this.
He's a mod so he has a few extra vBulletin privileges. That doesn't mean he should stop posting or should start acting like Digital Jesus.
- SolipsismX
Reply
He's a mod so he has a few extra vBulletin privileges. That doesn't mean he should stop posting or should start acting like Digital Jesus.
- SolipsismX
Reply
post #82 of 86
6/25/09 at 1:15pm
Quote:
Originally Posted by hmurchison View Post

So are you saying that a Microprocessor from the same family running at the same frequency will yield a performance increase strictly from a processes shrink?

No. I didn't write that. What I wrote is equivalent to: a microprocessor from the same family running at the same frequency will yield a performance increase per watt strictly from a process shrink -- and the performance increase per watt will be dramatic.

Given that laptops are limited in the maximum TDP they support, to maximize performance one takes the fastest processor within the TDP constraint. If that constraint is, say, 35W then a 35W 32nm processor is going to run circles around a 35W 45nm processor using the same microarchitecture. That is true regardless of the number of cores because adding cores, at any given technology level, requires slowing the clock to maintain TDP.
Mac user since August 1983.
Reply
Mac user since August 1983.
Reply
post #83 of 86
6/25/09 at 2:25pm
Quote:
Originally Posted by mcarling View Post

No. I didn't write that. What I wrote is equivalent to: a microprocessor from the same family running at the same frequency will yield a performance increase per watt strictly from a process shrink -- and the performance increase per watt will be dramatic.

Given that laptops are limited in the maximum TDP they support, to maximize performance one takes the fastest processor within the TDP constraint. If that constraint is, say, 35W then a 35W 32nm processor is going to run circles around a 35W 45nm processor using the same microcode. That is true regardless of the number of cores because adding cores, at any given technology level, requires slowing the clock to maintain TDP.

Yes yes yes I gotcha. Sorry and thank you for the clarification. I'm glad I'm not really in the market for a notebook because the computer vendors are going to have some tough decisions to make regarding product lineup in particular Apple because they have such thin and svelete notebooks.

Apple just might skip the initial Clarksfield rush and wait for Westmere based parts. I read an interesting article today about Intel remaining steadfast to full node process shrinks so they'll be skipping 28nm and shooting for 22nm next.

The article also stated that Larrabee will have the same "tick tock" architecture strategy where it'll be announced at 45nm and then migrate not only to 32nm but also ondie. Now I know that Intel hasn't been even close to stellar with GPU but they are putting a lot of work into Larrabee and should it be a success and quickly migrate to ondie it only gets rougher for Nvidia and ATI going forward.
He's a mod so he has a few extra vBulletin privileges. That doesn't mean he should stop posting or should start acting like Digital Jesus.
- SolipsismX
Reply
He's a mod so he has a few extra vBulletin privileges. That doesn't mean he should stop posting or should start acting like Digital Jesus.
- SolipsismX
Reply
post #84 of 86
6/25/09 at 2:39pm
Quote:
Originally Posted by hmurchison View Post

Yes yes yes I gotcha. Sorry and thank you for the clarification. ...

Apple just might skip the initial Clarksfield rush and wait for Westmere based parts.

I'm glad I finally found a satisfactory way to explain why I don't expect Apple to put Clarksfield processors in their laptops. Sorry that it took so long.
Mac user since August 1983.
Reply
Mac user since August 1983.
Reply
post #85 of 86
6/25/09 at 9:39pm
Quote:
Originally Posted by mcarling View Post

What some people here seem to overlook is that a 32nm Nehalem processor with any number of cores will necessarily and dramatically outperform a 45nm Nehalem processor with any number of cores if the clock rates are set such that the TDP is the same.

True, especially given that Intel has had a lot of head room lately due to the improved processes. What I wanted to stress though is that I see to much emphasis on clock rate comparisons, which frankly are not worth considering given the vastly different instructions per clock with C2D and Nehalam derived chips. It might be useful comparing within the same family and generation dual/quad but is useless when comparing a C2D with an i7 derived quad core.

I've just seen to many posts with a bunch of numbers with GHz trailing them.

Dave
Reply
Reply
post #86 of 86
6/26/09 at 2:10am
Applying the general rule to the specific question at hand, a 32nm Arrandale processor will compute circles around a 45nm Clarksfield processor if the clockspeeds are set such that the TDPs are equal.

Running both at max clockspeed would most likely result in the fastest Clarksfield being faster than the fastest Arrandale (at the cost of much higher TDP), but that's not very relevant to laptops.
Mac user since August 1983.
Reply
Mac user since August 1983.
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Future Apple Hardware
AppleInsider › Forums › Mac Hardware › Future Apple Hardware › Intel Core 2 Quad on MBP and iMac.