Original Link: https://www.anandtech.com/show/714
It may be a bold statement, but the truth of the matter is that the world of the notebook computer is evolving at a rate faster than that of the desktop computer. Although rapid evolution is clearly present on the desktop market, it is occurring to an even greater extent in the mobile platform. Functions previously reserved to desktops are quickly making their way into the realm of notebooks.
It was not long ago that everything but the most basic word processor functions were tied to the desk, along with a power outlet and a monitor. As a result of the technological advancements in the mobile industry, mobile computing has infiltrated a whole new market, freeing the computer from the confines of a cubicle. Although technology has steadily been changing the way we look at mobile computing, some barriers have historically held the notebook back from certain markets. For instance, no one in their right mind would consider replacing a $10,000 video editing computer with a $2000 to $3000 laptop.
Another arena where notebook computers have typically been dominated is that of 3D gaming. Any serious gamer who even considered switching from a high end desktop to a laptop of any sorts was thought of as crazy. Not only have notebook computers typically lacked the processor power, hard drive space, and memory requirements necessary for the most advanced games out there, the video card processor usually integrated on a notebook's motherboard has been far to slow to play recent games in a decent resolution. However, as processor speeds, storage space, and memory in laptop computers have slowly been catching up to those in desktop computers, the video solutions found in most notebook computers remain sub par compared to their full sized counterparts.
Lucky for us, those who crave fast 3D performance in any computer they plan on using, desktop trends in video card technology are finally trickling their way down to the laptop computer. Last month we heard a plan from 3D powerhouse NVIDIA on how they plan to bring full resolution 3D to the notebook market. Today we look at a product announcement from a company that is not new to the mobile market but aims to raise it to new levels: ATI. Their new product? The Mobility Radeon, a chip that promises to not only alter the face of 3D mobile computing but the face of ultra light computing as well, hitting the GeForce2 Go head on.
Bringing 3D to the Notebook
In order to bring high end 3D performance to the mobile platform, ATI turned to a chip that it has had much success with in the past: the Radeon. Although the Mobility Radeon shares many features with the desktop Radeon chip, the Mobility Radeon does not use the same core as the desktop version; rather ATI choose to implement some Radeon features in the Mobility Radeon. Let's see what ATI choose to use.
First off, the Mobility Radeon uses a crippled version of the Radeon's rendering pipeline, meaning that it has one of the Radeon's two, three texture pipelines, thus the chip is capable of processing 3 texels per clock. As we pointed out in our Radeon 64MB DDR review, this feature may be nice in future games but goes pretty much unused currently. The reason for this is that the Mobility Radeon's 3 texture pipeline will only be able to take advantage of the two textures that the vast majority of games out there now use. For example, when rendering a scene in ID's popular Quake III Arena game, the Mobility Radeon's core will only be able to process two texels per clock, as the extra texture unit on the chip's pipeline goes unused. ATI claims that support for three textured games is right around the corner, and people such as John Carmack support the idea of using the additional texture in games, but the truth is that the majority of stuff out there now will only be able to use two of the Mobility Radeon's three texels per pass. Let us point out once again that this is not necessarily a bad thing, especially considering that it is the memory bandwidth that provides the limitation, not the texel per second speed. What is bad, however, is the loss of the extra pipeline. This will certainly decrease chip performance.
Once again included in the Mobility Radeon's core is TMDS, a controller for driving digital video output. Naturally this is important when it comes to notebook computers, as they use this controller to power the LCD display. However the Mobility Radeon now gets dual display support courtesy of ATI's HydraVision software, meaning that an image or the desktop can be split among both a laptop's LCD and an external monitor.
We have not yet had a chance to play with HydraVision, but from the looks of it the software will provide support similar to what NVIDIA is currently offering in their TwinView enabled cards. This means that not only can a desktop be stretched across two monitors, but a game can be played from two different views (if there is enough power) and two applications can be run on different displays. This application of HydraVision seems to make the most sense in the notebook market, since it will now be easy to bring you mobile computer home, hook it up to your desktop monitor, and take advantage of HydraVision technology.
