Accelerated-X™ Summit Series
X server/driver Feature

MultiView - Dual, Triple, Quad          

When a graphics card can drive two, three, or four monitors, it is often referred to as a "DualHead," TripleHead," or "QuadHead" card. (When it can drive eight monitors, we call it an eight-head card). A monitor is obviously referred to as a "head." A graphics card with multiple outputs emables one to be creative in how the multiple outputs are used to make up a display.

More Heads, the Better
The Summit Series can support such graphics cards (with the permission of the manufacturer), depending upon the Series/Edition, and can also support mulltiple independent xscreens per GPU in some Series/Editions.

A card that has the ability to drive up to two monitors usually has only one graphics chip (GPU) with two monitor channels.

Some quad-head cards have only one GPU (Matrox QID), while others have two or four GPUs (e.g., ATI FireMV and Matrox G450MMX, respectively).

Triple-head cards generally have one GPU driving the three monitor channels.

The MultiView Feature is designed to support cards that have a GPU that are driving multiple monitor channels and provide full hardware acceleration for each channel..

BTW, what one sees on a monitor is referred to as a "view," not a "screen." The term 'screen' has a particular significance in 'X', and it is not necessarily what is seen on a single monitor. (See MultiChipSLS).

A card with a single GPU and two monitor channels can take advantage of the DualView Summit Feature, which can provide full hardware acceleration for each of the monitor views.

With DualView, one can have the same view on both monitors (the "Copy" mode). One monitor might be at a patient's bedside, while the other is at the nurse's station.

Or, one can have an xscreen for each monitor, with the ability to move ther cursor from one xscreen to the other (but not drag windows from one to the other).

Another DualView configuration (the sexy one) allows the two monitor views to be combined into one xscreen, and that xscreen is divided between the two monitors. This is a Single Logical Screen that is "stretched" across multiple monitors (two in this example). Actually the image is not actually stretched - instead think of it as the desktop being stretched, i.e., "stretched desktop."

In the DualView SLS mode, everything is still hardware acceleerated, and one can drag a window from one monitor to another. With a card driving a couple of large-resolution monitors (say1920x1200) from one graphics chip, the real estate is impressive, and relaitvely inexpensive. P

Few cards support three monitors from a single graphics chip, but they are handled very much as the dual-haad cards with a single GPU.

Matrox has graphics cards that can drive four monitors from a single GPU (QID Pro, for example). There are some quirks with at lest some of these cards that impose some limitations on how the four monitors can be configured with QuadView.

First, all monitors should be identical (don't mix resolutions). Second, the number of xscreens is limited to one, two or four.

MultiView applies to a single GPU. The DualView example of two 1920x1200 in SLS mode provides a large chunk of real estate with one xscreen.

If the system had two dual-head cards, this configuration can be repeated (using the HX Series, for example), providing two more large monitors sharing an xscreen. However, the four monitors cannot be configured into one single xscreen using MultiView.

That requires the Multiple GPU SLS feature, since the "desktop" is being "stretched" across graphics chips, a much more complicated SW task.