x86 Everywhere: Intel Announces Medfield Phones - HotHardware

x86 Everywhere: Intel Announces Medfield Phones

7 thumbs up
Scaling Power and Performance

Intel has sunk an enormous amount of effort into optimizing and improving Medfield's power consumption and frequency scaling. In order to understand why this matters, it helps to examine the relationship between voltage, frequency, and power consumption.



The graph above is for demonstration purposes only. It illustrates the relationship between CPU performance and total power consumption. At the lower end of the graph, CPU clock speeds increase much more rapidly than the chip's total power consumption. As voltage climbs, however, the graph flattens. At the right-hand side of the graph, power consumption increases more quickly than performance.

Saltwell is designed to take advantage of this curve. Clock control is fine-grained; adjustments are available in 100MHz increments. The CPU is designed to transition from sleep to active mode very quickly; Saltwell's worst-case exit latency for C1/C1E (CPU powered off) is just 350ns. The chip can wake up from C6 (deepest sleep, the entire SoC is powered down) in 70 microseconds.

Saltwell Flying Solo:

The ARM industry has put a considerable emphasis on using multiple cores inside a single SoC of late. Texas Instruments' OMAP4 platform incorporates a pair of Cortex-M3 cores for low-power operation, while Nvidia's fifth companion core is designed to lower the SoC's power consumption in stand-by or when performing low-level tasks.


Left, Nvidia's description of why using a Companion Core makes sense. Right, ARM's big.LITTLE concept.

The ability to transition quickly from one state to another is critical to Intel's strategy for mobile devices. Fine-grained power control allows Santa Clara to avoid the need for combining multiple CPUs (a strategy ARM refers to as big.LITTLE) or building specialized "companion" cores a la Nvidia. Intel has eschewed this policy in favor of fine-grained power control and extensive use of hibernation / sleep states.



This chart is from the Moorestown launch, but Saltwell's only difference is how it handles the L1 cache in C1/C2. Moorestown flushed the L1 when it dropped into these states--Saltwell / Medfield doesn't.

The following chart illustrates how fine-grained control can reduce total power consumption per task as well as improving device performance.


In this example, both CPUs begin in standby mode, drawing no power. CPU 1, when activated, establishes a constant frequency and begins data crunching. This essentially mirrors Nvidia's vSMP technology, where all cores run at the same clock speed. CPU 2, in contrast, starts off at 0.3W for an initial burst of calculations, ramps up to full power (0.8W) for a short period of time, falls back to 0.3W to close out its task, and then returns to standby.

Total power consumed over 15 seconds in this example is 4W for CPU 1 and 3.6W for CPU 2. In absolute terms, CPU 2 drew 10% less power.

Be aware, however, that this data is easily perverted--and the incentives for companies to do so are high enough that we're including the following as an early warning against future snake oil. If we examine average power consumption over the total time it takes for each CPU to calculate the workload, the tables turn. CPU #1 suddenly looks more efficient, with an average power consumption of 0.5W compared to 0.514W for CPU #2.

Because an arithmetic average is calculated in terms of power consumed per second, CPU #2 is mathematically penalized for completing the workload more quickly. The best way to avoid being tricked by bad graphs is to only trust data on average power consumption when the figures have been calculated against the same amount of time.
 

Article Index:

0
+ -

They are late to the game and they show up with a 32nm Atom is this a joke.

0
+ -

Power consumption on the device looks excellent --fully comparable to current phones. There's no effective difference between 32nm and 28nm, particularly given that Intel's 32nm is extremely solid. This is no joke.

0
+ -

Being late has nothing to do with whether they can compete. Plenty of products get introduced into the market all the time and it only depends whether they can compete and not when they are introduced. While being first is no guarantee of success either.

Nor does it matter if a company has failed before, other than stigma and getting consumers to believe in a product but that can all rapidly change. Really, company statuses go up and down all the time and just because a company is down now doesn't mean it can't become tops in very short order.

Never mind for a reality check, ARM is only coming out with 28nm products in the mid to later half of this year. The Tegra 3 for example is still a 40nm product that's even still based on Cortex A9 instead of the more capable Cortex A15. Nvidia may update to 28nm before the 3rd quarter of the year but they may decide to just wait till the Tegra 4 is ready at the end of the year or early next year.

ARM in general is still all 32bit, they're still a few years away from even sample 64bit products. Hardware fragmentation is still a serious issue. Until Windows 8 for ARM comes out in 2013, ARM still lacks a major desktop OS and even then it won't have full legacy support. While in terms of performance only the next gen 28nm chips actually rival Intel ATOM for CPU performance. Meaning that up till now there have been no overlap between ARM and Intel.

ARM's only advantage is it was designed from the start for low power usage that made it ideal for mobile and embedded devices. While Intel has from the beginning been designing all purpose processors intended to provide at least the basic performance needed to run a desktop OS. They have literally been competing from opposite sides of the performance spectrum and it's not easy for either of them to expand into the markets dominated by the other.

However, overlap is finally starting to happen and that means both stand to start making progress into the markets they each dominate right now. Meaning 2013 is the year we'll start seeing Intel make progress on mobile devices and when ARM will start being seen in what have been traditional PC products like laptops.

It's a long time coming for both but don't assume anything until we really start seeing them seriously compete.

Many of the game changers for Intel don't even kick in till they reach 22nm for example. Like the performance boosting Tri-Gate Transistors. Along with the full adoption of efficiency boosting SoC designs and other architectural changes that finally starts closing the gap between x86 power efficiency and ARM's. While similarly ARM is introducing some enhancements of their products over the next year as well.

So we'll see in 2013 to 2014 how this will play out.

0
+ -

if you read the article you'll see why this is no joke at all.. this should be a very solid first step to make

0
+ -

This is why I love technology :) go intel! .....well , i dont know if you guys didnt noticed but i see a iphone based design.

0
+ -

If you're implying that Apple might use Medfield for a future iPhone, that's not happening. Apple has heavily invested in its own version of an ARM core, and the A6 is already well into development.. This point was raised at our meeting with Intel, and the company wasn't willing to even speculate at what point Apple might consider a switch. My personal guess is that this doesn't even become remotely likely until the chip *after* next (let's call it Future Core 2.) 

If you mean that the prototype phone looks something like an iPhone, it looks more like a generic Samsung. In this case, the important factors are its size and weight. Both of these are within standard parameters for currently shipping devices.

0
+ -

It looks like x32 was on its way out. It will be for normal machines and servers soon but now our phones should be able to run 32 bit sometime soon.

0
+ -

All modern ARM and x86 processors are 32-bit. There's no plan to move to 64-bit phones on anyone's roadmap and several reasons not to do so at the moment -- 64-bit code is larger than 32-bit code, and cache / memory space is still at a premium in phones.

Login or Register to Comment
Post a Comment
Username:   Password: