LOSTCIRCUITS

Technical Briefs of the Quadro FX GPU Family

With increasing complexity and requirements on computer-generated images, the level of precision of the rendering engine had to increase to include subpixel precision. Historically, most graphics adapters have been working with a 4-bit subpixel precision that, at high resolution, can create artifacts in the form of spreckles. Spreckles, on the other hand, can be reduced by increasing the subpixel precision beyond 4 bits. The current Quadro FX family of GPU uses 12-bit subpixel precision.

Needless to say that the color scheme is using a true 128-bit floating point mode or 32-bit floating point formats for each of the RGB and alpha channel. One issue with 32-bit floating point operation is that it can slow down the rendering in very complex scenes. Particularly in some technical CAD/CAM (computer-aided design / manufacturing) applications, high color precision may not be required, rather it is real-time operation that is valuable. In this type of application, the Quadro FX can flip-flop back and forth between a 32-bit and a 16-bit mode per channel to optimally match the requirements. Finally, the Quadro FX GPUs support gamma-correction for reduction of antialiased line artifacts.

The nVidia Quadro FX1400 features two DVI outputs, however, those are not Dual Link DVI ports as required by some of the larger monitors used in CAD /CAM environements

The Midrange Quadro FX Lineup

The core of nVidia's midrange professional cards comprises three different models with the FX1500 as the top solution, featuring 256 MB of GDDR3 local frame buffer at 625 MHz clock speed (1250 MHz data rate), whereas the FX 1400 and 1300 use first generation DDR clocked at 600 MHz and 550 MHz data rate, respectively. The complete memory subsystem configuration and bandwidth data are shown below.

Next Page:    => Test Configuration =>