Some Notes About Our Recommendations

• This list is for gamers who want to get the most for their money. If you don't play games, then the cards on this list are more expensive than what you really need. We've added a reference page at the end of the column covering integrated graphics processors, which is likely more apropos.
• The criteria to get on this list are strictly price/performance. We acknowledge that recommendations for multiple video cards, such as two Radeon cards in CrossFire mode or two GeForce cards in SLI, typically require a motherboard that supports CrossFire or SLI and a chassis with more space to install multiple graphics cards. They also require a beefier power supply compared to what a single card needs, and will almost certainly produce more heat than a single card. Keep these factors in mind when making your purchasing decision. In most cases, if we have recommended a multiple-card solution, we try to recommend a single-card honorable mention at a comparable price point for those who find multi-card setups undesirable.
• Prices and availability change on a daily basis. We can't base our decisions on always-changing pricing information, but we can list some good cards that you probably won't regret buying at the price ranges we suggest, along with real-time prices from our PriceGrabber engine, for your reference.
• The list is based on some of the best U.S. prices from online retailers. In other countries or at retail stores, your mileage will most certainly vary.
• These are new card prices. No used or open-box cards are in the list; they might represent a good deal, but it's outside the scope of what we're trying to do.

Best PCIe Card: Under $85

Best PCI Express (PCIe) Card For ~$50:

Radeon HD 4650 (Check Prices)

Great 1280x1024 performance in most games, 1680x1050 with lowered detail

You will not find a card that packs more punch than ATI's Radeon HD 4650 under the alluring $50 price point. With solid stock performance and an overclockable GPU, this card is an excellent starting point for our recommendations, and a wholly worthwhile upgrade if you're currently stuck using a motherboard with integrated graphics.

Best PCI Express (PCIe) Card For ~$65: Tie

Radeon HD 4670 (Check Prices)

Good 1680x1050 performance in most games

With the release of Nvidia's GeForce GT 240, ATI's Radeon HD 4670 is no longer the most powerful reference card without a dedicated power connector. However, it remains a compelling solution under the $75 price point, which Nvidia's solution simply hasn't hit yet.

Performance is excellent and power usage is very low, making this product an impressive performer all-around. Its accelerated clock rates and modestly-higher price tag are worth considering if you originally had your eye on the Radeon HD 4650.

GeForce 9600 GSO (Check Prices)

Good 1680x1050 performance in most games

The GeForce 9600 GSO is seems to be getting quite hard to find, and is likely being end-of-life'd soon in favor of the new GeForce GT 240. Nevertheless, as long as it is available, the GeForce 9600 GSO remains a powerful competitor compared to the Radeon HD 4670. While the GeForce requires a dedicated PCIe power connector to supply more juice than the Radeon, it does offer better performance in some situations.

Best PCIe Card For ~$85:

GeForce 9600 GT (Check Prices)

Good 1680x1050 performance in most games

The GeForce 9600 GT is a great performer, thanks in part to its high-end 256-bit memory interface and speedy DDR3 memory. It's a great choice on an $85 budget, even if the architecture on which it centers is showing its age.

Certainly, this card's continued presence here is a testament to Nvidia's engineering work dating back almost two years ago. With that said, we'd certainly like to see the company's latest DirectX 10.1 cards drop in price to compete against ATI's strong offerings.

Best PCIe Card: ~$90 To $140

Best PCIe Card For ~$95: Tie

GeForce 9800 GT (Check Prices)

Exceptional 1680x1050 performance in most games, 1920x1200 in most games with lowered detail

The GeForce 9800 GT is essentially a rebadged GeForce 8800 GT, and offers the same great performance it has for years now (that sure sounds funny to say in reference to graphics cards).

With the rising price of ATI's Radeon HD 4850 giving it space to breathe, this legendary card is once again a recommended buy. But once again, we're looking forward to seeing technological progress put new, faster, and cooler products loaded down with more features in this space rather than revisiting history.

Fortunately, there's still PhysX and 3D Vision support to like about this aging board.

Radeon HD 4830 512MB (Check Prices)

Exceptional 1680x1050 performance in most games, 1920x1200 in most games with lowered detail

Just as the GeForce 9800 GT can once again be recommended due to the rising price of the Radeon HD 4850, so can the Radeon HD 4830. While availability is low, this Radeon is still a viable option under the $100 price point if you can find it. You'll discovered that it offers great performance on par with the GeForce 9800 GT, with the added benefit of DirectX 10.1 support.

Best PCIe Card For ~$110:

GeForce GTS 250 512MB (Check Prices)

Good 1920x1200 performance in most games

The dissapearance of the $100 Radeon HD 4850 has not only opened up the GeForce 9800 GT and Radeon HD 4830 for recommended status, but also the GeForce GTS 250.

At $110, the 512MB version of this card offers respectable performance, and nothing else in the price range can compare to it. As fast as the Radeon HD 4850 and new Radeon HD 5750 (and notably cheaper), the GeForce GTS 250 has no real competition from the rest of the sub-$150 market at this time.

Bear in mind that going this route instead of the Radeon HD 5750 will cost you DirectX 11 support and Eyefinity. But in the context of gaming, you'll need to make other quality sacrifices long before trying to enjoy either value-add in the $110 range.

