Notebook Video Graphics Card Guide 2006

by Dustin Sklavos

INTRODUCTION

Since 2005’s GPU review, notebook graphics hardware has done us the kindness of simplifying. The options on the market are somewhat fewer and it’s to our benefit, but it’s still somewhat confusing. I’ve seen questions on the forumsthat to some of us seem comical: Is an X1300 graphics card better than an X700? To the graphics geek, it’s a stupid question. But to a neophyte, it can be very confusing. After all, 1300 is greater than 700, isn’t it?

You see how it is.

This year is a good year for notebook graphics because a number of factors in the equation have simplified, and because gaming is becoming increasingly accessible in the notebook field, and conditions are constantly improving.

I encourage you to go back and read my 2005 GPU review, as it goes over certain subjects in depth, where most of them will only be touched on here. With the changes in the market, this 2006 GPU review sees some refinements from the prior year.

CHANGES

ATIand nVidiahave done us all a favor and cleaned up their lineups a bit. Notebooks have transitioned almost entirely to using the PCI Express bus instead of AGP and as a result, the notebook market has a cleaner slate. As a result, several parts have been retired: those parts are listed in the Retired section later on. If you’re looking for information on those, perhaps as a result of buying a used notebook or a refurbished one, please see my 2005 article, which has detailed information on most of those parts.

I’ve also changed the capsule information for each part in the article in order to simplify things for the neophyte. Hopefully there should be fewer questions and less confusion. For example, I no longer list video memory configurations outside of memory bus size, as most manufacturers now ship their parts with a healthy amount of dedicated memory attached to them.

With the change in the year comes a change in the graphics processing power required to run modern games. While my ATI Mobility Radeon X600 was fine and dandy in spring of 2005, the transitional period has brought games that increasingly punish the processing power of the average graphics card, games like Quake 4, F.E.A.R., and Oblivion. As a result, I’ve changed my my minimum recommendation for hardware as listed in the Need vs. Want section.

I’ve also elected to list Shader Model instead of DirectX version in the information for each part. They do mean roughly the same thing, but listing the Shader Model capability will make things much clearer and again, I’ll go over this in depth later on.

And finally, I’ve shifted parts between performance sections and actually changed the names of the performance sections in order to properly accommodate the changes in hardware and software demands.

INTEGRATED VS. DEDICATED

There are essentially two different types of graphics processors that can be found in computers in general: integratedand dedicated.

Integrated graphicsparts are a part of the motherboard, of the system itself. Motherboards have at least one core chip called a Northbridge, which controls functions like RAM, the PCI Express bus, and how the rest of the computer interfaces with the processor (I’m sure the diehards will probably correct me somewhere, but remember I’m just trying to make this extremely esoteric information palatable.) Integrated graphics have the graphics core built into the northbridge, and they share their video memory off of your RAM. As a result, they eat a little bit of the bandwidth your system is using to interface with the RAM, and reduce the amount of system RAM available. Most integrated parts tend to share that memory dynamically, though, meaning they only take what they need. Only the Radeon Xpress integrated parts have their own memory, and only in AMD systems. I’ll talk more about them later.

Despite the shared memory, there are many benefits to integrated parts. They improve battery life, reduce the cost of the notebook, and allow the notebook to run cooler as it doesn’t need to cool an extra dedicated chip.

The flipside is that they generally offer very poor performance for gaming, with the fastest integrated parts still being beaten handily by the slowest dedicated parts.

And speaking of dedicated parts: a dedicated graphics core is separate from the northbridge, and has its own video memory. This means it doesn’t siphon off the system RAM, although some parts (typically lower end) do dynamically share from system RAM as needed, though only when it would improve performance (and you’ll never see them do so in regular Windows operation, only during gaming).

Dedicated graphicsparts will always outperform integrated parts, albeit in varying degrees. If you’re going to game on your system, you’ll need one of these.

The problem with dedicated parts is that they drive up the cost of the notebook, can increase the temperature (as they need to be cooled by themselves), and reduce the battery life. Additionally, due to the power draw and cooling issues, dedicated parts seldom find themselves in ultraportable notebooks.

HYPERMEMORY AND TURBOCACHE

HyperMemory and TurboCache – boy they sure do sound fast, don’t they? Leave it to ATI and nVidia to use marketing speak to turn something that’s really actually bad for the consumer into something positive.

