laptop running about 10C hotter?

cfr · 2012-09-09 23:58:30

Recently, I've noticed that my laptop is running hotter than it was. I'm not sure if this is due to a kernel change, some other upgrade or my move to systemd. I would appreciate any diagnostic hints.

When I first got Arch set up, sensors reported 40ish on idle and now it reports 50ish. Similarly, when it is working intensively, the temperature tends to be approximately 10C higher than I would have expected previously. I caught it in the 70-80 C range recently. (I don't think I've seen it go over 80 though I am not looking all the time, obviously.)

According to sensors, high is 86+ and critical is 100+ so I don't think it is dangerously high. But since I'm not doing anything different, it is obviously rather annoying to be running through more energy, working with a hotter keyboard and draining batter power faster than need be.

I'm interested, in possible, in finding out what's changed.

Morn · 2012-09-12 15:57:55

Maybe take a look at the output of "cpupower frequency-info" and verify the ondemand governor is active? For me, automatic CPU frequency scaling by the kernel governor didn't work at all in the beginning (i.e., the CPUs were always running at max frequency when ondemand was active). It was only after moving to systemd and the 3.5 kernel recently that this issue was resolved. Perhaps your problem is related to that.

cfr · 2012-09-12 23:22:28

That's not it. The ondemand governor is still kicking in fine. My cores are currently, for example, all operating at the lowest frequency yet my machine is still 10C hotter than it used to be in similar circumstances.

bws · 2012-09-13 01:02:45

Could it be a fan has stopped running? Strange that it is always within 10 degrees of expected though.

Did you recently move the computer to a warmer pyhisical location?

cfr · 2012-09-13 02:35:16

bws wrote:

Could it be a fan has stopped running? Strange that it is always within 10 degrees of expected though.

Nope. The fan is definitely running. Sensors and a hair dangled in front of it agree.

Did you recently move the computer to a warmer pyhisical location?

Nope. If anything, it is cooler than it has been but I still see just the same difference. Always around 10C up from what it was before e.g. 50~ vs. 40~ on idle 60/70/80 vs 50/60/70. I don't understand why it is that consistent...

bws · 2012-09-14 01:26:06

Yeah, it seems a system variable was changed since the variance is THAT consistent. You could try incremental downgrades until you reach a point where you "know" the sensors were lower, but that sounds like a pain in the rear. Not to mention one of the upgrades you did could have been a bug fix, i.e. your system temp was miscalculated by 10 degrees before, but is now reported correctly.

You could look up the sensors package to see if there have been any recent bug fixes or other changes that could explain it, but outside of using incremental downgrades to pinpoint the specific package I'm not sure what other diagnostics you can do at this point.

edit:
Just a thought, are system temps logged anywhere? You could dig through /var/log to see if there's anything there. I don't know of one off hand, but it could at least give you a date/time to work with.

Last edited by bws (2012-09-14 01:30:22)

brebs · 2012-09-14 02:18:25

Are you blaming it on software, without having bothered to clean the dust out of the PC?

cfr · 2012-09-14 02:19:16

I suspect the logs will have been rotated out... This isn't the kind of thing which I immediately notice - only once the pattern emerges and I don't constantly sit running sensors...

Morn · 2012-09-14 08:39:13

Maybe you should install the linux-lts kernel (currently at 3.0.42) and see if running that lowers the temperatures. There was talk in another thread about a possible power regression in 3.5.

cfr · 2012-09-14 22:01:31

brebs wrote:

Are you blaming it on software, without having bothered to clean the dust out of the PC?

No. I had the machine open several times during this period and I always try to de-dust the fan though there wasn't much on it last time. (Probably because it wasn't long after I'd opened the machine up before.)

Pretty sure it isn't that.

cfr · 2012-09-14 22:02:28

Morn wrote:

Maybe you should install the linux-lts kernel (currently at 3.0.42) and see if running that lowers the temperatures. There was talk in another thread about a possible power regression in 3.5.

That sounds like a pleasanter experiment than trying to downgrade the kernel. I'll give that a shot - thanks.

tobyjwebb · 2012-09-18 19:23:44

cfr wrote:

brebs wrote:
Are you blaming it on software, without having bothered to clean the dust out of the PC?
No. I had the machine open several times during this period and I always try to de-dust the fan though there wasn't much on it last time. (Probably because it wasn't long after I'd opened the machine up before.)
Pretty sure it isn't that.

Another thing that can make the CPU go hotter (not sure if it's your case or not), is the thermal paste can dry up. What normally happens to my laptop is that it starts rebooting because of it being too hot. (I live in a pretty hot country.)

Other than that, could it be that you're comparing the temp in summer to the temp in winter?

Salutations

cfr · 2012-09-19 02:40:34

I'm comparing the temperatures in (hottish) July with those in (cooler) August and (cooler still) September.

I'd hope the thermal paste wouldn't have dried up yet but how would I check? (Where would I look and for what? I've read about this issue but have no idea what it actually looks like.)

Arch Linux

#1 2012-09-09 23:58:30

laptop running about 10C hotter?

#2 2012-09-12 15:57:55

Re: laptop running about 10C hotter?

#3 2012-09-12 23:22:28

Re: laptop running about 10C hotter?

#4 2012-09-13 01:02:45

Re: laptop running about 10C hotter?

#5 2012-09-13 02:35:16

Re: laptop running about 10C hotter?

#6 2012-09-14 01:26:06

Re: laptop running about 10C hotter?

#7 2012-09-14 02:18:25

Re: laptop running about 10C hotter?

#8 2012-09-14 02:19:16

Re: laptop running about 10C hotter?

#9 2012-09-14 08:39:13

Re: laptop running about 10C hotter?

#10 2012-09-14 22:01:31

Re: laptop running about 10C hotter?

#11 2012-09-14 22:02:28

Re: laptop running about 10C hotter?

#12 2012-09-18 19:23:44

Re: laptop running about 10C hotter?

#13 2012-09-19 02:40:34

Re: laptop running about 10C hotter?

Board footer