![]() |
Sharky Extreme : Video cards |
|
![]() |
![]() - Most Active Threads - Technical Support - CPUs & Overclocking |
![]() |
Video cards |
PNY Verto GeForce 7950 GT 512MB ReviewBy Vince Freeman November 22, 2006The high-end section of the video card market is a very competitive one, where both ATI and NVIDIA work hard to attract buyers at the all-important $250-$350 price range. One strategy employed by both companies is the tweener release, where a new video card is inserted between two existing models, either to shore up a gaping hole, or to head off a competitor's product. ATI has done this with the Radeon X1900 XT and X1950 Pro 256MB cards, both of which offer different price-performance ratios to the previous ATI line-up. NVIDIA has also performed the same maneuver a few times before, and the GeForce 7950 GT 512MB is an excellent example. This card is not only designed to compete well against similarly-priced ATI models, but also give high-end buyers an option between the GeForce 7900 GT and 7900 GTX, and to eventually replace the 7900 GT. The GeForce 7950 GT 512MB card features the same 90nm G71 GPU architecture found on the existing GeForce 7900 GT and GTX cards, and includes a full complement of 24 pixel pipes, 24 texture units, 24 pixel shaders, and 8 vertex pipelines. Other features such as CineFX 4.0, Intellisample 4.0, UltraShadow II, NVIDIA SLI, and PureVideo are all present and accounted for. The main difference between the GeForce 7900 GT/GTX and the GeForce 7950 GT is clock speed, with the latter card clocked at 550 MHz core and 1.4 GHz memory. This compares to 450 MHz core/1.32 GHz memory for the GeForce 7900 GT and 650 MHz core/1.6 GHz memory for the GeForce 7900 GTX. These clock speeds translate into some high-end fillrate and memory bandwidth specifications. The GeForce 7950 GT sports fill rates of 13.2 GTexel/s and 13.2 GPixels/s, and a memory bandwidth of 44.8 GB/s. These specifications compare well to the GeForce 7900 GT (10.8 GPixels/s and 42.2 GB/s of memory bandwidth) and the GeForce 7900 GTX (15.6 GPixels/s and 51.2 GB/s of memory bandwidth) cards. The 512MB of GDDR3 memory utilizes the same 256-bit memory interface and architecture. The only other change has been the addition of HDCP support, which should theoretically allow the GeForce 7950 GT to play back HD-DVD and Blu-ray media.
The basic design of the PNY Verto GeForce 7950 GT card is similar to a reference GeForce 7900 GT, with all the basics like GPU placement and GDDR3 memory location remaining consistent. The heatsink-fan is another nice holdover from the GeForce 7900 GT, and its diminutive size allows a single-slot design with no impact on adjacent peripherals. The heatsink-fan design is also very efficient and offers low noise and adequate cooling for our review card. PNY has chosen to go with the default clock speeds and the Verto GeForce 7950 GT 512MB is set at standard 550 MHz clock and 1.4 GHz memory speeds.
The card's physical design is also very standard and other than its PNY-branded heatsink-fan, its look is similar to reference designs. The backplate of the PNY Verto GeForce 7950 GT 512MB includes dual integrated dual-link DVI ports and an S-Video-out port. PNY also offers resolutions up to 2560x1600 and HDTV support up to 1080p. Like any other GeForce 7 Series card, the PNY Verto GeForce 7950 GT 512MB requires external power, and features a standard 6-pin PCI Express connector on the outside edge. PNY's retail bundle is ample, if not spectacular, and includes the Verto GeForce 7950 GT PCI Express card, a Quick Installation Guide, a driver/util CD, 2 DVI-to-VGA adapters, a S-video cable, a HDTV Break-out Pod, and a "Y" adapter power cable.
To give a better idea of the architecture of the various mainstream and high-end cards we see competing directly against the GeForce 7950 GT, here is a small chart outlining the key architectural features of each GPU:
Keep in mind that as many graphics designs have taken a modular view of the architecture, so the entire picture of the GPU needs to be taken into consideration. No more is performance relegated to the number of pixel pipelines or texel processing, but pixel and vertex shaders need to be taken into account, as does the number of ROPs (or Render Output units). Of course, the best performance metric is real-world testing, and to that end, we've assembled a wide range of game benchmarks in the next section.
|