Since graphics hardware has made the transition from AGP to PCI Express, it can now access system memory at much faster speeds, and as a result, graphics parts can dynamically share system memory. ATI’s implementation is called HyperMemory; nVidia’s is called TurboCache. By implementing these technologies, they can create cheaper parts that require less dedicated video memory. There are differences in the two implementations, however.

nVidia’s TurboCache has features dedicated to its operation built into the graphics chips themselves. As a result, their interface with system memory is usually quite fast. Unfortunately, this also results in a greater number of parts outright relying on this technology.

ATI’s HyperMemory, on the other hand, is done entirely in software, and is enabled on virtually all of their hardware. While conceivably this is a bonus and should at least marginally improve performance on any system with ATI hardware, the flipside is that most often the latency is too poor for it to be of any real benefit to any but the lowest grade of parts.

The major problem with these technologies – and nVidia is particularly guilty of this – is that they basically allow manufacturers and resellers to lie about the amount of dedicated video memory the part has. nVidia parts in the Low End and Mainstream categories are going to be guilty of this, and it or not matter to you as these parts are still fairly solid if unexceptional performers. The only way you can really know how much video memory these parts have, though, is if they say something like x Dedicated / x TurboCache or Shared.

ATI’s parts, apart from the X300 and X1300 are much less guilty of this, and I’ve found that ATI usually ships more than an adequate amount of video memory with their X1000 series parts. The flipside is that their HyperMemory technology is, as I stated before, pretty unremarkable.

MXM, AXIOM, AND UPGRADING NOTEBOOK GRAPHICS

I’ll get this out of the way right now: unless your computer is a Dell, there’s a 99% chance your graphics aren’t upgradeable. More than that, the question has already been asked on our forums, so please don’t ask it. One of my biggest pet peeves is people asking questions that have solutions easily found by just using the search feature.

But suffice to say, unless someone told you your graphics were upgradeable, they’re not.

That said, there do exist certain standards for interchangeable graphics parts, but they have very limited support, and upgrades can be hard to come by.

The first is ATI’s AXIOM standard, which was quite frankly pretty much stillborn. Not a whole lot of information is readily available on it, due in no small part to the lack of support by manufacturers.

The next is nVidia’s MXM standard, which while it hasn’t had the fortune to see widespread adoption, isn’t a half-dead standard like AXIOM either. For what it’s worth, upgrade parts for this can usually be found on eBay, but they tend to be rarefied and can be costly.

MXM parts are broken down into three Types in order of size. Type I parts are the smallest, with Type II being the middle ground and most mainstream. Type III tends to appear strictly in large desktop replacement notebooks. These are NOT interchangeable. You can’t fit a Type III into a Type II slot, nor vice versa, so don’t try.

The final isn’t a standard necessarily. Some notebooks, particularly Dell notebooks; whenever they have upgradeable graphics options when you order them, they tend to also have upgradeable graphics parts. This is NOT a guarantee so don’t mistake it for one, and if you buy a Dell notebook with integrated graphics and expect to upgrade to a dedicated part later, you’re most likely out of luck.

So why aren’t notebook graphics more upgradeable, like your desktop? Well, your desktop doesn’t really have to worry about thermal issues and power consumption like a notebook does. Notebooks are much more specialized due to those constraints. As such, generally the more powerful a part, the more power it draws and the more heat it generates, thus the more cooling is required and the worse the battery life. This is why you’ll never see a GeForce Go 7900 in a 12.1″ thin and light, and this is why notebook graphics are seldom upgradeable.

Ultimately, the part you buy is the part you’re stuck with, so you’d better make it a good one, and I’ll talk about it in the next section.

NEED VS. WANT

This is, in my opinion, one of the most crucial parts of the article. When you’re shopping for a notebook, you aren’t always going to be able to get what you want, so you need to focus on what you need to enjoy a game. The reality is that even though notebook graphics have taken some astonishing leaps in the past couple years (and the performance of the Mobility Radeon X1600 is particularly impressive, all things considered), their performance is still notably under their desktop cousins.

More than that, you have to take into account the kind of graphics power you’re going to be able to get in the form factor you want. If you’re looking at thin and lights and ultraportables, you’ll be lucky to get a dedicated part at all (though GeForce Go 7300 and 7400s in this segment have been spotted in the wild). In a mainstream notebook the most you’re likely to see is a performance class part. But if size isn’t an issue for you, desktop replacement notebooks run the entire gamut, from integrated to high end.

