Duncan Webb schreef:
Good to know, thanks for the info. However I'm pretty certain that this
does not stop the CPU going into a power saving mode.
I don't want him to sleep (it's a 24/7 server) but all steps I took to
lower the ticks saved me about 7 degrees Celcius... (and humans always
want
Jaap Struyk wrote:
Duncan Webb schreef:
It seems to be but may not be as reliable as rel-1-7 branch.
Actually it ran quite nicely apart from some trouble I got with the
recordserver due to kaa trouble.
Apart from that, 1.8 doesn't improve cpu calls, it's still ~205.
This is not not bad
Duncan Webb schreef:
This is not not bad and big improvement on the 1000 what was reported in
the bug number 1781831
Uhmmm, I yust upgraded to kernel 2.6.23.x and now my freevo is doing
1000 wakeups too...
I will see what kernel option I used that causes this.
--
Groetjes Japie
Jaap Struyk schreef:
I will see what kernel option I used that causes this.
CONFIG_HIGH_RES_TIMERS is responcible for this, without it I get ~205
again. (running 1.8.x fine and stable ;)
--
Groetjes Japie
-
This SF.net
On Sun, Nov 25, 2007 at 06:36:51PM +0100, Jaap Struyk wrote:
CONFIG_HIGH_RES_TIMERS is responcible for this, without it I get ~205
again. (running 1.8.x fine and stable ;)
Could the CONFIG_NO_HZ (Dynamic ticks) option be of any help? This
url[1] that it doesn't reduce much the power
Duncan Webb schreef:
It seems to be but may not be as reliable as rel-1-7 branch.
Actualy it ran quite nicely apart from some trouble I got with the
recordserver due to kaa trouble.
Apart from that, 1.8 doesn't improve cpu calls, it's still ~205.
b.t.w. John is right; You are doing a great job
Duncan Webb wrote:
Jaap Struyk wrote:
Duncan Webb schreef:
1.7.4 is the rel-1-7 branch, the up coming 1.8.0 (rel-1 branch) has the
kaa.notifier code.
Is the 1.8 usable for all day use?
It seems to be but may not be as reliable as rel-1-7 branch.
I've been
Jaap Struyk wrote:
Duncan Webb schreef:
1.7.4 is the rel-1-7 branch, the up coming 1.8.0 (rel-1 branch) has the
kaa.notifier code.
Is the 1.8 usable for all day use?
It seems to be but may not be as reliable as rel-1-7 branch.
Are there any values inside the code wich can be lowered?
Duncan Webb schreef:
1.7.4 is the rel-1-7 branch, the up coming 1.8.0 (rel-1 branch) has the
kaa.notifier code.
Is the 1.8 usable for all day use?
Are there any values inside the code wich can be lowered? (like media
polling intervals)
By configuration, no.
I understand that, I've been
Duncan Webb schreef:
kaa notifier is being used in the rel-1 branch, may be this version is
better.
1.7.4 is slightly bether, it has a proces of 100 ticks and one of ~105.
Funny is the value of 100, my nvidia drives has 50 (wich is the redraw
freq.) maybe freevo is also doing some screen
Jaap Struyk wrote:
Duncan Webb schreef:
kaa notifier is being used in the rel-1 branch, may be this version is
better.
1.7.4 is slightly bether, it has a proces of 100 ticks and one of ~105.
Funny is the value of 100, my nvidia drives has 50 (wich is the redraw
freq.) maybe freevo is
Hello,
Recently I started with powertop to see how I can bring down
powerconsumtion and therefore temp. on my server/freevo box and freevo
is responcible for most interupts; around 230 per second.
Is there a way to lower this? (polling intervals for media or remote?)
--
Groetjes Japie
On 2007-11-20 06:39, Jaap Struyk wrote:
Recently I started with powertop to see how I can bring down
powerconsumtion and therefore temp. on my server/freevo box and freevo
is responcible for most interupts; around 230 per second.
Is there a way to lower this? (polling intervals for media or
Jason Tackaberry wrote:
On 2007-11-20 06:39, Jaap Struyk wrote:
Recently I started with powertop to see how I can bring down
powerconsumtion and therefore temp. on my server/freevo box and freevo
is responcible for most interupts; around 230 per second.
Is there a way to lower this? (polling
14 matches
Mail list logo