Re: XFree86 4.1.0 r128 driver compilation
X11perf reportedly is able to benchmark 2-d performance.
would the r128 frame buffer kernel module be accelerrated? (included with
2.4)
- David Zhou
dzhou@glue.umd.edu
aim: Lightstatic
On Tue, 23 Oct 2001, Michael D. Crawford wrote:
> Whether or not you get accelleration in the fbdev depends on whether the
> framebuffer driver you're using supports it, and whether you enable it as a
> kernel option when you boot your machine.
>
> On the x86, the VESA2 framebuffer driver is easy to set up but does not have
> accelleration.
>
> My compaq laptop uses the Mach64 chip. This is documented to be accellerated,
> but when I try to use it, the most spectacular thing happens to the LCD screen,
> it looks like someone is taking a blowtorch to it from the back. It gives on
> the urge to turn the power off, and quickly.
>
> I am not entirely certain that XFree86 uses framebuffer accelleration if it is
> available though. But I would suspect that r128 is probably the one where
> framebuffer accelleration is most likely to work.
>
> Note that there are two kinds of accelleration to consider, 2D and 3D. 2D
> accelleration speeds up conventional X drawing, like drawing lines as vectors
> rather than as a sequence of pixels. 3D accelleration requires that you get DRI
> working. Most people have an easier time getting 2D accelleration working than
> #d accelleration.
>
> Here's a question - what's a good way to make a simple benchmark of 2D
> accelleration? For 3D, most people usually report the framerate of glxgears, or
> for more in-depth testing, quake running in demo mode. But I haven't found a
> simple way to test 2D accelleration.
>
> Mike
> --
> Michael D. Crawford
> GoingWare Inc. - Expert Software Development and Consulting
> http://www.goingware.com
> crawford@goingware.com
>
> Tilting at Windmills for a Better Tomorrow.
>
>
> --
> To UNSUBSCRIBE, email to debian-powerpc-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
>
>
Reply to: