Homework: How Turbo Mode Works

AMD and Intel both figured out the practical maximum power consumption of a desktop CPU. Intel actually discovered it first, through trial and error, in the Prescott days. At the high end that's around 130W, for the upper mainstream market that's 95W. That's why all high end CPUs ship with 120 - 140W TDPs.

Regardless of whether you have one, two, four, six or eight cores - the entire chip has to fit within that power envelope. A single core 95W chip gets to have a one core eating up all of that power budget. This is where we get very high clock speed single core CPUs from. A 95W dual core processor means that individually the cores have to use less than the single 95W processor, so tradeoffs are made: each core runs at a lower clock speed. A 95W quad core processor requires that each core uses less power than both a single or dual core 95W processor, resulting in more tradeoffs. Each core runs at a lower clock speed than the 95W dual core processor.

The diagram below helps illustrate this:

  Single Core Dual Core Quad Core Hex Core
TDP
Tradeoff

 

The TDP is constant, you can't ramp power indefinitely - you eventually run into cooling and thermal density issues. The variables are core count and clock speed (at least today), if you increase one, you have to decrease the other.

Here's the problem: what happens if you're not using all four cores of the 95W quad core processor? You're only consuming a fraction of the 95W TDP because parts of the chip are idle, but your chip ends up being slower than a 95W dual core processor since its clocked lower. The consumer has to thus choose if they should buy a faster dual core or a slower quad core processor.

A smart processor would realize that its cores aren't frequency limited, just TDP limited. Furthermore, if half the chip is idle then the active cores could theoretically run faster.

That smart processor is Lynnfield.

Intel made a very important announcement when Nehalem launched last year. Everyone focused on cache sizes, performance or memory latency, but the most important part of Nehalem was far more subtle: the Power Gate Transistor.

Transistors are supposed to act as light switches - allowing current to flow when they're on, and stopping the flow when they're off. One side effect of constantly reducing transistor feature size and increasing performance is that current continues to flow even when the transistor is switched off. It's called leakage current, and when you've got a few hundred million transistors that are supposed to be off but are still using current, power efficiency suffers. You can reduce leakage current, but you also impact performance when doing so; the processes with the lowest leakage, can't scale as high in clock speed.

Using some clever materials engineering Intel developed a very low resistance, low leakage, transistor that can effectively drop any circuits behind it to near-zero power consumption; a true off switch. This is the Power Gate Transistor.

On a quad-core Phenom II, if two cores are idle, blocks of transistors are placed in the off-state but they still consume power thanks to leakage current. On any Nehalem processor, if two cores are idle, the Power Gate transistors that feed the cores their supply current are turned off and thus the two cores are almost completely turned off - with extremely low leakage current. This is why nothing can touch Nehalem's idle power:

Since Nehalem can effectively turn off idle cores, it can free up some of that precious TDP we were talking about above. The next step then makes perfect sense. After turning off idle cores, let's boost the speed of active cores until we hit our TDP limit.

On every single Nehalem (Lynnfield included) lies around 1 million transistors (about the complexity of a 486) whose sole task is managing power. It turns cores off, underclocks them and is generally charged with the task of making sure that power usage is kept to a minimum. Lynnfield's PCU (Power Control Unit) is largely the same as what was in Bloomfield. The architecture remains the same, although it has a higher sampling rate for monitoring the state of all of the cores and demands on them.

The PCU is responsible for turbo mode.

New Heatsinks and Motherboards Lynnfield's Turbo Mode: Up to 17% More Performance
POST A COMMENT

340 Comments

View All Comments

  • Gary Key - Wednesday, September 09, 2009 - link

    Actually the manufacturers wanted Clarkdale desperately for the school/holiday shopping seasons. It is delayed as they are still debugging the platform, unofficially I think that means the drivers are not ready. ;) Believe me, if we had a stable Clarkdale platform worthy of a preview, you would have read about it already. Reply
  • justme2009 - Wednesday, September 09, 2009 - link

    You are incorrect sir. The manufacturers were complaining to Intel that they couldn't get rid of the current stock before Intel released mobile Nehalem, so Intel caved.

    http://techreport.com/discussions.x/16152">http://techreport.com/discussions.x/16152

    http://www.techspot.com/news/33065-notebook-vendor...">http://www.techspot.com/news/33065-note...-pushing...

    http://www.brighthub.com/computing/hardware/articl...">http://www.brighthub.com/computing/hardware/articl...

    http://gizmodo.com/5123632/notebook-makers-want-in...">http://gizmodo.com/5123632/notebook-mak...o-delay-...

    Needless to say, I'm waiting for mobile Nehalem (clarkdale/arrendale). With a 32nm manufacturing process, plus starting in 2010, Intel will begin to move both the northbridge and southbridge chips onto the processor die. The move should complete some time around 2011 as far as I can tell.
    It will be far better than what we have today, and I'm really ticked off at the manufacturers for holding back progress because of their profit margin.
    Reply
  • Gary Key - Wednesday, September 09, 2009 - link

    I spoke directly with the manufacturers, not unnamed sources. The story is quite different than the rumors that were posted. I will leave it at that until we product for review. Reply
  • justme2009 - Wednesday, September 09, 2009 - link

    Of course the manufacturers wouldn't fess up to it. It's bad business, and it makes them look bad. It already angered a great many people. I don't think they are rumors at all. Reply
  • justme2009 - Wednesday, September 09, 2009 - link

    Personally I'm holding off on buying a new system until the northbridge/southbridge migration to the processor die is complete, ~2 years from now. That will definitely be the time to buy a new system. Reply
  • ClagMaster - Tuesday, September 08, 2009 - link

    “These things are fast and smart with power. Just wait until Nehalem goes below 65W...”

    I surely will Mr Shimpi with this exceptional processor. I am going to wait until the summer of 2010 when prices are the lowest, rebates are the sweetest, before I buy my i7 860. By that time, hopefully, there would be 65W versions available on improved stepping. It’s worth the wait.

    I would wager the on-chip PCIe controller could use some additional optimization which would result in lower power draw for a given frequency.

    Intel sure delivered the goods with Lynnfield.
    Reply
  • cosminliteanu - Tuesday, September 08, 2009 - link

    Well done Anandtech for this article... :) Reply
  • ereavis - Tuesday, September 08, 2009 - link

    great article. Good replies to all the bashing, most seem to have misread.

    Now, we want to see results in AnandTech Bench!
    Reply
  • MODEL3 - Tuesday, September 08, 2009 - link

    Wow, the i5 750 is even better than what i was expecting...

    For the vast, vast majority of the consumers, (not enthusiasts, overclocking guys, etc...) with this processor Intel effectively erased the above 200$ CPU market...

    I hope this move to have the effect to kill their ASP also... (except AMDs...) (not that this will hurt Intel much with so many cash, but it is better than nothing...)


    I see that the structure/composition in this review and in many others tech sites reviews is very good, maybe this time Intel helped more in relation with the past regarding info / photos / diagrams / review guide etc...


    One question that i have (out of the conspiracy book again...) is,
    if the integration of the PCI-Express controller in the CPU die on the mainstream LGA-1156 platform will be a permanent strategy from now on...
    and if the recent delay for the PCI-Express standard 3.0 has a connection with the timing of the launch of mainstream LGA-1156 based CPUs with PCI-Express 3.0 controller integrated...

    Sure, they can launch future LGA-1156 motherboard chipsets with PCI-Express 3.0 controller, but doesn't this contradict the integration strategy that Intel just started with the new processors?
    Reply
  • MODEL3 - Tuesday, September 08, 2009 - link

    I can't edit...
    I just want to clarify that the PCI-Express 3.0 question is for LOL reasons, not taken serious...

    Reply

Log in

Don't have an account? Sign up now