But you have to pay for that performance. Dedicated graphics ALWAYS drive the cost upon a notebook, and the better they are, the more expensive they get. Likewise, the more out of place they are (i.e. the GeForce Go 7400 in an ultraportable), the more expensive the notebook gets.

My point is to be realistic about what you’re expecting to get.

It’s also important to consider how much power you actually need. The casual gamer not need more than the low end. The trick is to neither overbuy nor underbuy. Though overbuying hurts substantially less in the long run, it also cause you to spend more than you really need (or want) to. Don’t listen to the people on the forums that scream bloody murder about how you NEED a Mobility Radeon X700 to play games like Doom 3 and Half-Life 2. Mainstream grade parts onward play most modern games very comfortably and at higher resolutions.

Last year I issued a recommendation for the bare minimum in terms of notebook graphics you’ll need to game for the next few years, and that recommendation was a Mobility Radeon 9600, or a GeForce FX Go 5700; basically, anything that can comfortably run Doom 3 with all the settings turned on.

Since then, more demanding games have taken the fore, and I’m forced to raise my minimum recommendation to at least the Mobility Radeon X600, but ideally you’ll want something from the Mainstream section or better. I also have to suggest a minimum dedicated video memory requirement of 128MB. 64MB just won’t get you very far anymore; this is excepting nVidia’s GeForce Go 7400, which still performs quite handily even with only 64MB of dedicated video memory, due to the efficient nature of its TurboCache implementation.

CAPSULE INFORMATION

On each graphics part, I’ll list the following information; this section is a summary of what each of those pieces of information means.

Shader Model:

There are three different shader models that we’ll see in modern notebook graphics parts: Shader Model 2, 2b, and 3. The higher the shader model, the increased number of graphic features that part supports. This is sort of esoteric, but it means the difference between High Detail and Very High Detail in games like Far Cry and Age of Empires III. A Shader Model 3 part allow you to add additional details to parts of a game that supports it, provided the part itself has enough horsepower to keep the game running playably. (To wit: nVidia’s GeForce FX series all supported Shader Model 2, but the FX Go 5200’s performance was too poor to actually run games using it.) Note that Model 2b is a strictly ATI standard, and is very minimally supported.

Pipelines:

The number of pixel pipelines (shaders) a given GPU has. What these do is generally esoterica; they essentially refer to the number of pixels that can be shaded per cycle in the chip. More is generally better, but the bare minimum is typically 4, though 8 pipeline parts and greater are becoming a lot more common. Again, this line is being blurred and this section very well be obsolete by next year’s article.

Memory Bus Width:

This is a specification that can have an incredibly profound effect on performance. A bigger bus width allows more data from the RAM to be transferred at once. Bus width comes in 64-bit, 128-bit, and 256-bit, and generally 128-bit is the minimum you want. What’s different now as opposed to last year is that nVidia’s low end TurboCache parts have 64-bit bus width, but still offer very acceptable performance.

Specification Clock Speeds:

When a graphics part is released, it typically has a specification for how fast the part and its memory run. On integrated parts, the memory clock won’t be listed as these parts (again, excluding certain Radeon Xpress parts) share system memory. While I didn’t list these last year, I’ve noticed companies have tightened up and standardized these clocks a bit more, so I’m a little more comfortable producing these. These are REFERENCE, however. The MacBook Pro’s Mobility Radeon X1600 is notoriously underclocked, while the Mobility Radeon X600 in my own notebook runs 10MHz over specification on the memory. There is still some variance.

RETIRED GRAPHICS CARD PARTS

The following parts are being retired from the article, as they are no longer modern and rarely if ever appear in modern notebooks. For information on these, please refer to my 2005 article.

Integrated:Intel Extreme Graphics, Intel Extreme Graphics 2, ATI 320M/340M IGP, ATI Mobility Radeon 9000/9100 IGP

Dedicated:ATI Mobility Radeon, ATI Mobility Radeon 7500, ATI Mobility Radeon 9000, ATI Mobility Radeon 9200, nVidia GeForce 4 Go series, ATI Mobility Radeon 9600 (9550), ATI Mobility Radeon 9700, ATI Mobility Radeon 9800, nVidia GeForce FX Go series, ATI Mobility Radeon X800 (XT), nVidia GeForce Go 6800 (Ultra)

Note that a couple of these seem to be retired prematurely while others listed below seem to you as they should’ve been retired. I’ve chosen these largely at my own discretion, but also based on their market penetration. For example, the instant the Go 7800 was released, nVidia pretty much stopped shipping the Go 6800. X700s and 6600s, on the other hand, while gradually disappearing from the market, are still present.

