SPCR Turion 64 MT-ML power draw comparison.

[carlosd]

Originally posted by: Markfw900
The links provided either didn;t work, or were not valid. The one IBM only had 256 meg memory, 40 gig and cd burner vs my 512/80 gig/DVD burner. And the one day Dell deal that I didn;t even pursue that was valid one day last december. Find me one today with comporable equip. I already tried and I didn;t find any. And it was still $200 more !!!!

Yeah! mark is right, neither could find similarly configured Pentiums M at same price or cheaper than Turions. Please find it I am dying to see that.

 

[Viditor]

Originally posted by: coldpower27

Originally posted by: Viditor

Originally posted by: Furen
Now, let's take see... is the northbridge part of Dothan? No. So you can't include it as part of the CPU's power draw because it's simply not there. Yes, I know that you NEED a northbridge in the system but it is not part of the CPU in Intel's design. I alsosaid that the difference diminishes when you look at total system power draw, which includes the northbridge for both, in addition to the other stuff needed for the systems to work.

Exactly my point...so when you compare efficiencies, you must add the power draw of the Centrino Northbridge to the Dothan to compare apples to apples. I wasn't saying you were wrong, just clarifying...

I am going to have to agree with Furen on this one, regarding power consumption, you can't compare apples to apples with regard to CPU because Intel didn't include the NB in their design, so hence you don't add the NB into the CPU power draw, if you want "apples to apples", you might as well compare Turion and Centrino in a platform style, because while Dothan needs the memory controller to run, Turion still needs the chipset to be able to run in a system as wel. It also isn't relevant, to compare CPU to CPU anyway as they both need platforms to run on. It's only interesting from a theoretical standpoint.

Let me draw an anology to help explain my point here...
If you had a CPU with on-board L2 cache and one with the L2 cache on the mobo, would you compare the power draw of the 2 processors equally (i.e. seperating out the L2 for one and not the other)?

For Accord...

Turion MBs still have a northbridge, it just doesn't have a memory controller

While it's allowable to add a Northbridge to an AMD mobo (Via does it because they prefer to use VLink), it's not usually done or necessary. For instance, with an ATI or Nforce chipset, there is only a Southbridge which connects directly to the CPU.

 

[Accord99]

Originally posted by: Viditor
For Accord...

Turion MBs still have a northbridge, it just doesn't have a memory controller

While it's allowable to add a Northbridge to an AMD mobo (Via does it because they prefer to use VLink), it's not usually done or necessary. For instance, with an ATI or Nforce chipset, there is only a Southbridge which connects directly to the CPU.

http://www.techreport.com/reviews/2004q4/radeon-xpress200/index.x?pg=1

The NForce merely integrates the functionality of the northbridge and southbridge into the same chip.

 

[Furen]

The northbridge is integrated into AMD CPU's in the name of the SRI, the crossbar and the memory controller. It is basically the hub though which the CPU communicates with the rest of the system and other CPUs. If you asked me, I'd say that the chipsets used for A64s are actually southbridges, they're just split into two parts in some cases. And it is usually done this way, only Nvidia and ULI make single-chip core logic and Nvidia is actually moving away from this in its newer chipsets.

Viditor: If one chip had external L2 I would not count it as part of the CPU's power draw BECAUSE it isn't in the chip. Now, if it was in an MCM then I'd cosider it the same since it's packaged together but trying to account for something that is PART of another chip (so we can't accurately measure) is not very feasible.

dexvx: The HD is SLIGHTLY more power-hungry (.1-.2W more at idle idle and 1.2W more while seeking, and I'm sure the HD was pretty idle during CPUburn) so this difference should be negligible. The ram issue is a different matter altogether (I'd say that the difference is perhaps 2-3W), though this may be moot, too, considering that CPUburn doesn't really stress the ram too much.