Hi!
Running latencytest I have found a quite bad behaviour with the high
X11 load test when the X server has the DRI module loaded and active.
Is anyone there using DRI and lowlatency-patch at the same time?
Have you experienced this kind of problems?
Any known solution? I use OpenGL for real time visualization of audio,
and it is not nice if I need to disable HW accel!!!
Regards,
Enrique.
---- Some output from latencytest:
Using 3x512 fragmets, with DRI loaded and direct rendering active, I
got:
> fragment latency = 2.902494 ms
> cpu latency = 2.321995 ms
> 8.1ms ( 80)|
> 1MS num_time_samples=23155 num_times_within_1ms=22451 factor=96.959620
> 2MS num_time_samples=23155 num_times_within_2ms=22451 factor=96.959620
Or in other words, 3% frame dropouts...
After removing the load "dri" line from the XF86Config-4, it goes
fine:
> fragment latency = 2.902494 ms
> cpu latency = 2.321995 ms
> 3.3ms ( 0)|
> 1MS num_time_samples=21477 num_times_within_1ms=21477 factor=100.000000
> 2MS num_time_samples=21477 num_times_within_2ms=21477 factor=100.000000
Or in other words, clean sound, and a beautiful latency graph.
I have tried a few times, just commenting and uncommenting the 'load
"dri"' line, and the results are always almost identical.
---- System Info:
* kernel 2.4.17 with
** SMP disabled
** Andrew Morton's low latency patch active.
** agpgart active (VIA chipset)
** mga drm driver active.
* ALSA 0.9.0rc1.
* latencytest0.42-png
* XFree86 version: 4.1.0.1
** mga driver
** GLX Direct Rendering working when DRI is loaded.
* Processor: Dual Pentium III 800, (I did this tests with SMP disabled
in kernel).
* G. Card: Matrox Graphics, Inc. MGA G400 AGP (rev 04)
* S. Card: Ensoniq ES1371 [AudioPCI-97] (rev 08)
* AGP chipset: VIA Technologies, Inc. VT82C598/694x [Apollo MVP3/Pro133x AGP]
