More power than a stock GTX 680

Every GPU generation has its flagship videocards: the ones with the top-of-the-line GPU with all cores enabled, loaded for bear. In this generation, those cards are Nvidia’s GTX 680 (with a full GK104 GPU inside) and AMD’s Radeon HD 7970 (with a full Tahiti GPU). These cards are monstrously fast, but they’re also expensive and tricky to manufacture. Not all parts come off the line fully functional. So a few months after each flagship GPU launch, the vendors come out with a slightly stripped-down version that uses binned top-end GPUs with a few parts disabled, or lower clock speeds. AMD’s Radeon HD 7950, for example, uses the same GPU as the 7970, but with 28 GCN units instead of 32, and with an 800MHz reference clock instead of 925MHz. The cheaper, lower-powered video cards appeal both to gamers with shallower pockets and also to vendors, who clock those stripped-down, less expensive GPUs right back up to within spitting distance of their full-powered peers. Thus we arrive at the Asus GeForce GTX 670 DirectCU II TOP, a factory-overclocked GTX 670 with a custom cooling solution.

The DirectCU II cooler’s three direct-contact heat pipes keep the GPU cool.

Asus has been coming on strong in graphics cards for several years now, though it never offers quite the variety of versions as companies like XFX and EVGA. Typically, Taiwan-based Asus will ship a reference card under its main brand, and then a custom-built, high-end card under its DirectCU brand. At a later date, the company might ship a super-high-end card using the company’s Matrix or Mars sub-brands. Price differences between Asus’s high-end and standard versions are wider, too, so it’s a little easier to figure out which card really is the premium version.

Unpacking the XFX HD 7950 Black Edition caused a bit of déjà vu. The card bears a strong resemblance to its big brother, the HD 7970 Black Edition (reviewed April 2012), clad in svelte brushed aluminum. If graphics cards dressed up for black tie galas, then the XFX Black Edition is ready to attend.

XFX pushes the reference clocks higher than stock, hitting 900MHz for the core clock and 1,375MHz for the memory clock. The additional memory cycles translate to a peak memory bandwidth of 5.5 gigabytes per second—the same as the HD 7970, and higher than the 5GB/s of the stock 7950. The question is: Can the GPU keep up? There’s always a balance between memory bandwidth and how much of that bandwidth the GPU cores can actually use. Plus, as more games become shader- and tessellation-intensive, bandwidth isn’t as big a part of the equation.