The following table shows the technical characteristics of the most recent solutions from NVIDIA and ATI:

The Radeon X1950 PRO is based on the famous code named RV570 architecture. The GPU is completely new and is the first ATI GPU constructed under an 80 nanometer process. The GPU is based on the R580 architecture, which is used for the Radeon X1900 and X1950 video cards. The main difference between the two architectures, however, is that the RV570 comes with 36 pixel shaders while the R580 has 48. The number of texture mapping units and ROPs has also been cut down from 16 to 12. The amount of vertex shaders remains untouched at 8.

The X1950 PRO also uses a completely new way of handling Crossfire. As you will see later on in this article, the Radeon X1950 PRO does not require the use of a Master Card in order to implement a Multi-GPU Crossfire setup. ATI has also done away with the external Crossfire connector cable. More is to come about this later in the article.

The card has a 256bit memory bus width and uses GDDR3 memory clocked at 1,380 MHz. The maximum bandwidth per second is 44.1 Gbytes. As can be seen, the Radeon X1950 PRO has technical specifications that are very similar to the Radeon X1900GT. Both video cards have 8 units of vertex shaders and 36 pixel shaders. The main difference between the two cards is that the X1950 PRO has a core clock of 1.38 GHz while the Radeon X1900 GT has a core clock of 1.2GHz.

Radeon X1950PRO (GPU RV570)    -    Radeon X1950XTX (GPU R580)

Comparing the RV570 GPU to the R580 GPU, the GPU which is being used in the high-end Radeon X1950XTX cards, it is evident that with their more recent GPU, the RV570, ATI has significantly reduced the dimensions of the chip’s surface. The RV570 has dimensions approximately equal to 17x14 millimeters, while for the R580 GPU the dimensions are 19x19 millimeters. The area of the RV570 GPU is 238 millimeters. This is a reduction of 34% in respect to the 361 square millimeters of the RV580 GPU.