HyperZ Revisited
ATI has also included HyperZ technology in the Mobility Radeon, meaning that this mobile processor can also benefit from the memory bandwidth optimizations that come with the technology. You may recall that HyperZ technology, made up of Hierarchical Z, Z-Compression, and Fast Z-Clear, was able to bring new performance to the 3D graphics market. Luckily, the Mobility Radeon is able to bring this technology to the mobile platform, especially considering what kind of memory bandwidth limitations will surely exist.
Why, you may ask, will memory bandwidth be even more important in a laptop? The reason is a result of ATI's choice in memory bus widths. The Mobility Radeon core is able to support two different memory bus widths: 32-bit and 64-bit wide configurations. Now compare this to the 128-bit memory bus that is currently limiting us in desktop 3D and you see where the problem comes in.
The most likely configurations for the Mobility Radeon's memory subsystem are 32-bit DDR and 64-bit SDR solutions. Although a 64-bit DDR solution is possible, it is unlikely that many manufacturers will put the extra money into producing such a solution, as it is much cheaper to use the 32-bit solution with DDR memory than it is to use either a 64-bit SDR or a 64-bit DDR solution.
In either popular configuration, both 32-bit DDR and 64-bit DDR, the memory system is left with very little memory bandwidth. ATI suggested that possible memory clock speeds were up to 166 MHz or 183 MHz. Both ways leave the system severely crippled: with only 1.3GB/s (166 MHz) or 1.5GB/s (183 MHz), both solutions fall short of the 5.8GB/s that current DDR desktop cards posses.
It is for this reason that HyperZ technology is going to play a crucial role for the Mobility Radeon. By essentially decreasing the amount of data that must travel over the memory bus, the Mobility Radeon will not be as memory bandwidth crippled as other solutions out there, including NVIDIA's GeForce2 Go. However, the GeForce2 Go does feature support for a 128-bit bus, meaning that theoretically the memory bandwidth could be brought up to the level we see in current desktop solutions. The sad truth of the matter is that it is extremely unlikely that we will see any manufacturers take advantage of this high end memory bus support, as the amount of traces required for it would be very costly and take up quite some space.
On the subject of clock speeds, ATI mentioned to us that the Mobility Radeon is able to operate at clock speeds from 166 MHz or 183 MHz all the way down to 66 MHz. More on that low number in a minute, but it is clear that the Mobility Radeon is able to clock right up there with its desktop counterpart.
Some Specialties
The Mobility Radeon also has some aspects that ATI is betting will be extremely popular in the mobile market. First off is a hardware feature of the chip known as inverse discrete cosine transform, or iDCT for short. As we showed in our DVD Roundup, iDCT is able to significantly reduce the amount of work that a CPU has to do when performing DVD playback. By offloading some of the work normally placed on the CPU, ATI claims that iDCT is able to increase battery life by 20 to 30 minutes when playing back DVD movies.
The second new technology that the Mobility Radeon brings to the table is its power management technology. Much like Intel's SpeedStep technology, the power management on the Mobility Radeon allows the core to operate at two different clock speeds and voltages: a "low" speed when running on battery and a "high" speed when plugged in. Also added is the ability for the user to dynamically change the clock speed and voltage to anything he or she wants. Naturally, we will have to wait before we see what kind of difference this technology makes, but it should prove to be interesting.
A Few Downsides
You may have noticed that while discussing what features the Mobility Radeon gets from the desktop Radeon, the Radeon's Charisma engine was left out. The reason for this is that ATI essentially cut this part of the chip out in order to save space, money, and power.
By cutting out the Charisma engine, ATI essentially removed the Mobility Radeon's T&L support. Considering how much the Mobility Radeon shares with the Radeon, we were surprised to hear that the T&L engine had been cut. It was only six months ago that ATI was proclaiming that their T&L engine was superior to all others.
The biggest disappointment with the loss of the T&L engine is the Mobility Radeon's inability to move the common transform, lighting, and clipping calculations from the CPU to the graphics chip. Doing so would undoubtedly offloaded some strain placed on the CPU and diverted it to the less power hungry and much more available graphics chip, in a similar manner that iDCT was able to do during DVD playback.
