Thinkpad X61s (Core 2 Duo L7500)

You can write down your success stories here (temperature decreasements, power savings, etc).
Post Reply
Posts: 5
Joined: Sat 28. Nov 2009, 12:53

Thinkpad X61s (Core 2 Duo L7500)

Post by Emme » Sat 28. Nov 2009, 13:39


This is really a great tool! As a first test subject, I tried it on my work laptop, a Thinkpad X61s with a Core 2 Duo L7500. I run Ubuntu 9.10 amd64. I installed the generic-phc kernels from the ppa at, linux-phc intel 0.3.2-2 and PHCTool 0.5.1.

The standard VIDs are:
9:31 8:22 6:21 136:11

As this is already a low voltage CPU, I did not expect big savings, but so far I reached these values:
9:18 8:15 6:13 136:11

Too bad that the lowest speed (800MHz) is already at the minimum allowed VID, but the savings for the higher speeds are still not bad, and might give me still some more battery life.

The next test subject is my Thinkpad X200, which I am currently testing out.

Thanks for PHC!


Posts: 6
Joined: Sat 29. May 2010, 03:25

Re: Thinkpad X61s (Core 2 Duo L7500)

Post by bluesceada » Sat 29. May 2010, 03:36


this is strange, my X61s with a L7500 shows the following default controls:

9:36 8:26 6:21 136:11
compared to your:
9:31 8:22 ...

Yours are lower by default for 1601MHz(?) and 1600MHz(?), it seems..

And actually I can change the VID for 136 to something lower, I just don't know if it really affects anything (but it is shown reading out phc_controls and phc_vids)

What kind of CPU is shown in /proc/cpuinfo for you?
Mine is this (stripped potentially uninteresting information):

Code: Select all

cpu family      : 6
model           : 15
model name      : Intel(R) Core(TM)2 Duo CPU     L7500  @ 1.60GHz
stepping        : 10
siblings        : 2
cpuid level     : 10
I'm using a Debian stock kernel with externally compiled phc-intel module:
Installed: 2.6.32-13

I didn't try how low I can go with the VIDs. I just went a little bit lower, and so far running stable, but I didn't do any stress testing so far. I will post my results here, when I've done so
(stress testing every frequency with cpuburn, and going down in voltages stepwise, until it becomes unstable, then go up 2-3 steps again - this should be the best practice to get the right voltages, isn't it?)

Posts: 6
Joined: Sat 29. May 2010, 03:25

Re: Thinkpad X61s (Core 2 Duo L7500)

Post by bluesceada » Sat 29. May 2010, 15:14

I was quite low but had some crashes not sure related to which settings it was (did test single frequencies, but then later when switching frequencies with ondemand it seemed to make problems)

So currently I am on this, could probably go lower, but no time to test now:
9:20 8:18 6:15 136:11

User avatar
Posts: 20
Joined: Wed 9. Jul 2008, 20:38

Re: Thinkpad X61s (Core 2 Duo L7500)

Post by aar3Angahqueiquibohc » Tue 8. Jun 2010, 23:23

hey bluesceada,

the reason for the different default fid vid pairs is that every cpu (including that ones from the same waver) have different production margins and so the default voltages would be adjusted for every cpu itself.

so it may happen that cpus with same type and family could have the same default vid fid pairs but its not very probable.

regards Gnom
Life is sometimes just like a bad computer game, bad story but very nice graphics.

if you have questions meet me (DonGnom) (almost daily) in #linux-phc on

Posts: 6
Joined: Sat 29. May 2010, 03:25

Re: Thinkpad X61s (Core 2 Duo L7500)

Post by bluesceada » Thu 17. Jun 2010, 02:00

Umm, but how would those then behave like the same type of CPU, being low-voltage, needing low-energy? Or are those vids then fine-grained tuning? I wonder why they are exported as changeable registers then? Hmmm...

So far I only knew about the other way around. The CPUs from one wafer are rated differently. Some become a L7500, others become only a L7400, or some ULxxxx or some Txxxx, and so on, depending on their physical analog-electrical characteristics. Otherwise they are identical in (digital) hardware design. Some get some cache disabled, because parts of it are damaged and therefore have a smaller cache, and so on ....
And for example AMDs Phenom with 3 cores exist also because of errors in production - a quad core with one core damaged (most of the time).

Some one can explain this about VID/FID in more detail? So far, I can't believe it.

Posts: 346
Joined: Wed 9. Jul 2008, 19:57

Re: Thinkpad X61s (Core 2 Duo L7500)

Post by the-fallen » Fri 18. Jun 2010, 16:36

DonGnome is right.

The default VIDs are calibrated at manufacturing time (I am sorry for not having the source of that information here but I am sure it was in one of the CPU's white-papers available from intel).

For sure, those DIE's that meets the requirements for a LV / ULV CPU may be used for that ( I am not sure if they produce them on the same wafer or if they use different ones in general ).

The registers where the default values are stored in are not changeable. But the register for the current voltage is to give the operating system the control of the power saving capabilities.

Post Reply