2013-11-16, 05:24
Hi all
I am new to XBMC so : congratulations to all people having contributed
to this HTPC system!
I built bleeding edge gentoo intel XBMC with nvidia graphics and see
xbmc.bin running 100% CPU on one core, even when nothing happens. I
have read a little bit and I have adapted algorithmdirtyregions
nofliptimeout in advancedsettings. It does not change behaviour.
vmstat shows -MANY- interrupts and context switches per second:
# vmstat 1
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
...
1 0 0 248120 265336 2933444 0 0 0 0 71403 142516 25 1 74 0
...
pidstat points to XBPyThread
# pidstat -wt | egrep '(__XBPyThread|cswch)'
03:57:49 UID TGID TID cswch/s nvcswch/s Command
03:57:49 1000 - 4437 7073.50 0.03 |__XBPyThread
03:57:49 1000 - 4440 0.73 9.33 |__XBPyThread
03:57:49 1000 - 4441 3898.50 0.01 |__XBPyThread
powertop also spots xbmc.bin with similar amount of wakeups
Summary: 7982.1 wakeups/second, 0.0 GPU ops/seconds, 0.0 VFS ops/sec and 55.2% CPU use
Usage Events/s Category Description
548.7 ms/s 7862.5 Process /usr/lib64/xbmc/xbmc.bin
FWIW: /proc/interrupts reports "rescheduling"
# watch -tdn1 cat /proc/interrupts
RES: 15097764 10079972 12758152 4994793 Rescheduling interrupts
Is this a known issue?
Thanks!
I am new to XBMC so : congratulations to all people having contributed
to this HTPC system!
I built bleeding edge gentoo intel XBMC with nvidia graphics and see
xbmc.bin running 100% CPU on one core, even when nothing happens. I
have read a little bit and I have adapted algorithmdirtyregions
nofliptimeout in advancedsettings. It does not change behaviour.
vmstat shows -MANY- interrupts and context switches per second:
# vmstat 1
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
...
1 0 0 248120 265336 2933444 0 0 0 0 71403 142516 25 1 74 0
...
pidstat points to XBPyThread
# pidstat -wt | egrep '(__XBPyThread|cswch)'
03:57:49 UID TGID TID cswch/s nvcswch/s Command
03:57:49 1000 - 4437 7073.50 0.03 |__XBPyThread
03:57:49 1000 - 4440 0.73 9.33 |__XBPyThread
03:57:49 1000 - 4441 3898.50 0.01 |__XBPyThread
powertop also spots xbmc.bin with similar amount of wakeups
Summary: 7982.1 wakeups/second, 0.0 GPU ops/seconds, 0.0 VFS ops/sec and 55.2% CPU use
Usage Events/s Category Description
548.7 ms/s 7862.5 Process /usr/lib64/xbmc/xbmc.bin
FWIW: /proc/interrupts reports "rescheduling"
# watch -tdn1 cat /proc/interrupts
RES: 15097764 10079972 12758152 4994793 Rescheduling interrupts
Is this a known issue?
Thanks!