Accelerated-X v. 2.1 & Metro-X 3.1.2

This has not alleviated the pain as much as I hoped. The appearance of Metro-X and Accelerated-X in the offices of Linux Journal, begging for a review, has turned my worst nightmare into an exercise in rebellion: now I can reconfigure X for the hell of it, in about twenty seconds.
So Which Server is Better?

Before we get to The Chart, I did notice one interesting thing while I was playing around with the two servers. If /usr/X11R6/bin/X is a soft link to an XFree86 server, such as XF86_SVGA, AccelX will move the XFree86 server to X.LINUX. If, however, /usr/X11R6/bin/X is linked to, oh, say, Xmetro (the Metro-X server), then AccelX will turn Xmetro into a link to Xaccel, destroying the Xmetro server. Hmmmm.



Accelerated X

Graphics card Support

38 manufacturers

51 manufacturers


up to 4 displays

Additional Cost add-on





GUI and curses

curses only

System Requirements

8MB RAM 12 MB HD space

4MB RAM, 3 MB min HD space

Require XFree installed







printed, web

printed, web






I am currently running Metro-X. I have two friends who swear by X Inside. Which server is better sort of depends on your needs. Metro-X has more support for commercial applications, including multi-head display and touch-screen capabilities. Accelerated X supports more graphics cards, making it more likely to suit your hardware.

Bottom line? Check your hardware against both lists, and if both servers will meet your needs, flip a coin—they're both cheaper than a day spent calculating modelines.

Accelerated X Update

The mouse trouble noted in our review of Accelerated X—the mouse flinging applications off the desktop and causing a disruptions to the server connection—have been resolved.

Linux kernel version 2.0.0 introduced a problem with the Async mouse feature of Accelerated X. (This feature allows the mouse to move around the screen even when the X server is busy doing something else—a usability trick Microsoft has been using for a while on their Windows product.) X Inside notified the Linux kernel team and worked with them to correct the problem. With release 2.0.26 of the kernel, the error was resolved.

Jon Gross is a scuba diver, a marine biologist, a writer, a cyclist, a sysadmin, and works for Seattle Software Labs Inc. In his “spare” time, he likes to sleep and cook (not concurrently). He can be reached at