It wasn’t so long ago that we would see only one or two graphics chips migrate from the desktop to mobile; now, though, not only are we seeing more desktop ASIC’s (presently 5 discrete PCI Express ASIC’s with RV370, RV380, RV410, R430 and R480) but more of these are being brought across to mobile platforms, including the very high end desktop chips. What do you think has been the catalyst for this?

With the rapid growth in the notebook space we have seen increased amount of segmentation with new notebook form factors emerging. This increased growth and segmentation has necessitated expanding our product lines to better meet our customer’s needs. We now have to offer a product line to accommodate everything from 4lb thin and light commercial notebooks to 12lb desktop replacement mobile gaming platforms. With a top-to-bottom line of notebook PCI Express products we feel we have the right mix to support our customer needs.

Some time ago ATI announced HyperMemory, which leverages system RAM for graphics utilization, although we have yet to see any discrete desktop or mobile platforms using it; what importance is this, particularly in the mobile segment?

In the mobile space this offers a number of advantages for our OEM partners. The adoption of HyperMemory technology enables our customer to reduce costs while still utilizing high performance graphics. In addition, technologies like HyperMemory and CSP enable notebook makers to reduce power consumption. In the coming months you will see a number of our partners using ATI discrete and integrated graphics adopt HyperMemory technology.

We’ve also noticed that along with the Mobility Radeon X300, ATI are listing HyperMemory capabilities for Mobility Radeon X700 and X800 (although, curiously not X600), however these graphics chips are fairly high performance solutions and you wouldn’t really want to be reliant on much slower system RAM as it is much lower performance. Do you think we’ll see these types of configurations utilizing HyperMemory?

That was actually an error on the site. In theory the MRX300, MRX600, MRX700 and Radeon Xpress 200M all support HyperMemory. The reality is that the majority of mobile designs using HyperMemory will utilize MRX300 and Radeon Xpress 200M.

One of the features of PowerPlay 5, your mobile graphics power saving features, was control of the PCI Express lane utilisation such that it would shut down many of the lanes from the 16 lane graphics connection when in low power mode. We’ve heard some suggestions that this doesn’t save much in the first place as PCI Express inherently has power saving functionality and the chipset would need a reset to alter from one lane to 16. Can you give us any indications of how your solution operates and what type of power savings can be gained?

This is simply not the case. When PowerPlay 5.0 is enabled DLCS enables the dynamic reduction of GPU power, allowing the PCI Express lane to reduce from a X16 to X1. This, in tandem with reductions in clocks and overall GPU voltage enable up to 30 per cent reduction in GPU power consumption. For example, with MRX600 we have seen power consumption reduce from 1.59W static screen power to 1.22W, just by enabling this feature with all else being constant. That’s almost 400mW savings just from going from x16 to x1 without any other changes. This feature alone can provide an additional 15-20 minutes of battery life.

On the subject of power, under present Windows environments, when operating on the desktop you can rely on just the VGA engine, which uses far less power than the 3D processing elements of the chip, however Microsoft next generation OS, codenamed Longhorn, will feature a fully 3D enabled desktop environment. Presumably this change will have ramifications in terms of power utilisation for graphics, which will inevitably impact mobile platforms the most. Is this the case and if so what type of changes do you think we’ll see in power management technology from the graphics vendors?

This is really a question that’s better suited for Microsoft to answer.

Another technology area that is approaching fairly fast is that of High Definition Media and HD optical players. Presently we can get at least a full DVD movie worth of battery power in most laptops, however inevitably Blue Ray and / or HD-DVD players will turn up in notebooks, and users will expect similar battery performances even though they are playing HD content, which will require more processing power and, hence, impact of power draw.

What is ATI doing from the video decoding and display side in order to minimize power draw in order to maintain a battery life of at least a movie when playing HD content?

ATI has long been a leader in using GPU technology to offload processing from the higher power CPU. This is an area where ATI has long lead and innovated in and you can expect some significant developments in this area over the coming months.