very slow performance of glxgears (68 fps)

[John Tapsell]

2009/2/5 Matthias Hopf <mhopf at suse.de>:
> On Jan 30, 09 18:11:57 -0800, Bryce Harrington wrote:
>> On Fri, Jan 30, 2009 at 01:29:49PM -0800, Eric Anholt wrote:
>> > > $ glxgears
>> > > Failed to initialize TTM buffer manager.  Falling back to classic.
>> > > 300 frames in 5.0 seconds = 59.884 FPS
>> > > 299 frames in 5.0 seconds = 59.621 FPS
>> > > 300 frames in 5.0 seconds = 59.818 FPS
>> >
>> > glxgears is not a benchmark.
>> >
>> > We sync to vblank because running glxgears at 1000fps is dumb.
>>
>> I am going to go out on a limb and guess we're going to see a crapload
>> of reports of "performance regression" due to reduced glxgears frame
>> rates.
>
> At openSUSE we print out a warning now (well, this change went into
> *after* 11.1, unfortunately), that this is not a benchmark. We got really
> tired of these statements.

Except that it _is_ a benchmark.  Or rather waas before the change.
You can argue that it's not an accurate benchmark, but even on the
kernel mailing lists I've seen it used as a crude benchmark.

If for example, without vsync, a user gets just 100fps on their new
nvidia card, then that is clearly showing that something is wrong.

>
> Matthias
>
> --
> Matthias Hopf <mhopf at suse.de>      __        __   __
> Maxfeldstr. 5 / 90409 Nuernberg   (_   | |  (_   |__          mat at mshopf.de
> Phone +49-911-74053-715           __)  |_|  __)  |__  R & D   www.mshopf.de
> _______________________________________________
> xorg mailing list
> xorg at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/xorg
>