Specifications
| GeForce 7900GTX | GeForce 7950GX2 |
---|
Transistor count | 278 million | 278 million x 2 |
---|
Manufacturing process | 0.09 micrometer | 0.09 micrometer |
---|
Die Area | 196 mm² | 196 mm² x 2 |
---|
Core clock speed | 650 MHz | 500 MHz |
---|
Number of pixel shader processors | 24 | 24 x 2 |
---|
ROPs | 16 | 16 x 2 |
---|
Number of vertex pipelines | 8 | 8 x 2 |
---|
Peak pixel fill rate (theoretical) | 10.4 Gigapixel/s | 16 Gigapixel/s |
---|
Peak texture fill rate (theoretical) | 15.6 Gigatexel/s | 24 Gigatexel/s |
---|
Memory interface | 256-bit | 256-bit x 2 |
---|
Memory clock speed | 1.6GHz GDDR3 | 1.2GHz GDDR3 |
---|
Available Memory | 512MB | 512MB x 2 = 1024MB |
---|
Peak memory bandwidth | 51.2GB/s | 76.8GB/s |
---|
While the manufacturing process and the die has remained the same, the GeForce 7950GX2 GPUs run at the specification using a lower voltage than single GPU configurations. The results are 143W peak power consumption for the GeForce 7950GX2 compared to 95W for a GeForce 7900GTX and 125W for a Radeon X1900XTX. The minimum PSU requirement has been raised from 22A on 12V of the GeForce 7900GTX to 27A on 12V for the GeForce 7950GX2.
The clock speed has been dropped from th 650MHz of the GTX model to 500MHz. The standard 7900GTs run at 450MHz, so the new GPU is placed in between the two models. As mentioned before, some manufacturers have decided to offer factory overclocked cards. The number of Pixel Shaders, ROPs and Vertex Pipelines has doubled as there are two GPUs. The pixel and texture fill rate has not quite doubled because the 7950GX2 uses 1.2GHz DDR3 memory, while the 7900GTX uses 1.6GHz DDR3 memory. In total the memory bandwith has gone up from 51.2GB/s to 76.8GB/s
But besides speed and performance, the GeForce 7900GX2 brings HDCP to the high-end.