Best PCIe Card For ~$120:

GeForce GTS 250 1GB (Check Prices)

Good 1920x1200 performance in most games

For $10 more than the 512MB version, an interested gamer can get the benefit of a full gigabyte of memory. At the highest resolutions and levels of anti-aliasing, this extra memory might provide a performance boost, though it's unlikely the GeForce GTS 250 is powerful enough to run at those detail levels. Still, many buyers might find the slight $10 price increase worthwhile in something like Grand Theft Auto IV.

Best PCIe Card: ~$150 To $290

Best PCIe Card For ~$155: Tie

Radeon HD 5770 (Check Prices)

Great 1920x1200 performance in most games

While the new Radeon HD 5770 isn't any faster than its older Radeon HD 4870 cousin (we've found that it's even slightly slower in many instances), it does have something the Radeon HD 4870 doesn't have: full DirectX 11 and Eyefinity support. Indeed, while the Radeon HD 5770 doesn't run away with any performance crowns in this category, it does look good from a longevity/value standpoint.

Read our full review of ATI's Radeon HD 5770 for more information on the card and its accompanying architecture.

GeForce GTX 260 (Check Prices)

Great 1920x1200 performance in most games

Like many cards, the GeForce GTX 260 is becoming very hard to find, and may soon be end-of-life'd. In any case, it does offer advantages in titles that run better on Nvidia's GT200 architecture, and it sports some GeForce-only value-added features like PhysX compatibility and support for GeForce 3D Vision.

Once again, a little diligence is required on the part of the buyer to find out which card is best adapted for his or her favorite titles, and whether or not your motherboard supports SLI, CrossFire, or both multi-card technologies.

Best PCIe Card For ~$200:

Radeon HD 4890 (Check Prices)

Excellent 1920x1200 performance in most games

The Radeon HD 4890 is essentially an overclocked Radeon HD 4870. However, the tweaks that AMD made to the newer RV790 die result in much higher overclocking headroom. At stock speeds, this card is worth the $200. But to get the most out of it, some overclocking is in order. And now that the prices on Radeon HD 5850 cards are through the roof, there's not much between this board and ATI's next-fastest solution.

Read our full review of ATI's Radeon HD 4890 for more information on the card and its accompanying architecture.

Best PCIe Card For ~$240:

2 x GeForce GTS 250 1GB in SLI Configuration (Check Prices)

Exceptional 1920x1200 performance, 2560x1600 in most games with lowered detail

Two GeForce GTS 250 cards in SLI pack a punch and make a strong case for multi-card setups. With the Radeon HD 4850s going up in price and down in availability, these GeForce cards replace them as the weapon of choice for sub-$300 brute force power.

Best PCIe Card: ~$300 To $400

Best PCIe Card For ~$310: None

Honorable Mention: Radeon HD 5850

Exceptional 1920x1200 performance, 2560x1600 in most titles

The new Radeon HD 5850 has some definite advantages over a pair of GeForce GTX 260s in SLI or a pair of Radeon HD 5770s in CrossFire. It doesn't need a CrossFire-compatible motherboard, it uses a lot less power thanks to its 40nm manufacturing process, and it sports DirectX 11 capabilities (plus Eyefinity).

Unfortunately, scant availability forces us to relegate the Radeon HD 5850 to Honorable Mention status until it can be purchased without having to fight for it.

Read our full review of ATI's Radeon HD 5850 for more information on the card and its accompanying architecture.

Best PCIe Card For ~$330: Tie

At approximately the same price, these options retain the same advantages of their single-card counterparts: two GeForce GTX 260s offer SLI compatibility, PhysX, and GeForce 3D Vision support, and two Radeon HD 5770s offer DirectX 11, Eyefinity, and high-def audio bitstreaming to CrossFire-compatible motherboard users. A good case can be made for either of these options, and none of them are a poor choice. Just pick your poison.

2 x GeForce GTX 260 in SLI (Check Prices)

Exceptional 1920x1200 performance, good 2560x1600 performance in most titles

Nvidia doesn't have a DirectX 11-class architecture yet, so if you're going to sink $300+ into DirectX 10 hardware, do so knowing that there are competing DirectX 11 boards available in the same price range.

A pair of Radeon HD 5770s won't offer quite the same level of performance as two GeForce GTX 260s in SLI; that's the trade-off for more modern functionality, though.

2 x Radeon HD 5770 in CrossFire (Check Prices)

Exceptional 1920x1200 performance, good 2560x1600 performance in most titles

The new Radeon HD 5770 sports added benefits over its GeForce GTX 260 competition: DirectX 11, triple display outputs, and the ability to bitstream high-definition audio content from Blu-ray movies contribute significant value to ATI's newest mainstream graphics cards. For those seeking long-term DirectX 11 compatibility, this might be the more attractive option.

Read our full review of ATI's Radeon HD 5770 for more information on the card and its accompanying architecture.

Best PCIe Card For ~$400:

Two Radeon HD 4890 cards in CrossFire Configuration (Check Prices)

Good 2560x1600 performance in most games

Two Radeon HD 4890 cards should, on average, perform on par or better than a single GeForce GTX 295, and will definitely beat out a single Radeon HD 5870. Plus, these Radeons cost less than either option. If you have a CrossFire-compatible motherboard and want some serious performance at high resolutions, this route is the way to go.

Read our full review of ATI's Radeon HD 4890 for more information on the card and its accompanying architecture.

Best PCIe Card: ~$400 And Up

Best PCIe Card For ~$410: None

Honorable Mention: Radeon HD 5870

Good 2560x1600 performance in most games

For $10 less, a couple Radeon HD 4890s will easily beat a single Radeon HD 5870 in the titles that matter today (perhaps this will change when DirectX 11 software becomes more pervasive). From a raw price/performance standpoint, this makes the Radeon HD 5870 a hard sell. But that is not to say this card is underpowered: it sports the fastest single GPU on the planet, relatively low power usage (remarkably low at idle), and DirectX 11 support. For folks without a motherboard that supports CrossFire and a hefty power supply, the new Radeon HD 5870 is definitely a more-than-viable option. Unfortunately, availability is still quite rare.

Read our full review of ATI's Radeon HD 5870 for more information on the card and its accompanying architecture.

Best PCIe Card For ~$465: None

Honorable Mention: GeForce GTX 295 (Check Prices)

Good 2560x1600 performance in most games

Despite ATI's new Radeon HD 5970 taking its place as the fastest graphics card on the planet, Nvidia's GeForce GTX 295 (with SLI-on-a-board) remains an extremely powerful graphics card. Essentially two conjoined GeForce GTX 275s, the GeForce GTX 295 offers very notable gains over a single Radeon HD 5870 in the great majority of game titles, although the Radeon will use far less power doing so. The GeForce GTX 295 does have an advantage in that it it still quite easy to find and purchase. Moreover, ATI's release has forced prices on these cards down by a significant chunk.

Read our full review of Nvidia's GeForce GTX 295 for more information on the card and its accompanying architecture.

Best PCIe Card For ~$625: None

Honorable Mention: Radeon HD 5970

Great 2560x1600 performance

3,200 shader processors. There isn't much more we need to say about the brutal grace of execution that characterizes the world's fastest graphics card, the Radeon HD 5970. With two Radeon HD 5870 GPUs onboard, the only things we can complain about are scant availability and an extremely high price tag. Availability should improve over time; the price not so much, but if you're in the market for this card price probably isn't an issue.

Read our full review of ATI's Radeon HD 5970 for more information on the card and its accompanying architecture.

There you have it folks; the best cards for the money this month. Now all that's left to do is to find and purchase them.

Don't worry too much about which brand you choose, because all of the cards out there are close to Nvidia's and ATI's reference designs. Just pay attention to price, warranty, and the manufacturer's reputation for honoring the warranty if something goes wrong.

Also remember that the stores don't follow this list. Things will change over the course of the month and you'll probably have to adapt your buying strategy to deal with fluctuating prices. Good luck!

And you have to figure that's pretty far behind, since the Larrabee launch timeframe was 2009/2010. The only way you'll be able to touch Larrabee now is as a development platform for graphics engines or high-performance computing, in order to develop for future Intel products.

Intel says they're going to announce new plans for discrete cards some time in 2010—mayyybe CES, where we talked to former Intel Chairman Craig Barrett about Larrabee last year? But, more likely at the Intel Developer Forum later in the year. [Intel]

MaximumPC put the 5970 to the test (check out their review here), and found that it lives up to its promise. The 2GB dual-GPU card is the first to support DirectX 11, and basically doubles its wholly respectable predecessor (the 5870) in specs, capable of delivering nearly 5 teraflops of raw processing power. It's a massive card, about a foot long, designed mostly for heat dissipation, at which aim it apparently succeeds. It's also got easy access to overclocking via AMD's OverDrive, and can drive up to three displays simultaneously with a maximum resolution of 7680x1600. So it's pretty much the greatest thing ever, and it's got a pricetag to match: $600 upon its undisclosed release. Yow. [MaximumPC]

Insane Graphics Card Packs 2GB Memory and Six (Yes, Six) DisplayPort Ports

Rumor has it that AMD is set to announce a new, and 100% ridiculous, flagship card: The Radeon HD 5870. The 5870, if the rumors are correct, will have 2GB of GDDR5 memory and a whopping six DisplayPort ports, not to mention dual DVI ports and an HDMI. Clock speed and GPU speed are relatively normal at 4800MHz and 850MHz, respectively. It'll require two PCI-E slots, which is reasonable considering the spec excess. It ends up here in Remainders because the only source we've seen is a post on Softpedia, and while it's certainly crazy, we're not sure how big the market is for a six-port graphics card. [Softpedia via Tom's Hardware via Crunchgear]

Volcano Prevention Efforts May Result in Volcano Eruption

Geologists working in Campi Flegrei, a caldera in Naples, Italy, are concerned about the volcano's possibility of eruption, so understandably they're doing a little drilling to figure out exactly what the risk is. Unfortunately, the seven 4-km holes they have to drill could have exactly the opposite effect from preventing an eruption—there's a distinct risk that the drilling itself could trigger an eruption, which could well wipe out the entirety of urban Naples. If they hit magma, the sudden rise in temperature would vaporize their drilling liquid, causing an explosion which could in turn trigger an eruption. It's a sexy and dangerous headline, but it winds up in Remainders because, well, the risk doesn't seem all that great: 4km wouldn't even be halfway deep enough to reach any known reservoirs of magma, so it's probably a moot point. [New Scientist via Pop Sci]

Makeshift Bike Rack Attaches to Street Signs

The CycleHook is a dual-looped piece of metal that locks securely onto pretty much any signpost, creating two sweet parking spots for bicyclists. It's a pretty good idea—I'm a cyclist, and I can confirm that most cities don't provide enough spots to lock up a bike, and sometimes traffic cops get snippy when you chain up to a signpost on the sidewalk. The problem with the CycleHook is that it's not new. Montreal, for one, has had this exact design attached to parking meters for years, and I'm sure it's not the only city to have the same thought. On the other hand, Montreal's parking spots aren't pretty pink. [Wired]

PC Built Inside a Classy-as-Hell Leather and Wood Chest

We love PC case hacks. Anything you can shove a PC into, please do it, and send us pictures. There are several inviting spaces that haven't yet been used—and this here is one of them. Rob Higardea crammed a PC into this beautiful, vintage wood and leather chest. Yeah, it's got wires hanging out the back, but from the front? Pure class. I'd love one of my own, except that much class would probably stick out in my apartment more than a slab of anonymous grey plastic. [UnPluggd]

At more than a foot long, this card requires eight- and six-pin power connectors and a equally huge rig to live in. Seriously, the folks at Alienbabel Tech said they had issues fitting this card into a full size Antec 1200 chassis. Luckily, the card they had was apparently an engineering prototype, and may shrink down slightly for retail.

When we brought you word of this card in September, pricing and release date were unknown, and that hasn't changed today. Just some additional pictures for you PC gamers to drool over as you fantasize about that six 30-inch monitor setup. [Alienbabel Tech via PC Perspective via Engadget]

DirectX 67? Tongue-finity? My favorite quote: "Pulling a trigger is just as easy as pushing a button, only a lot more fun and healthy...You say babies, we say bootcamp" Nice.

There are three "commercials" in the playlist below, with the most interesting one loaded up first. A little too self-consciously viral, but too interesting not to share. [XFX 2118AD] Thanks Milo!

First, the Windows 7 stuff. The ability to add any old CableCARD tuner to your setup is realized, as we just shoved on a tuner and Windows 7 automatically recognized it and installed the correct drivers. The setup with Comcast was fairly simple, as all we had to do was run through the Tuner setup program, and all Comcast had to do was read some numbers back to home base and have them activate the CableCARD.

The copy freely features work just as you'd imagine. There's not much to "show", in that all it enables is the ability to watch your recorded shows on any other PC. So we copied a recording off of the Zone Pro, onto another Windows 7 PC, and it played back fine.

To enable this, you need Windows 7 and you need to upgrade your firmware on your ATI Tuner—the latter will be available for free from ATI and other OEMs, depending on where you bought your current tuner. The new format for recorded TV, .WTV, is also key, so if you recorded something on a Vista machine, you wouldn't be able to share it with other computers on your network (obviously). Eventually there will be third-party transcoders that can take the .WTV file and make it capable of being read on Zunes and iPhones. And with the Zune HD, you'll be able to handle HD recordings as well.

As for Windows 7 media center itself, the Zone Pro handles it just fine. All the effects are there, such as transparent overlay, are there, and the box can handle two simultaneous tuners at once. Pretty amazing, seeing as the size of one of the external tuners is basically the same size of the Zone Pro itself.

Niveus will start shipping the Zone Pro with Windows 7 to arrive by Win 7's launch. [Niveus]

They Call It "Open" For a Reason
One of the more excellent aspects of Snow Leopard, actually, is its full-scale deployment of OpenCL 1.0—Open Computing Language—a framework that allows programmers to more easily utilize the full power of mixes of different kinds of processors like GPUs and multi-core CPUs. (Much of the excitement for that is in leveraging the GPU for non-graphical applications.)

OpenCL lives up to its name: It is a royalty-free open standard managed by the Khronos Group, and supported by AMD/ATI, Apple, ARM, IBM, Intel, Nvidia, among others. Interesting thing about this open industry standard is that it was developed and proposed by... Apple.

What Is a Standard?
By "standard," we're talking about a format, interface or programming framework that a bunch of companies or people or organizations agree is the way something's going to get done, whether it's how a movie is encoded or the way websites are programmed. Otherwise, nothing works. A video that plays on one computer won't play on another, web sites that work in one browser don't work in another, etc. With increased connectedness between different machines and different platforms, standards are increasingly vital to progress.

Standards can range from open (anybody can use them, for free) to open with conditions (anybody can use them as long they follow conditions X, Y and Z) to closed (you gotta have permission, and most likely, pay for it). Some companies view standards strictly as royalty machines; others don't make much money on them, instead using them to make sure developers do things the way they want them to. Apple falls into this latter category, by choice or possibly just by fate.

Kicking the Big Guy in the Shins
Of course, OpenCL isn't the only open standard that Apple's had a hand in creating or supporting that actually went industry-wide. When you're the little guy—as Apple was, and still is in computer OS marketshare, with under 10 percent—having a hand in larger industry standards is important. It keeps your platform and programming goals from getting steamrolled by, say, the de facto "standards" enforced by the bigger guy who grips 90 percent of the market.

