Thinkpads Forum

First rule is pick a common tool so we are all using the same metric. I am going to use NHC which currently shows a CPU temperature of 47C and motherboard temperature of 47C. Everest shows the same values (its also using ACPI for CPU and motherboard temps), but it also shows core temperatures from the Intel on-board die sensors of 44C and 47C respectively. Core temps are actually better measure because they are most consistent across hardware lines but the calculations are unfortunately different for each software tool, and since we are just looking at T6x the ACPI should be the most consistent.

This laptop is using the on-demand power management scheme, and has the low-power BIOS options enabled. It's mounted in one of the Advanced Docking Stations which gives it good airflow underneath. I am using BIOS fan control when taking this reading, and TPFanControl is NOT running and is NOT being used for temperature reading (the ACPI sources appear to be in error).

Sometimes when the computer resumes from standby the CPU and core temperatures will spike to ~60C and stay there, like the power management has lost its mind, but after a reboot it's back to the "normal" temps listed above. See this graph for more information.

I would like to get this down to lower 30s but I'm not sure if it's possible due to the compact design. FWIW my workstation rig has idle core temperatures of ~18C

Right this moment, my CPU is idling at 49C. My normal is about 46-47, same as your machine. It's pretty F-ing hot in my apartment right now, which accounts for the slight difference. I use NHC on Dynamic Switching with no manual changes in the CPU voltage.

Anyone know if there is a reason for the higher temps after restoring from suspend? Mine also hovers around 58-59 at idle after I un-suspend it sometimes, which lasts until I reboot it

The T60p was notorious for high operating temps (CPU and GPU). Under load some peoples' temps are approaching 100oC!!! The cooling system was redesigned for the T61p - idle temps are lower now but full load temps are still pretty high. If your workstation is a desktop then you can't compare them. What are your temps on the T60p under load?

These temperatures are quite alright. Sorry, guys, but no machine with such a GPU will ever run really cool.

None of my workstation-class-ThinkPads (A31p-T43p-T60p-W500-Z60m, all 15/15.4 units) is ever allowed to go into standby. I just turn the LCD off when they're not in use, and have never seen the issue that is being referred to in this thread, although I don't doubt for a split second that it's very real.

My experiences only.

Does NHC work with Windows 7?

Does NHC work with Windows 7?

Yes, with some minor issues. See this: http://forum.thinkpads.com/viewtopic.php?f=57&t=76480

In the past year, I have looked through every T60-temp thread on this forum, as I work to optimize my own T60p for heat and power.

But very few of the temp-report posts I've seen are useful for comparisons.

Why do I say this?

Because the posted reports fail to include any data about critical test conditions such as ambient-temp and most especially CPU-consumption-%.

A report of one's cpu-temp being at 50C, without also saying how hard the cpu was working at that time, is not very meaningful.

That's because a cpu-% difference of just 25% can make a cpu-temp difference of 20-30C ! ....a big difference indeed.

And 25% differences in cpu-consumption between posters' thinkpads are very common. The normal idle cpu-% for a clean XP installation will be in the 2-5% range; whereas a system with typical OS/service and autoruns bloat may be running the cpu at 20-30% even at so-called "idle"...

It should be obvious that the ongoing great debate between the "T60's run so hot at 70-80c" camp, and the "no, no, T60's only run at 50c" camp, is in large part due to exactly this difference.....the difference in 'idling' cpu-%.

Further, the measured cpu/gpu temps will quite closely follow changes in ambient-temp. An ambient rise of 10C will typically cause a directly proportional cpu-temp rise of 7-9C. Thus, a cpu-temp report which doesn't include a semi-accurate ambient-temp condition is also pretty useless for comparing between members' machines.

Finally, it's just as critical that the Fan conditions be stated with every measurement/report. The airflow of a centrifigual fan or pump varies as the cube of the rpm. That means that if you double the RPM, you get eight times the airflow. So even small changes of fan-speed make a big, big difference.

As a point of refernce, a fan-speed difference of only 500rpm, from 3000 to 3500 rpm, can drop cpu-temp by almost 10C.

