Monday, 14 January 2013

The Tech Detail Behind Intel's 4th Gen 7W Ivy Bridge CPU



Intel introduced a handful of new, lower-power Y-series Ivy Bridge CPUs designed to fit into thinner and lighter Ultrabooks and tablets. The slides in Intel's keynote called these "7 watt" Ivy Bridge CPUs, and the company compared them directly to the 17 watt U-series chips in wide use today.
The actual thermal design power (TDP) of those processors is in fact 13 watts—still lower than before, but less miraculous than had previously been implied. A new measurement, scenario design power (SDP), was actually being used to achieve that 7 watt figure.
There was one particular element of that write-up not entirely accurate: based on our conversations with Intel reps, we thought SDP was purely a marketing ploy, a measurement of the amount of power the processor would use on average. It turns out there is an element of marketing to these new 7 watt CPUs, but there's a technical element, too—Intel is simply giving a name to and publicizing a measurement previously left behind-the-scenes. We talked with an Intel engineer to get a better explanation.
TDP explained
To fully understand what's going on here, you need to understand TDP. In Intel's case, a specified chip's TDP has less to do with the amount of power a chip needs to use (or can use) and more to do with the amount of power the computer's fan and heatsink needs to be able to dissipate while the chip is under sustained load. Actual power usage can be higher or (much) lower than TDP, but the figure is intended to give guidance to engineers designing cooling solutions for their products.
In laptops especially, balancing power usage and performance is all about sliding clock speeds, active CPU cores, and other values up and down dynamically depending on what the computer is doing. Older mobile processors could simply ramp their clock speeds down when idle to save power, but newer Turbo Boost-equipped CPUs can actually exceed their stated clock speeds (and their TDP) for a short amount of time. The idea is that the sooner the CPU can get its work done, the sooner it can go back into a power-sipping idle state. If the CPU needs to work hard for an extended period of time and the laptop gets warmer, it will slowly ramp down its speed until it's operating at its stated TDP.
There are two OEM-configurable "power level" states that define how quick the CPU can be in these situations: PL2 tells the processor how much power it's allowed to use when it needs a short burst of speed, and PL1 defines how quickly the processor can run under sustained load. Both of these PL states are given in terms of watts, and generally speaking PL1 is set to the CPU's maximum TDP value.
Here's where we start to get closer to what Intel is doing with its Y-series CPUs: for its U-series 17 watt TDP CPUs, Intel actually tests them and validates them for usage at three different PL1 values: 17 watts (the max TDP), 20 watts, and 14 watts. The latter two values aren't officially defined anywhere on the CPUs' official product pages, but they're there for OEMs whose cooling solutions either exceed or undershoot the TDP value.
Under what Intel considers to be "normal" workloads—short bursts of activity followed by a relatively prompt return to an idle state—the same CPU will perform at exactly the same level regardless of the PL1 value set by the OEM. It's under sustained heavy workloads—things like gaming, video and photo editing, crunching big databases, and so on—that the PL1 value makes a difference. Depending on what it's set at, you might have two different laptops using the same processor that actually perform differently under load.
SDP: Taking it one step further
This is at the heart of what Intel is doing with the Y-series processors: their maximum TDP has been lowered four watts, from 17 to 13. Intel is also validating them for use at two lower PL1 values: 10 watts and 7 watts. This is where the marketing we discussed earlier comes in—rather than keeping these values under the covers as it has so far been content to do, Intel has taken that lowest value, put it on its product pages, and called it SDP.
Intel told me that these SDP values are only intended for use with current and future Y-series processors. U-series, M-series, and desktop class processors will continue to use Intel's TDP ratings, at least for the foreseeable future, and while they may be validated for use at lower and higher values it won't be a core part of those CPUs' specifications as it is with the Y-series chips.
For systems that do use these processors, they will enable thinner and longer-lasting laptops and tablets, but they make it more difficult to tell exactly how fast a given PC will be when under sustained load. Short bursts of CPU activity will run equally quickly on a Core i5-3339Y whether the PC OEM sets its PL1 value to the SDP of 7 watts or the full TDP of 13 watts, but for long-running CPU-heavy activities like gaming and video encoding there's a chance that two different PCs running the same processor may perform the same task at different speeds. The near-50-percent drop in TDP is also likely to cause a much more noticeable performance hit than a 17 watt processor with a PL1 value set to 14 watts.
Is any of this deal-breaking? Almost certainly not—even when running at the 7 watt PL1 value, the systems will still be much quicker than anything based on Atom and ARM would be. The increased efficiency will also enable better performance and battery life to even smaller and slimmer devices than we currently have. Even if OEMs stick to the max TDP of 13 watts, you're still saving power over the older 17 watt processors while maintaining performance similar to that found in Sandy Bridge-era Ultrabooks. This is a testament to both the increasing maturity of Intel's 22nm manufacturing process and processor binning, which Intel assured me it was doing in order to find the most power-efficient processors.
It also helps the tasks that will expose speed discrepancies between systems with disparate PL1 settings are usually best left to beefier systems. Serious gamers, professional video editors, and others to whom top-flight performance is vital will likely want to have larger but faster laptops and desktops to speed those tasks along.
What these processors do unfortunately add to the mix is another value the technically savvy need to be aware of when they buy their systems—one that we'll need to be particularly vigilant about in reviews of PCs that use them. Intel isn't being specific about the clock speeds or performance levels we can expect from PCs with differing PL1 values. Instead, we were told, essentially, that we would need to benchmark systems as we received them and find this out ourselves. Especially in PCs using Y-series processors, Ars will be careful to do this in our reviews going forward. [Arstechnica]

You can follow me on Twitter, add me to your circles on Google+ or Subscribe to me on facebook or YouTube. You can also check my website and blog to keep yourself updated with what is happening in the ever changing world of technology

No comments:

Post a Comment