And for some odd reason, OEMs are still madly in love with the X600.

MODERN AVAILABLE GRAPHICS CARD PARTS

INTEGRATED:

Intel Graphics Media Accelerator (GMA) 900/950

• Shader Model 2
• 4 Pipelines

The Intel GMA 900 and 950 are some of the worst integrated parts on the market; this you’ve probably heard. What you don’t really expect is just how bad they are. While they can certainly benchmark fairly close to the Radeon Xpress series, in actual practice their performance is much less fluid, and they have much less compatibility. While these are fine for the average computer user, if you plan on playing any games released after Unreal Tournament 2004 (which itself doesn’t run that well on either part), these are not for you.

ATI Radeon Xpress 200M / 1100 / 1150

• Shader Model 2
• 2 Pipelines

The 1100 and 1150 are refreshes of the 200M; the 1150 has a slightly higher clock while the 1100 is basically the 200M. Outside of the Go 6100 and 6150, these are the fastest integrated parts available. These sometimes appear with dedicated memory attached to them (usually on HP/Compaq notebooks); those versions tend to run about 33% faster. Still, if you plan on doing any more than the most casual gaming, these aren’t a good fit for you, as they won’t run Doom 3 very comfortably – even the parts with dedicated memory. Thus, don’t expect to be playing even more complex games, like Prey or Oblivion, on these. Still, for older games, they’ll be fine. Note that these are usually found in AMD systems and on lower end Intel systems.

nVidia GeForce Go 6100 / 6150

• Shader Model 3
• 2 Pipelines

It took them long enough to return to the integrated graphics market, but with AMD’s release of the Turion 64 X2, nVidia has entered with a very powerful integrated part. The GeForce Go 6150 can be considered the fastest integrated part on the market, albeit winning in sometimes small margins against the X200. The Shader Model 3 support is borderline irrelevant, however, as the core just simply isn’t powerful enough to really take advantage of it. It is, however, the only integrated part that can comfortably run Doom 3, and it can play most games at low resolutions and medium/low settings. While a casual gamer will want to get a dedicated part, if you’re really strapped, you could do worse.

S3 Unichrome, Etc.S3 and SiS parts are the lowest of the low, the cream of the crap. If you don’t plan on doing any gaming on your notebook, these are fine, but note that these only appear on the very cheapest of machines and are completely inadequate for any gaming. As such, they should be avoided at all costs.

Words about the Radeon Xpress 1250

• Shader Model 2b
• 4 Pipelines
• Clocks Not Yet Available

Later this year, ATI is planning to release a new integrated chipset, code-named RS600. The RS600’s core is going to be based on the X700, and is expected to have 4 pipelines. Early performance expectations are around the level of the X600 – no small feat for an integrated part. Keep your eyes open for this one; while I don’t think it’ll hit that X600 target, it will nonetheless be a very important and notable piece of hardware.

Words about the Intel Graphics Media Accelerator X3000

• Shader Model 3/4
• ? Pipelines
• Clocks: 400MHz Core (Expected)

Intel’s parts in the past have been notoriously bad, and anyone who says otherwise is kidding you and themselves. That all stands to change, however, with the GMA X3000. The GMA X3000 will have full hardware texture and lighting (finally) and feature unified shaders. While the GMA 900/950 were supposed to be Intel’s push for full compatibility with Windows Vista, anyone who has used the beta can tell you that just isn’t happening, which has resulted in the design of the X3000. The catch is that while these are expected to appear in the next couple months for Intel’s desktops, their introduction into their notebooks isn’t expected until Q1 2007, so we’ve got a wait on our hands, and by that time, the performance of this part not be enough. If nothing else, though, it’ll be able to run Windows Vista Aero Glass.

LOW END DEDICATED GRAPHICS CARDS

ATI Mobility Radeon X300

• Shader Model 2
• 4 Pipelines
• 64-bit/128-bit Memory Bus
• Clocks: 350MHz Core, 300MHz Memory (600MHz DDR Effective)

The Mobility Radeon X300 is on its way out, having been replaced by the Mobility Radeon X1300. The asterisk notes that some of these have dedicated memory, while others rely on the embarassingly slow HyperMemory. Either way, the X300 isn’t really fit for playing modern games on, and is in no way futureproof. If you’re even a casual gamer, this part is not recommended.

