Re: gkrellmd question

On Mon, 11 Jan 2010 20:14:03 -0500
Rio <[email protected]> wrote:

> On Monday 11 January 2010, you wrote:
> > On Mon, 11 Jan 2010 06:36:16 -0500
> > Rio <[email protected]> wrote:
> > 
> > > I am trying gkrellmd and using a 'remote' client as my local monitor in an attempt to 
> > > thwart this 'duplicating' effort kde appears to be doing on just the local monitor. 
> > > however, i find when I use gkrellmd I lose the local nvidia temp monitors for my vid 
> card 
> > > and tesla.
> > > 
> > > is gkrellmd not capable of watching those modules?
> > > anything special I need to do that was not in the gkrellmd.conf?
> > > here is my module list
> > 
> > nvidia temps are monitored using the nvdia-settings command and as far as I
> > know the nvidia module does not report temps directly in the kernel /sys or /proc
> > filesystem.
> > 
> > If you have the nvidia-settings command then it should be used automatically
> > regardless if you are running gkrellm or gkrellmd and you will see it in the config
> > if it detects gpus that report temps.  Also run "gkrellmd -d 0x80" and look for
> > the nvidia-settings line.  You should get the same result as running
> > "gkrellm -d 0x80" on the same machine.  If there are differences regarding
> > nvidia, what are they?
> > 
> 
> the 2 modules are kernel modules
> 
> it87                   26936  0
> hwmon_vid               3648  1 it87

I don't think hwmon_vid is for getting any graphics card temps.  The "vid" has to
do with "voltage identification" and not video.  If in gkrellm where you are getting
gpu temps, the sensors config names for the GPU sensors is like "nVidia GPU:0 core",
then the gpu temps are coming from nvidia-settings.

> these are my results from the above test
> 
> gkrellm -d 0x80                       
> --- GKrellM 2.3.4 ---                                 
> hddtemp_daemon_read (once a minute):                  
>         |/dev/sda|WDC WD7502ABYS-02A6B0|30|C||/dev/sdb|WDC WD7502ABYS-02A6B0|30|C|
> nvidia-settings gpus = 2                                                          

> 
> gkrellmd -d 0x80
> Starting GKrellM Daemon 2.3.4
>         [gkrellm_connect_to: (2,1,6) 127.0.0.1:7634]
> hddtemp_daemon_read (once a minute):
>         |/dev/sda|WDC WD7502ABYS-02A6B0|30|C||/dev/sdb|WDC WD7502ABYS-02A6B0|30|C|
> nvidia-settings gpus = 0

The problem is here.  You are getting different results between gkrellm and gkrellmd.
It's strange because both programs use the same code which runs nvidia-settings to
determine the number of gpus you have.  What do you get when you run by hand:

   nvidia-settings -q gpus

I'm afraid this is something that would take more detailed debugging.  I could add
some debugging output if you can compile a test tarball and run some more tests?

Bill

______________________________________________
Gkrellm mailing list
[email protected]
Read archives-join-leave-set RSS feed:
http://archive.lists.netservicesgroup.com
This service is provided by The Network Services Group:
http://www.netservicesgroup.com