View previous topic :: View next topic |
Author |
Message |
colinbarnette n00b
Joined: 07 Mar 2007 Posts: 17 Location: Louisville, Kentucky
|
Posted: Wed Apr 25, 2007 3:43 pm Post subject: A question on system usage. |
|
|
After having my system's heat alarm go off the other night while playing Enemy Territory, I decided I needed to monitor my temperatures. So, using lm_sensors and top, I gathered a little data and I'm needing some help understanding. The data was taken over a 5 minute period for each situation.
I am using kernel 2.6.19.2, my CPU is an Athlon XP 2500+.
System Temp. was steady the whole time at 29C.
Code: | Running X, FVWM, gnome-terminal:
CPU: 31 - 33C
Usage: 0.0 - 1.0%
With Rhythmbox, no music playing:
CPU: 31 - 33C
Usage: 0.0% - 1.0%
Rhythmbox, music playing:
CPU: 44 - 46C
Usage: 2.0 - 3.0%
Rhythmbox is closed, running Enemy Territory:
CPU: 53 - 55C
Usage: 70 - 99% |
So my curiosities are: Why does playing music use only a percent or so more of the CPU, but increase the Temp. by ~13C? Why on full load with ET does it only increace another 9 - 10C?
I'm not really looking for specifics, just wondering if this seems normal, or if it indicates I need a new cooling solution? Really I just want to understand why. I hope this peaks someones interest, and thanks in advance!
P.S. As I write this with music playing in Rhythmbox and Bon Echo, CPU: 47C, Usage: 3 - 7%. |
|
Back to top |
|
|
asturm Developer
Joined: 05 Apr 2007 Posts: 8936
|
Posted: Wed Apr 25, 2007 4:52 pm Post subject: |
|
|
Uhm, the temperature increase while playing music is quite strange.
You don't have PowerNow enabled, dynamic fan speed or something similar, do you?
Although, a 5 minute period is a bit on the weak side for conclusions. |
|
Back to top |
|
|
colinbarnette n00b
Joined: 07 Mar 2007 Posts: 17 Location: Louisville, Kentucky
|
Posted: Wed Apr 25, 2007 5:18 pm Post subject: |
|
|
Thats a no on the dynamic fans, but I'm not sure about PowerNow so I'll assume that's a no too.
I know 5 minutes is a bit weak, but even now, Firefox and Rhythmbox, if I pause the music, I have an immediate drop to 35C from about 44C. This is according to lm_sensors, which I believe is configured correctly, as when I tested it at full load I let it get to my alarm level and it reported correctly.
So what is it about the music? Is it a sensor error, or is it really spiking my CPU temp, and why?
Strange indeed. Any other thoughts? |
|
Back to top |
|
|
colinbarnette n00b
Joined: 07 Mar 2007 Posts: 17 Location: Louisville, Kentucky
|
Posted: Wed Apr 25, 2007 5:28 pm Post subject: |
|
|
A quick follow up:
I just viewed a video on YouTube with sound and I got about a 5 degree increase over just having sound.
When I played a video in mplayer that did not have sound, I only saw about a 5 degree increase from the "idle" of 32. With the YouTube video I was around 50C. It's got to be something lower level than Rhythmbox or Firefox... |
|
Back to top |
|
|
vibrokatana Guru
Joined: 09 Feb 2006 Posts: 328 Location: o0o0oo
|
Posted: Wed Apr 25, 2007 5:30 pm Post subject: |
|
|
CPU temperature monitors are known to be inaccurate. you should open the case and feel if the heatsink is hot. 50-60C should feel almost unbearable to the touch on the heat sink and ~30C slightly warm. You may need to remount the heatsink with new paste and make sure it is mounted in the right direction (anthlon xp heatsinks only go on 1 way and will be slanted if you manage to get it on backwards). _________________ My Systems - "I suggest the whole thing be coded in whitespace. Henceforth the code will be obscure and functional at the same time." |
|
Back to top |
|
|
asturm Developer
Joined: 05 Apr 2007 Posts: 8936
|
Posted: Wed Apr 25, 2007 5:31 pm Post subject: |
|
|
Couldn't think of any sane reason. For a try, you could watch your temperatures during execution of an other music player.
Playing music via Amarok changes just about nothing at my idle system temperatures. First, my CPU needs a very long time going up from 25°C to an operating temperature of 33°C (which you should await too, before interpreting temp-figures) - and there it stays, regardless of running Amarok or not. Even a vastly watercooled system as mine would show at least some increase, when yours jumps up 11-13°C. |
|
Back to top |
|
|
colinbarnette n00b
Joined: 07 Mar 2007 Posts: 17 Location: Louisville, Kentucky
|
|
Back to top |
|
|
bcmm Tux's lil' helper
Joined: 27 Jun 2004 Posts: 100 Location: Oxford, UK
|
Posted: Wed Apr 25, 2007 6:40 pm Post subject: |
|
|
Different instructions cause different amounts of heat production. 100% CPU usage does not equal maximum heat output.
Also, ET is probably GPU-bound.
If you applied the thermal paste yourself, there are a number of possible reasons for a poor thermal connection. The most likely one is too much paste. If it's spread too thick, it makes a bad connection, as it's nothing like as thermally conductive as the heatsink's copper. Arctic silver themselves publish the best guide I've found to applying thermal paste. I followed that, and after a couple of messed up attempts, I managed to get my heatsink mounted so that I never get any cooling issues (always below 50C).
I've found it works well to scrape loads off after applying it, even if that wastes some. There is so little Silver involved that it doesn't cost much.
If you do reseat it, clean it really carefully. Isoproyl alcohol is supposed to be best. I've used isoproyl-based CD wipes on CPUs with some sucess.
"some sort of Arctic Silver"==Arctic Silver or Arctic Silver 5? The former is a thermal adhesive and is probably less good (as well as horrid to remove). _________________
Code: | bash-2.05b# cat /dev/mem | strings | grep -i llama |
Damn, my RAM is full of llamas! |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|