In summary, I suggest and ask that all Temp-Reports posted here at TP-forums in the future include the following info, in order to make them much more useful to everyone from now on...

- CPU and GPU temp (but preferably the whole tpfc device-set)
- Ambient Temp at time of this measurement
- CPU% as shown in Task Manager at time of measurement!
- Fan conditions (off or on, and what rpm)
- Hardware disablings; i.e, if Wifi, LCD, or HD is shut off, report it

If we all structure our temp-reports in this way, then they'll be far more useful to all of us as we move forward.

Richard

Notes:

- The various softwares used to measure temps are prone to bugs. It's important that one cross-check with a different package and/or technique. If two packages aren't giving the same temp-reading within 1-2 degrees, then something's wrong with your software-setup, and that should definitely be part of your reports.

- RMclock is known to jump 10C after 'sleep' cycles. It appears to be 'losing' its cpu temp-sensor offset value. I.e., after a fresh boot, rmclock will agree almost exactly with tpfc. But after a sleep cycle, suddenly it's reading 5-6deg higher than tpfc. After more sleep cycles, it's up to 10deg higher than tpfc. Beware. This is why you should always use more than one 'instrument' to take experimental measurements....crosscheck.

- TPfancontrol is highly subject to user-setup errors in re editing the ini file. Screwing up an offset number, or a temp-sensor address, will produce temp-readout lists that are wildly off.

- Some tpfancontrol reports posted here have shown cpu/gpu temps in the 45C range while simultaneously showing battery temps of 50-55c ! Obviously, something's badly wrong there. Either the poster's tpfc setup is screwed up, or they were fast-charging their battery while taking temp-measurements (which produces 'outlier' reports that aren't valid comparisons for other members).

Batt-charging has the same effect on internal temps as a very hot ambient-temp does. Bottom-line: make sure your battery isn't charging, whenever you take temp measurements.

- As mentioned, CPU/GPU temps will track changes in ambient-temp fairly closely; meaning the temp of the air entering the fan; which air comes from inside the case. Thus, anything adding heat to the air as it flows through the case to the fan-inlet will cause a rise in your measured cpu/gpu temps (e.g., the battery-charging example above).

Running WWAN, Gigabit Enet, and any PC-Cards, has the exact same effect as batt-charging. The moral here is that the only useful temp-report is one that mentions ALL test conditions. And also, that the tests one does before/after a change (like adding Arctic Silver) must always use ALL the same conditions, or neither the tests nor the reports will be accurate or useful.

The normal idle cpu-% for a clean XP installation will be in the 2-5% range; whereas a system with typical OS/service and autoruns bloat may be running the cpu at 20-30% even at so-called "idle"..

Apparently my optimizations are first rate, because I am running Avira anti-virus, Yahoo IM, Firefox with 10 tabs open, Comodo firewall, Clipmate, Pawclock, Coretemp, and DUmeter, not only on startup, but all the time, and my CPU utilization is 0-1% whenever I check in taskmanager, perhaps I need to sell these 100 or so XP registry tweaks I wrote!

my T60 pretty much stays at 35°-39°C when using it in a 24°C room (75°F)

wild_bill wrote: Apparently my optimizations are first rate, because I am running Avira anti-virus, Yahoo IM, Firefox with 10 tabs open, Comodo firewall, Clipmate, Pawclock, Coretemp, and DUmeter, not only on startup, but all the time, and my CPU utilization is 0-1% whenever I check in taskmanager,

perhaps I need to sell these 100 or so XP registry tweaks I wrote!
.....
(75°F)

perhaps so!

yes, that sounds like good optimization to me.

my T60p runs also runs at 0-1% when truly 'idling'...and about 2% avg while reading something while an mp3 is playing (1by1 player) and a download is going in the background (FDM/FFox). This is on win2k/sp4, although I'm getting the same results on XP in my wife's T61/intel, and our R60/intel 'connection server' running xp/ics 24/7 (at under 10 watts draw avg. )

PS: even better than selling those jewels of optimization...POST THEM !