If you succeed in creating a standard, you're making everybody else do things the way you want them done. If you're doubting how important standards are, look no further than the old Sony throwing a new one at the wall every week hoping it'll stick. Or Microsoft getting basically everybody but iTunes to use its PlaysForSure DRM a couple years ago. Or its alternative codecs and formats for basically every genuine industry standard out there. To be sure, there is money to be made in standards, but only if the standard is adopted—and royalties can be collected.

Web Standards: The Big Headache
The web has always been a sore spot in the standards debate. The web is a "universal OS," or whatever the cloud-crazy pundits call it, but what shapes your experience is your browser and in part, how compliant it is with the tools web developers use to build their products. Internet Exploder shit all over standards for years, and web programmers still want IE6 to die in a fiery eternal abyss.

Enter WebKit, an open source browser engine developed by Apple based off of the KHTML engine. It's so standards-compliant it tied with Opera's Presto engine to be the first to pass the Acid3 test. What's most striking about WebKit isn't the fact it powers Safari and Google Chrome on the desktop, but basically every full-fledged smartphone browser: iPhone, Android, Palm Pre, Symbian and (probably) BlackBerry. So WebKit hasn't just driven web standards through its strict adherence to them, but it has essentially defined, for now, the way the "real internet" is viewed on mobile devices. All of the crazy cool web programming you see now made is made possible by standards-compliant browsers.

True, OpenCL and WebKit are open source—Apple's been clever about the way it uses open source, look no further than the guts of OS X—but Apple is hardly devoted to the whole "free and open" thing, even when it comes to web standards.

All the AV Codecs You Can Eat
The recent debate over video in the next web standards, known collectively as HTML5, shows that: Mozilla supports the open-source Ogg Theora video codec, but Apple says it's too crappy to become the web's default video standard—freeing everyone from the tyranny of Adobe's Flash. Apple says Ogg's quality and hardware acceleration support don't match up to the Apple-supported MPEG-4 standardized H.264 codec, which is tied up by license issues that keep it from being freely distributed and open. (Google is playing it up the middle for the moment: While it has doubts about the performance of Ogg Theora, Chrome has built-in support for it and H.264.)

Apple has actually always been a booster of MPEG's H.264 codec, which is the default video format supported by the iPhone—part of the reason YouTube re-encoded all of its videos, actually—and gets hardware acceleration in QuickTime X with Snow Leopard. H.264 is basically becoming the video codec (it's in Blu-ray, people use it for streaming, etc.).

Why would Apple care? It means Microsoft's WMV didn't become the leading standard.

A sorta similar story with AAC, another MPEG standard. It's actually the successor to MP3, with better compression quality—and no royalties—but Apple had the largest role in making it mainstream by making it their preferred audio format for the iPod and iTunes Store. (It saw some limited use in portables a little earlier, but it didn't become basically mandatory for audio players to support it until after the iPod.) Another bonus, besides AAC's superiority to MP3: Microsoft's WMA, though popular for a while, never took over.

FireWire I Mean iLINK I Mean IEEE 1394
Speaking of the early days of the iPod, we can't leave out FireWire, aka IEEE 1394. Like OpenCL, Apple did a lot of the initial development work (Sony, IBM and others did a lot of work on it as well), presented it to a larger standards body—the Institute of Electrical and Electronics Engineers—and it became the basis for a standard. They tried to charge a royalty for it at first, but that didn't work out. It's a successful standard in a lot of ways—I mean, it is still on a lot of stuff like hard drives and camcorders still—but USB has turned out to be more universal, despite being technically inferior. (At least until USB 3.0 comes out, hooray!)

Update: Oops, forgot Mini DisplayPort, Apple's shrunken take on DisplayPort—a royalty-free video interface standard from VESA that's also notably supported by Dell—which'll be part of the official DisplayPort 1.2 spec. Apple licenses it for no fee, unless you sue Apple for patent infringement, which is a liiiiittle dicey. (On the other hand, we don't see it going too far as industry standard, which is why we forgot about it.)

That's just a relatively quick overview of some of the standards Apple's had a hand in one way or another, but it should give you an idea about how important standards are, and how a company with a relatively small marketshare (at least, in certain markets) can use them wield a lot of influence over a much broader domain.

Shaping standards isn't always for royalty checks or dominance—Apple's position doesn't allow them to be particularly greedy when it comes to determining how you watch stuff or browse the internet broadly. They've actually made things better, at least so far. But, one glance at the iPhone app approval process should give anybody who thinks they're the most gracious tech company second thoughts about that.

Still something you wanna know? Send questions about standards, things that are open other than your mom's legs or Sony Ultra Memory Stick XC Duo Quadro Micro Pro II to tips@gizmodo.com, with "Giz Explains" in the subject line.

ATI's first, with the RV870-based Radeon HD 5800 series shipping out next month (no surprise, since they were showing it off a couple months ago), while Nvidia's following with the GT300 series that'll apparently hit in December, according to DigiTimes' sources. On the other hand, Nvidia seems to have the lead on the actual Windows 7 front, since their GPUs are already Windows Hardware Qualification Lab-certified with support for the new DirectCompute API. Bonus: Your existing Nvidia graphics works with it, if it ain't ancient.

