EVGA GeForce GTX 295 Review :: Features

• NVIDIA SLI Technology
• Full Microsoft DirectX 10 Shader Model 4.0 Support
• True 128-bit Floating Point High Dynamic Range lighting
• PCI Express 2.0/1.1 Support
• 2nd Generation Unified Shader Architecture
• NVIDIA CUDA support
• NVIDIA PhysX ready
• 2-way SLI Support
• NVIDIA PureVideo HD Technology
• OpenGL 3.0 Support
• Digital Vibrance Control technology (DVC)
• NVIDIA GeForce GTX295
• 576MHz GPU clock
• 2x240 SPs
• 400MHz RAMDAC
• 1792MB GDDR3 memory
• 896-bit memory interface
• 1998MHz memory clock speed
• 223.8GB/second memory bandwidth
• PCI-E 2.0 16x
• DVI-I DVI-I, HDMI
• SLI Capable
• 240MHz Max Refresh Rate
• 2048x1536 Max analog
• 2560x1600 Max Digital
• 680W PSU Minimum (Minimum recommended PSU with +12 Volt current rating of 46 Amps
• An available 6-pin PCI-E power connector and an available 8-pin PCI-E power connector

The EVGA card is based upon NVIDIA's GeForce GTX 295 design which basically takes two of the GeForce GTX 275 chips and combines them on one board. The GeForce GTX 275 was actually released after the GTX 295 as a counter to ATI's HD 4890 chip. Each of the two chips on the card has 1.4 billion transistors, a 448-bit memory interface, 896MB of GDDR3 memory and 240 Stream Processors.

NVIDIA implemented a 448-bit memory interface with their GTX 295 and 275 chips. The GeForce GTX295 having two chips each with a 448-bit memory interface and clocked at 1998MHz has a memory bandwidth of 223.8GB/second an impressive amount of bandwidth exceeded only by overclocked versions of this card at present. The Core Clock is set at 576MHz with the Shader Clock at 1242MHz. Those with the original GeForce GTX 280 cards will note that the clock speed is the default clock speed for that card as well.

NVIDIA chips do not as of yet support Microsoft DirectX 10.1 as virtually the entire ATI lineup does from top to bottom. They do, however, support DirectX 10.0 features like Pixel Shader 4.0, Vertex Shader 4.0 and Geometry Shaders. NVIDIA chips will have to support 10.1 with the release of their DirectX 11 cards in the future, but today ATI is the only company that supports 10.1 in hardware. NVIDIA cards have the advantage of supporting PhysX which is a technology they purchased when they bought Ageia.

PhysX was developed by a company named Ageia which released the PhysX accelerator and was purchased by NVIDIA in early 2008. PhysX allows the computer to do hardware physics on NVIDIA based video cards and on the CPU on other video cards from ATI or Intel. Games like Darkest of Days, Cryostasis, Mirror's Edge and others use PhysX to show off particle effects, realistic smoke, realistic cloth effects, and much more. NVIDIA released a new driver to support OpenCL 1.0 today which is the first driver to support OpenCL conformance in the world with ATI sure to follow soon.

NVIDIA video cards have supported working together in two card solutions since the release of the nForce 4 SLI motherboard chipset in 2004. The GeForce GTX 295 uses an internal SLI connector to bridge the two graphic chips on the card. This allows the graphics card to display 32x and 32XQ FSAA modes. 32X FSAA is actually each graphics card doing 16x FSAA while 32XQ is each graphics card doing 16XQ mode. 16x FSAA on NVIDIA cards is 4x MSAA+12 additional coverage samples. 32x would be 8x MSAA+ 24 coverage samples. 32XQ would be 16x MSAA+ 16 coverage samples.