Llano is not AMD’s solution to its processor performance woes. The quad-core A8-3500M processor in the testbed system received benchmark scores 2 to 3 times lower than those of second-gen Intel Core i7-QM processors, and also failed to best Intel’s Core i5. This isn’t terribly surprising, as the processor is simply another revision of the K10 architecture that AMD has been using since 2007, but this does little to curb my disappointment. It’s clear that AMD needs a new architecture to compete with Intel in processor-intensive applications. With luck, Bulldozer will be the refresh the company is desperate for, but it isn’t available yet.
Since competing with Intel on processor performance is out of the question, this entire platform instead must rely on its graphics performance and its portability. Fortunately, these are two areas where Llano shows great strength. Even with dual graphics disabled, the APU was capable of out-performing Intel’s current HD 3000 IGP by a significant margin.
It should be noted, of course, that the testbed was endowed with the quickest graphics part available on an A-series APU, which skews the results. However, those laptops shipping with less powerful APUs will be targeting price points below $600. This means the competition from Intel will likely be in the form of a second-gen Core i3 processors and various products based on Nehalem, including not only the first-gen Core i3/i5 but also various Celeron and Pentium processors. This means Fusion laptops should retain a graphics edge at any given price point, although predicting how the market shakes out is always a foolhardy task.
Portability is also in AMD’s favor. The endurance of the testbed system was excellent, particularly when the entirely average size of the system battery is considered. If ASUS (for example) decided to throw a 70Wh+ battery into an A-series ultraportable, as it has done with similar Intel systems like the ASUS U33JC, I would not be surprised to see battery life figures of six to eight hours in light use. I’m anxious to see what the less powerful Fusion APUs are capable of, as well. If this quad-core variant with a powerful graphics component is capable of such battery life, a dual-core APU with fewer stream cores may offer even more endurance.
There are many factors that will determine the ultimate appeal of mainstream Fusion laptops when they hit the market, not least of all is the quality of the systems that will be slapped with the Fusion logo. But there is at least reason for excitement. The market for laptops above $600 has been an Intel monopoly for several years, as AMD hasn’t had a product that could compete on the basis of performance or portability. Although the Fusion A-series doesn’t beat Intel’s finest, it does provide unique capabilities, and does so at a low price. These APUs will make it possible for manufacturers to offer more choice, and that’s never bad for consumers and tech enthusiasts.
Ryan’s Thoughts: After having spent a good amount of time with the AMD A-Series notebook I do believe that the platform as a whole is a life-changing development for AMD in the notebook field. They have never been on the cutting edge of innovation in this area even when they were in the world of desktop computers and the mainstream APU is likely just the right thing to get OEMs interested and thus consumers as well.
However, I do think that the x86 cores on Llano are going to be a hindrance to the adoption of this platform in very high end notebooks as well as the desktop market later in the summer. It is much easier to get away with mediocre CPU performance in the mobile markets (just see the popularity Atom garnered) but certain users will always take a look at our CPU-centric benchmarks on the performance page and scoff. AMD will just NOT be able to convince them otherwise until perhaps the next generation of APU based on the Bulldozer core is released in 2012. Until then, I think AMD has a winner on its hands for the sub-$700 notebook but continues to have an issue beyond that.