Then again, there's no excuse like a new operating system for buying hundreds of dollars in new computer gear. [DigiTimes]

It actually boils down fairly simply to a mixed bag: ATI's Stream tends to be outright faster and pulls more of the load off of the CPU, but Nvidia's CUDA tends to produce better quality results. Interestingly, PC Perspective seems to like ATI's Avivo video transcoding application more than they used to, saying they're impressed by its simplicity. But which side are you on? [PC Perspective via Engadget]

The MSI R4890 Cyclone SOC, which shares as much aesthetically with a motorcycle engine as it does a PC component, is a perfect example of everything that is confusing, irritating and unnecessarily obfuscated with graphics cards today.

There's a lot of impressive tech here—an overclocked 1GHz ATI Radeon HD 4890 GPU and 1GB of GDDR5 RAM— but instead of putting that into meaningful terms, MSI is content to just throw the numbers out there and brag about the card's "200-year lifespan" and "largest ever" fan. I'm sure this is fun for the hardest of the hardcore enthusiasts, but for the rest of us it's just a headache. The card should ship any day now, priced somewhere in the ballpark of $250. [Slashgear]

Whether you're buying a new computer, building your own or upgrading an old one, the process of choosing a new graphics card can be daunting. Integrated graphics solutions—the kind that come standard with many PCs—have trouble playing games from three years ago, let alone today, and will put you at a disadvantage when future technologies like GPGPU computing, which essentially uses your graphics card as an additional processor, finally take hold. On top of all this, we're in the middle of a price dip—it's objectively a great time to buy. (Assuming you're settled on a desktop. Ahem.) The point is, you'll want to make the right choice. But how?

Set Specific Goals, Sight Unseen
Your first step to finding the right graphics card is to just step back. Just as graphics card specs are nigh-on impossible to understand, naming conventions and marketing materials will do nothing except give you a headache. The endlessly higher numerical names, the overlapping product lines, the misleadingly-named chip technologies—just leave them. For now, pretend they don't exist.

Now, choose your goals. What games do you want to play? What video output options and ports do you want? What resolution will you be playing your games at? Do you have any use for the fledgling GPGPU technologies that are slowly permeating the marketplace? And although you may have to adjust this, set a price goal. Ready-built PC buyers will have to consider whatever upgrade cost your chosen company is charging, and adjust accordingly. For people upgrading their own systems, $150-$200 has been something of a sweet spot: It'll get you a card with a new enough GPU, and sufficient VRAM to handily deal with mainstream games for a solid two years. If you want to spend less, you can; if you want to spend more, fine.

These are the terms that matter most. Seriously, disregard any allegiance to Nvidia or ATI, prior experiences with years-old graphics hardware or some heretofore distant, unreleased and unspec'd game franchise. Be decisive about what you want, but as far as hardware and marketing materials go, start blind.

Don't Get Caught Up In Specs
Now that you've laid out your ambitions, as modest or extreme as they may be, it's time to dive into the seething, disorienting pool of hardware that you'll be choosing from. The selection, as you'll find out, is daunting. The first layer of complexity comes from the big two—Nvidia and ATI—whose product lines read more like Terminator robot taxonomies than something generated by humans. Here's Nvidia's desktop product line, right now:

It seems like you ought to be able glean a linear progression of performance (or at least price) out of that alphanumeric pile, right? Not at all. How in the world are we to know that the 9800GTX is generally more powerful than the GTS 250, or that the 8800GTS trumps a 9600GT? A two letter suffix can mean more than a model number, and likewise, a model number can mean more than membership in a product line. These naming conventions change every couple years, and occasionally even get traded between companies. For example, I've personally owned two graphics cards that bore 9x00 names—you just won't see them on the chart above, because they were made by ATI. Point is: You don't need to bother with this nonsense.

The next layer of awfulness comes from the sundry OEMs that rebrand, tweak and come up with elaborate ways to cool offerings from the big two. This is what Sapphire, EVGA, HIS, Sparkle, Zotac and any number of other inanely named companies do. They can, on occasion, cause some sizable changes to the performance of the GPUs they're built around, but by and large, the Nvidia or ATI label on the box is still the best indication of what to expect from the product, i.e., a Zotax Gtx285 won't be that much better or worse than an eVGA or stock model. You'll get a different fan/heatsink configuration, different hardware styling, and possibly different memory or GPU frequency specs, but the most important difference—and the only one you should really concern yourself with—is price.

Graphics cards' last, least penetrable line of defense against your comprehension is hardware jargon. Bizarre, unhelpful spec sheets are, and always have been, a common feature in PC hardware, from RAM (DDR3-1600!) to processors (12 MB L2 cache! 1333MHz FSB!).

The image associated with this post is best viewed using a browser.Graphics cards are worse. Each one has three MHz-measured speeds you'll see advertised—the core clock, the CPU (shader) clock and the memory frequency. VRAM—the amount of dedicated memory your card has to work with—is another touted specification, ranging from 256MB to well beyond the 1GB barrier for gaming cards. On top of frequency, memory introduces a whole slew of additional confusing numbers: memory type (as in, DDR2 or DDR3); interface width (in bits, the higher the better); and memory bandwidth, nowadays measured in GB/s. And increasingly, you'll see processor core numbers trotted out. Did you know that Nvidia's top-line card has 480 of them? No? Good.