As we have mentioned before, T&L support does not drastically alter the nature of gameplay currently, but it seems that it is just a matter of time before it does. Although right now the lack of T&L support may cost the Mobility Radeon very little when it comes to performance and battery life, things may change as games continue to evolve.
More Than Just Performance
Although the Mobility Radeon is clearly able to match mobile graphics performance with the best of them, ATI is not solely committed to serving the needs of power craving gamers. Although the Mobility Radeon is able to serve these users rather ready handedly by offering support of up to 64MB of DDR memory, it is also able to find its way into ultraportable market with the help of up to 16MB of on-die memory.
Contrary to popular belief, use of on-die memory does not increase performance over external memory; all that it does is save space, and in the ultraportable market, space is more valuable than gold. Currently ATI is in a unique position to provide a video solution for the ultraportable market, as they currently have one of the lowest voltage, high performing processors out there currently: the Rage Mobility 128.
With the introduction of the Mobility Radeon, ATI is poised to take even more of the ultraportable market. For example, the ultraportable Sony Picturebook, which uses a Transmeta Crusoe processor, uses the Rage Mobility for its graphics chip due to its integrated memory solution. It is only a matter of time before the more powerful Mobility Radeon, clocked at a low clock speed, finds its way into this ultraportable and more. Plus, with the ability to run at different speeds when plagued in and not, 3D gaming may finally find its way into the ultraportable market, as long as you are plugged in.
The Packaged Deal
ATI has more plans for the Mobility Radeon than just those of high end 3D graphics. ATI is going to many venders with a new idea: a mobile All-in-Wonder of sorts. Rather than be forced to create an integrated All-in-Wonder solution, ATI seems to be suggesting that venders package Mobility Radeon computers with USB TV Wonder. By doing so, ATI can produce an All-in-Wonder type product for use on laptops. Included is the ability to "time shift" live TV and record shows for later playback. For more details regarding the All-in-Wonder idea and what it can do, check out our Radeon All-in-Wonder review.
From the looks of it, ATI has an interesting idea on their hands by coupling the Mobility Radeon with the TV Wonder. Since the product is USB based, it can easily be used in any laptop out there today. This does result in a problem, however, as nothing is stopping a manufacturer from using the NVIDIA GeForce2 Go product along with the TV Wonder. One advantage to using both the Mobility Radeon and the TV Wonder together is that it simplifies things for manufacturers by allowing them to work with only one vender. Then again, the end user can just as easily purchase a TV Wonder for use with another graphics chip if desired.
Conclusion
There is no question that the mobile market is growing at an exponential rate. Current estimates place the mobile market at 35 million computers in 2001, and that number keeps growing. As the technology from desktops slowly make its way down into a computer the size of a binder. Luckily, with both ATI's and NVIDIA's entrance into the mobile market, we can look forward to a bright future.
There is no question that the Mobility Radeon is a step up from the current notebook solutions. Offering not only more power, but smarter use of the power, the Mobility Radeon will finally start bringing 3D to the notebook market. However, the question of whether the Mobility Radeon can take on NVIDIA's GeForce2 Go still stands.
Although we will not know which solution will rise to the top until we get some products in our hands for testing, both solutions seems as if they will provide similar performance. ATI claims that they offer "significantly lower" power consumption than the GeForce2 Go's 2.8 watts, however it is unclear whether this "significantly lower" number was generated while plugged in in "high" performance mode or while on battery power in "low" performance mode. Once again, only hands on tests will tell.
The only downside that the Mobility Radeon seems to have when compared to NVIDIA's GeForce2 Go is its lack of T&L support. As discussed above, T&L certainly would not hurt performance and could very well increase battery life to a small amount. Then again, the Mobility Radeon has iDCT, something that the GeForce2 Go does not have. This feature seems that it will increase battery life to at least some extent when performing DVD playback.
With the introduction of the Mobility Radeon, have we reached the point of equality where notebook computers are just as powerful as their full sized brothers? Not yet, but we are coming close. And the competition created in an all out battle between ATI and NVIDIA will only make us advance faster.
Facebook
Twitter
RSS