ATI Mobility Radeon X600

• Shader Model 2
• 4 Pipelines
• 128-bit Memory Bus
• Clocks: 400MHz Core, 250MHz Memory (500MHz DDR Effective)

The Mobility Radeon X600 is really the bare minimum for comfortable gaming, and that’s really only the 128MB version. Note that these have appeared on the market with HyperMemory – make sure you know how much dedicated memory you’re really getting. The X600 plays all modern games; some better than others. As an X600 owner at the time of this wriitng, I can tell you that once I hit Quake 4, it started showing its age, and I have an Athlon 64 backing mine up. If you’re planning on using your notebook for gaming, even moderately casual gaming, this is your bare minimum, but you’re realistically going to want to look more at the Mobility Radeon X1400, which is quickly supplanting the X600 in the market anyways.

ATI Mobility Radeon X1300

• Shader Model 3
• 4 Pipelines
• 64-bit/128-bit Memory Bus
• Clocks: n/a

The performance of the Mobility Radeon X1300 resides squarely between the X300 and X600, which unfortunately doesn’t bode that well for it. More than that, HyperMemory versions of it can be found in the wild; those parts are bad and should be avoided. The only blessing is that the Mobility Radeon X1400 is substantially more common than the X1300, which is surprisingly rare.

nVidia GeForce Go 6200/6400

• Shader Model 3
• 4 Pipelines
• 64-bit Memory Bus
• 6200 Clocks: 300MHz Core, 300MHz Memory (600MHz DDR Effective)

  6400 Clocks: 400MHz Core, 350MHz Memory (700MHz DDR Effective)

These are both TurboCache parts, but nVidia’s TurboCache implementation isn’t awful. Still, these parts don’t come recommended, but if you’re stuck, the Go 6400 is certainly a passable, capable part, in some cases approaching the performance of the X600. The Go 6200, however, is probably best avoided entirely. In games based on the Doom 3 engine, the Go 6400 will actually be preferable to the X600, as nVidia notoriously offers superior performance in those games. These parts, though, like the X300 and X600, are being phased out for their newer 7-series siblings.

nVidia GeForce Go 7200/7300

• Shader Model 3
• 4 Pipelines
• 64-bit Memory Bus
• Clocks: 350MHz Core, 350MHz Memory (700MHz DDR Effective)

The performance of the Go 7300 is about around the 6400, the Go 7200 even less desirable. Honestly, these parts are probably best avoided, but fortunately, the 7200 and 7300 are pretty rarefied, as manufacturers have tended more towards the Go 7400, which is perfectly acceptable for moderate gaming.

MAINSTREAM MID-LEVEL DEDICATED GRAPHICS CARDS

ATI Mobility Radeon X700

• Shader Model 2b
• 8 Pipelines
• 128-bit Memory Bus
• Clocks: 350MHz Core, 350MHz Memory (700MHz DDR Effective)

Despite being a Shader Model 2b part, the X700 is still plenty fast and more than adequate for notebook gaming. While it won’t allow you to turn on every feature in new games (since some features require Shader Model 3), it provides plenty of horsepower for gaming and comes well recommended. This, along with the Go 6600, would be the minimum for continued comfortable play for the next couple years. It has enough performance to push most modern games at WXGA (1280×800).

ATI Mobility Radeon X1400

• Shader Model 3
• 4 Pipelines
• 128-bit Memory Bus
• Clocks: n/a

Primarily a faster clocked version of the Mobility Radeon X1300, the X1400 sits squarely between the X600 and X700 performance-wise, but also includes full support for Shader Model 3. While the X700 would still be preferable, the X1400 is very common in notebooks and offers acceptable performance. Not great, but not terrible either.

nVidia GeForce Go 7400

• Shader Model 3
• 4 Pipelines
• 64-bit Memory Bus
• Clocks: 450MHz Core, 450MHz Memory (900MHz DDR Effective)

The Go 7400 is the slowest part in the Mainstream category, achieving performance between the Mobility Radeon X600 and the Mobility Radeon X1400. Truthfully, I quibbled about which category to place this part before settling it into the Mainstream. While it still offers fine performance, especially in games using the Doom 3 engine, it’s probably not the most ideal part. Still, the upshot of the Go 7400 is how frequently it can be found in thin-and-lights, impressively including the 12.1″ Dell XPS M1210, bringing playable gaming to a smaller form factor than traditionally available.