The best way to approach these numbers is to ignore them. Sure, they provide comparative evaluation and yes, they do actually mean something, but unless you're a bonafide graphics card enthusiast, you won't be able to look at a single spec—or a whole spec sheet—and come to any useful conclusions about the cards. Think of it like cars: horsepower, torque and engine displacement are all real things. They just demand context before they can be taken to mean anything to the driver. That's why road tests carry so much weight.

Graphics cards have their own road testers, and they've got the only numbers you need to worry about.

The image associated with this post is best viewed using a browser.Respect the Bench, or Trust the Experts
In the absence of meaningful specs, names or distinguishing features, we're left with benchmarks. This is a good thing! For years, sites like Tom's Hardware, Maximum PC, and Anandtech have tirelessly run nearly every new piece of graphics hardware through a battery of tests, providing the buying public with comparative measures of real-word performance. These are the only numbers you need to bother yourself with, and where those goals you settled on come into play.

Here's how to apply them. Say you just really want to play Left 4 Dead, and have about a hundred dollars to spend. Navigate over to Tom's, check their benchmarks for that particular game, and scroll down the list. You're looking for a card that is a) an option on whatever system you're buying and b) can handle the game well—at a high resolution and high texture quality—which, generally speaking, is a comfortable 60 frames per second. Find the card, check the price and you're practically done. Once you've zeroed in on a card based on your narrow criteria, expand outward. You can check out more games benchmarks and seek out standalone reviews, which will enlighten you on other, less obvious considerations, like fan noise, power draw and reported reliability. (Note: resources for notebook users are a little more sparse. That said, Notebook Check [click the British flag for English] does good work.]

From there, your next worry will be buying for the future. You shouldn't buy the bare minimum hardware for the current generation of games—there's no need to spring for a card that'll be obsolete within a few months, no matter how cheap it is. But buying the latest, greatest dual-GPU graphics cards is an equally bad value proposition. As generations of video hardware have come and gone, one thing has remained constant: A company's midrange offerings, usually pegged at about $150-$200, are your best bet, period. Sometimes they'll be new products, and sometimes they'll have been around a while. What you'll be buying, basically, is the top end of the last generation. This is fine, and will keep the vast majority of users happy for the lifecycle of their PC. Those of you who live on the bleeding edge probably don't need this guide anyway.

Your alternative route is to just trust the experts. Sites like Ars Technica and Maximum PC regularly assemble system guides at various pricepoints, in which they've made your value judgments for you. Tom's even assembles a "Best Cards for the Money" guide each month, which is invaluable. At given price points, the answer will often be obvious, and these guys know what they're talking about.

But keep in mind, they're applying the same formula you can, just with a slightly more knowing eye. The matter truly is as simple as broadly deciding what you need, consulting the right sources and floating far enough above the spec-ravaged landscape so as to avoid getting a headache. Good luck.

What a Difference a Letter Makes
GPU sounds—and looks—a lot like CPU, but they're pretty different, and not just 'cause dedicated GPUs like the Radeon HD 4870 here can be massive. GPU stands for graphics processing unit, while CPU stands for central processing unit. Spelled out, you can already see the big differences between the two, but it takes some experts from Nvidia and AMD/ATI to get to the heart of what makes them so distinct.

Traditionally, a GPU does basically one thing, speed up the processing of image data that you end up seeing on your screen. As AMD Stream Computing Director Patricia Harrell told me, they're essentially chains of special purpose hardware designed to accelerate each stage of the geometry pipeline, the process of matching image data or a computer model to the pixels on your screen.

GPUs have a pretty long history—you could go all the way back to the Commodore Amiga, if you wanted to—but we're going to stick to the fairly present. That is, the last 10 years, when Nvidia's Sanford Russell says GPUs starting adding cores to distribute the workload across multiple cores. See, graphics calculations—the calculations needed to figure out what pixels to display your screen as you snipe someone's head off in Team Fortress 2—are particularly suited to being handled in parallel.

An example Nvidia's Russell gave to think about the difference between a traditional CPU and a GPU is this: If you were looking for a word in a book, and handed the task to a CPU, it would start at page 1 and read it all the way to the end, because it's a "serial" processor. It would be fast, but would take time because it has to go in order. A GPU, which is a "parallel" processor, "would tear [the book] into a thousand pieces" and read it all at the same time. Even if each individual word is read more slowly, the book may be read in its entirety quicker, because words are read simultaneously.

All those cores in a GPU—800 stream processors in ATI's Radeon 4870—make it really good at performing the same calculation over and over on a whole bunch of data. (Hence a common GPU spec is flops, or floating point operations per second, measured in current hardware in terms of gigaflops and teraflops.) The general-purpose CPU is better at some stuff though, as AMD's Harrell said: general programming, accessing memory randomly, executing steps in order, everyday stuff. It's true, though, that CPUs are sprouting cores, looking more and more like GPUs in some respects, as retiring Intel Chairman Craig Barrett told me.

