1 |
On Wed, 20 Jun 2012 22:47:08 -0400 |
2 |
"Walter Dnes" <waltdnes@××××××××.org> wrote: |
3 |
|
4 |
> On Wed, Jun 20, 2012 at 07:04:13AM -0700, walt wrote |
5 |
> |
6 |
> > I just did the same update. I thought at first everything was okay, |
7 |
> > then I ran glxgears, which runs at 1/5 normal speed. Going back to |
8 |
> > 295.59 fixed it. Very puzzling, since there was nothing unusual in |
9 |
> > Xorg.0.log. |
10 |
> |
11 |
> What does glxgears prove anyways? With 295.59 I get 5300 fps, and |
12 |
> 302.17 gives me 3 spf (yes... seconds per frame!!!). But Youtube, etc |
13 |
> seems to function the same. So I'm wondering about the point of |
14 |
> glxgears. |
15 |
> |
16 |
|
17 |
glx gears shows you that stuff works or not. |
18 |
|
19 |
So it's a quick go-no go test, just not useful for measuring anything. |
20 |
It's designed for the scenario where you want to judge relatively how |
21 |
well things are working today video-wise. Now, you could run a 3D |
22 |
rendering app, or play a video, or something like that and see how it |
23 |
performs compared to the same thing yesterday. But those apps all do |
24 |
many other things as well as drive the video card so you don't really |
25 |
know what is influencing the mix. |
26 |
|
27 |
With glxgears, all it does is drive the video and draw spinning gears, |
28 |
nothing else. So if version X-1 yesterday worked fine and the video was |
29 |
smooth, but version X today plays very jerkily, and nothing else |
30 |
changed, then you know your system does not like version X. |
31 |
|
32 |
It's a perfectly valid test, just don't trust the numbers - trust what |
33 |
your eyes tell you. |
34 |
|
35 |
|
36 |
|
37 |
-- |
38 |
Alan McKinnnon |
39 |
alan.mckinnon@×××××.com |