nVidia GeForce Go 6600

• Shader Model 3
• 8 Pipelines
• 128-bit Memory Bus
• Clocks: 375MHz Core, 350MHz Memory (700MHz DDR Effective)

In my last article there was debate over which notebook part was preferable – this or the Mobility Radeon X700 – I think there’s less debate. While the Go 6600 is slightly slower than the X700, Shader Model 3 support is becoming increasingly relevant. Thus I maintain my recommendation from 2005. Mercifully, though, this part is being rarefied as it’s phased out in favor of the substantially more powerful Go 7600.

PERFORMANCE GRAPHICS CARDS

ATI Mobility Radeon X1600

• Shader Model 3
• 12 Pipelines
• 128-bit Memory Bus
• Clocks: 470MHz Core, 470MHz Memory (940MHz DDR Effective)

The Mobility Radeon X1600 is one of the most desirable notebook parts on the market. It offers performance approaching last generation’s top end, running many games at native, and frequently appears with a large amount of dedicated video memory. Most impressive about the X1600 is its appearance in mainstream and even 14″ thin-and-light notebooks. If you’re at least a semi-serious gamer, this part is easily recommended, and will play most if not all games at native with most details maxed out.

nVidia GeForce Go 7600

• Shader Model 3
• 8 Pipelines
• 128-bit Memory Bus
• Clocks: 450MHz Core, 500MHz Memory (1GHz DDR Effective)

Comparable in performance to a desktop 7600GS, the Go 7600 is a little crippled next to its desktop cousin, which has a full 12 pipelines. Performance wise, it’s slightly slower than the X1600 and thus not as desirable, but as I’ve stated before with nVidia parts, if your favorite games run off of the Doom 3 engine, this is going to be the better choice. This part is still a very strong performer, blowing away competitors in the lower classes, and can be found in the same notebook segments as the X1600. Thus, it remains easy to recommend.

HIGH END PERFORMANCE GRAPHICS CARDS

ATI Mobility Radeon X1800

• Shader Model 3
• 12 Pipelines (16 in XT Version)
• 256-bit Memory Bus
• X1800 Clocks: 450MHz Core, 500MHz Memory (1GHz DDR Effective)

  X1800XT Clocks: 550MHz Core, 650MHz Memory (1.3GHz DDR Effective)

Unlike the old Mobility Radeon X800 line, these can actually be found, albeit again in small numbers. These basically trade off with the high end GeForce Go 7 series parts as being the fastest notebook graphics processors available. However, the Go 7 series remains more common and easier to find, with these almost exclusively appearing in boutique notebooks from manufacturers like Alienware.

nVidia GeForce Go 7800

• Shader Model 3
• 16 Pipelines (24 in GTX Version)
• 256-bit Memory Bus
• 7800 Clocks: 350MHz Core, 550MHz Memory (1.1GHz DDR Effective)

  7800 GTX Clocks: 400MHz Core, 550MHz Memory (1.1GHz DDR Effective)

These are gradually being phased out for the faster Go 7900 line, but are still incredibly powerful chips, running even modern games at WXGA or better resolution, with detail settings maxed out. They’re also fairly easy to come by in desktop replacement notebooks, but they demand a premium.

nVidia GeForce Go 7900 GS

• Shader Model 3
• 20 Pipelines (24 in GTX Version)
• 256-bit Memory Bus
• 7900 GS Clocks: 375MHz Core, 500MHz Memory (1GHz DDR Effective)

  7900 GTX Clocks: 500MHz Core, 600MHz Memory (1.2GHz DDR Effective)

Basically just faster versions of the Go 7800 line.

A BRIEF WORD ABOUT NOTEBOOK SLI DUAL GRAPHICS CARDS

I’m sure some of you have heard by now about SLI (Scalable Link Interface) technology, which allows a system to run two graphics processors in tandem, resulting in a substantial improvement in performance. Naturally, some of you have also heard that it’s been making its way into the notebook sector.

This is rumor control. Here are the facts.

SLI is available in some notebooks, but appearing only in the largest, heaviest of desktop replacements. These machines will be loud, heavy, oversized, and have comically low battery life. They barely qualify as notebooks and are better described as portable computers.

Additionally, these notebooks demand a high premium. If you simply must have the fastest gaming performance available in a portable package, and you’re willing to spend for it, then you want to seek one of these beasts out.

SLI appears only using nVidia Go 7800 and 7900 series parts, as it is an nVidia technology. ATI has a similar technology in the desktop market called CrossFire, but it has yet to make the transition, and I haven’t heard of any plans to do so. This is an extremely niche market.

A crucial point to make about SLI on either platform is that the performance increase is NOT linear. Two graphics processors will NOT double performance. Typical gains are between 30% and 50%. This is due to latency with the two graphics processors communicating with one another, and driver support. Some games run better with SLI than others; in some isolated cases, SLI even slightly reduce performance on a game that the driver hasn’t been coded to run it on. However, most mainstream games will see a substantial performance increase.

WINDOWS VISTA AND THE FUTURE

Everyone’s talking about Windows Vista, and they’ve been for the past couple years. Everyone wants their machine to be fully compatible with it.

Honestly, having played with Windows Vista Beta 2, I feel I can comfortably say that it will indeed be worth the wait, provided you have hardware that can run it in all its glory. This means running the most demanding piece of Vista: Aero Glass.

Aero Glass, for those uninitiated, is a visual style for Vista that adds transparency and other features to the windows. It’s VERY attractive in practice, and gives the whole interface a very modern look.

However, it demands a graphics processor with:

• Shader Model 2
• 64MB of Video Memory

So most of you probably think, Oh, well the GMA 900/950 has those, so it can run it. You’re mistaken. Aero Glass will only run on integrated parts at the level of the X200M (1100/1150) or better.

This makes buying a notebook a little bit trickier for some users, as I can’t honestly suggest buying a notebook that won’t run Aero Glass. Without it, you miss out on a major feature of Vista, and though the changes to it under the hood are much more relevant, it will feel largely like a glorified Windows XP.

POPULAR GAMES

Like last year, I’ll list a few of the more popular games on the market right now, along with the minimum needed to run them.

HALF-LIFE 2: Will run even on the GMA 900/950, but really needs at least a low end part to show its true colors.

HALF-LIFE 2 (EPISODE ONE): Really needs a Shader Model 3 part to look halfway decent, and requires around a mainstream part to look good and playable.

DOOM 3: Can run on a dedicated memory X200M, or on the GeForce Go 6150, but really wants at least an X600 to run attractively, and can be a bit of a video memory hog.

QUAKE 4/ PREY: Needs at least a mainstream part to look good; runs playably if unspectacularly on low end parts. Quake 4, in particular, gets substantially more demanding as the game progresses.

WORLD OF WARCRAFT: Runs on almost anything except the slowest integrated parts. Knock yourself out.

F.E.A.R.:Requires at least an X600, but really wants a mainstream part at least.

OBLIVION:Playable on an X600; do yourself a favor and get at least a mainstream part to make it comfortably playable.

CONCLUSION

My final recommendations are as follows:

The casual gamer will want at least a Mobility Radeon X600 with at least 128MB of video memory.

The more serious gamer will want a part from at least the Performance category with, again, a minimum of 128MB of video memory. (256MB is much more common with these parts, and much more desirable.)

It’s unsurprising that games have become more demanding; this has been the subject of heated debate on our forums. The next year is going to see some more big leaps in performance, with games like Crysis and Unreal Tournament 2007 on the horizon.

I strongly recommend avoiding buying a notebook with a GMA 900/950 in it, as these parts aren’t the least bit futureproof and can’t run Windows Vista in all its glory, much less play modern games. Understanding that these are the vast majority of parts in the market right now, they’re still very hard to recommend.

We’re approaching a serious transitional period; during this time I’d recommend erring on the side of caution and be spending up a little bit on the graphics hardware of your notebook, unless you don’t game at all, in which case you should at least look at an integrated part from ATI or nVidia. Gamers would do well to only buy Shader Model 3 parts at this point, and follow my recommendations.

As always, with questions, please post to the forums. Alternatively, we’re going to be piloting a Q&A column here, so you can also ask me directly and your question will likely appear in that column in the near future.

Other Resources:

• Graphics Mobile Card Info Page (by Charles Jefferies)
• Updating Mobile Video Card Drivers (by Charles Jefferies)
• Notebook Video Card Guide 2005(by Dustin Sklavos)
• NBR Gaming forum for hardware and software