Explosions Are Cool, But Where's the General Part?
Okay, so the thing about parallel processing—using tons of cores to break stuff up and crunch it all at once—is that applications have to be programmed to take advantage of it. It's not easy, which is why Intel at this point hires more software engineers than hardware ones. So even if the hardware's there, you still need the software to get there, and it's a whole different kind of programming.

Which brings us to OpenCL (Open Computing Language) and, to a lesser extent, CUDA. They're frameworks that make it way easier to use graphics cards for kinds of computing that aren't related to making zombie guts fly in Left 4 Dead. OpenCL is the "open standard for parallel programming of heterogeneous systems" standardized by the Khronos Group—AMD, Apple, IBM, Intel, Nvidia, Samsung and a bunch of others are involved, so it's pretty much an industry-wide thing. In semi-English, it's a cross-platform standard for parallel programming across different kinds of hardware—using both CPU and GPU—that anyone can use for free. CUDA is Nvidia's own architecture for parallel programming on its graphics cards.

OpenCL is a big part of Snow Leopard. Windows 7 will use some graphics card acceleration too (though we're really looking forward to DirectX 11). So graphics card acceleration is going to be a big part of future OSes.

So Uh, What's It Going to Do for Me?
Parallel processing is pretty great for scientists. But what about those regular people? Does it make their stuff go faster. Not everything, and to start, it's not going too far from graphics, since that's still the easiest to parallelize. But converting, decoding and creating videos—stuff you're probably using now more than you did a couple years ago—will improve dramatically soon. Say bye-bye 20-minute renders. Ditto for image editing; there'll be less waiting for effects to propagate with giant images (Photoshop CS4 already uses GPU acceleration). In gaming, beyond straight-up graphical improvements, physics engines can get more complicated and realistic.

If you're just Twittering or checking email, no, GPGPU computing is not going to melt your stone-cold face. But anyone with anything cool on their computer is going to feel the melt eventually.

[Back to our Complete Guide to Snow Leopard]

So that's a bit of a cheat: In theory, you can buy the 850MHz version and overclock it yourself without paying the premium for the 1GHz version. Or maybe you can buy the factory-overclocked 1GHz version and overclock it again so your PC melts and reaches the Earth's core. [Hot Hardware]

• AMD tried to kill the megahertz myth before Intel. During the Pentium 4 days Intel kept pushing clock speeds higher and higher, before it hit a wall and abandoned the Prescott architecture. The message was clearly, "more megahertz is more better." AMD's competing Athlon XP chips, while clocked slower, often beat their Pentium 4 rivals. Ironically, AMD was the first to 1GHz, as some commenters have pointed out (don't know how I forgot that). Obviously though, AMD's performance lead didn't last forever.

• AMD beat Intel to 64-bit in mainstream computers. And we're not just talking about its Opteron and Athlon 64 processors. AMD actually designed the X86-64 specification, which Intel wound up adopting and licensing—so AMD's spec is used Intel's 64-bit processors to this day.

• AMD was first to consider energy efficiency in processor designs. Okay, this is kind of an extension of point number one, but during Intel's Pentium 4 'roid rage period AMD's processors consistently used less power than Intel's. Intel's performance per watt revelation didn't really start until the Pentium M (which was actually a throwback to the P6 architecture), which set the tone for Intel's new direction in its successor, the Core line of chips.

• AMD beat Intel to having an integrated memory controller. A tech feature AMD lorded over Intel for years: AMD's processors started integrating the memory controller with its processors years ago, reducing memory latency. Intel's first chip to use an integrated memory controller is the Core i7—before, the memory controller was separate from the processor. (Here's why Intel says they held off.)

Athlon XP and Athlon 64—those were the good old days, AMD's cutthroat competitive days. The days they were ahead of Intel. I miss them—at one point, every hand-built computer in my house ran AMD processors. I felt like a rebel—a rebel with faster, cheaper computers.

Unfortunately, I don't run AMD chips anymore. Intel came back, and came back hard. But here's hoping for another resurgence, and another 40 years, guys. Share your favorite AMD memories in the comments.

I meant to post this a few days ago, but it fits in really nicely with our benchmark testing to explain what's going on under Windows 7's hood. Microsoft obviously focused a lot on the user experience in Windows 7, so a lot of work went into improving desktop responsiveness—smoothing out the little snags or hangs up that made people feel like Vista was too slow. Which is apparently a hard thing to do, since a million different things can cause slowdown. But the most frequent cause of hangups is a bottleneck caused by one graphics device interface application—an app that taps your graphics card—waiting on another GDI app that's being all slow and crappy.

In Vista, this could happen because the way the GDI was designed, a single app could hold a system-wide global lock, so apps running simultaneously constantly jockey for the lock in order to render on the screen, and if one asshole app doesn't let go, it screws every other app waiting in line. So Microsoft re-designed the way this stuff is orchestrated, so multiple apps can "reliably" render at the same time, meaning less bottlenecks. Besides improving reliability, the redesign actually improved performance with multiple GDI apps running simultaneously on multi-core processors, so you'll see real benefits from going multi-core, which no doubt makes Intel's Craig Barrett happy.

Oh yes, they also reduced the memory footprint, but anybody running Windows 7 already noticed this. So yes, Windows 7 really is more responsive, even if run-of-the-mill benchmarks can't exactly measure how that is. [Engineering Windows 7]