View previous topic :: View next topic |
Author |
Message |
numlock n00b
Joined: 17 Sep 2004 Posts: 58
|
Posted: Sat Aug 16, 2008 11:24 am Post subject: == Patch to reduce notebook cpu temperature: -20°C for me == |
|
|
Hi all,
For 2 years, my Acer Ferrari 4005WLMi notebook has been running with a custom patch.
Here is the current version, which applies cleanly to linux 2.6.26-gentoo-r1:
Code: |
--- linux/arch/x86/kernel/cpu/cpufreq/powernow-k8.c.orig 2006-02-01 01:12:32.732294624 +0100
+++ linux/arch/x86/kernel/cpu/cpufreq/powernow-k8.c 2006-02-01 01:20:10.560694128 +0100
@@ -863,6 +863,8 @@
fid = data->acpi_data.states[i].control & FID_MASK;
vid = (data->acpi_data.states[i].control >> VID_SHIFT) & VID_MASK;
}
+ const u32 cooler_vid = 0xC; /* 0xE = 1.200 V, 0xD = 1.225 V, 0xC = 1.250 V */
+ if (vid < cooler_vid) vid = cooler_vid;
dprintk(" %d : fid 0x%x, vid 0x%x\n", i, fid, vid);
|
Startup log (WITHOUT PATCH):
Code: |
Aug 15 21:21:33 killdozer powernow-k8: Found 1 AMD Turion(tm) 64 Mobile Technology ML-37 processors (1 cpu cores) (version 2.20.00)
Aug 15 21:21:33 killdozer powernow-k8: 0 : fid 0xc (2000 MHz), vid 0x4
Aug 15 21:21:33 killdozer powernow-k8: 1 : fid 0xa (1800 MHz), vid 0x6
Aug 15 21:21:33 killdozer powernow-k8: 2 : fid 0x8 (1600 MHz), vid 0x8
Aug 15 21:21:33 killdozer powernow-k8: 3 : fid 0x0 (800 MHz), vid 0x16
|
Startup log (WITH PATCH APPLIED):
Code: |
Aug 16 10:26:42 killdozer powernow-k8: Found 1 AMD Turion(tm) 64 Mobile Technology ML-37 processors (1 cpu cores) (version 2.20.00)
Aug 16 10:26:42 killdozer powernow-k8: 0 : fid 0xc (2000 MHz), vid 0xc
Aug 16 10:26:42 killdozer powernow-k8: 1 : fid 0xa (1800 MHz), vid 0xc
Aug 16 10:26:42 killdozer powernow-k8: 2 : fid 0x8 (1600 MHz), vid 0xc
Aug 16 10:26:42 killdozer powernow-k8: 3 : fid 0x0 (800 MHz), vid 0x16
Aug 16 10:26:42 killdozer powernow-k8: ph2 null fid transition 0xc
|
(the "null fid transition" warning is displayed because several p-states share the same voltage)
In plain English (using the AMD formula: Voltage = 1.55V - 0.025V * vid) this means:
Code: |
FREQUENCY 2 GHz 1.8 GHz 1.6 GHz 800 MHz
without patch: 1.450V 1.400V 1.350V 1.000V
with patch : 1.250V 1.250V 1.250V 1.000V
|
In short, this patch will "undervolt" your AMD K8 cpu, while keeping it at the same frequencies.
What matters most, is the temperature at full cpu load. In this case, we go down from 1.450V to 1.250V.
Do you expect a small impact on the temperature ? Well, on my 3 1/2-year old Acer Ferrari 4005WLMi, after air-cleaning (to get rid of the dust), I get:
Code: |
no patch : ~ 95°C and still rising
with patch: ~ 75°C stable
|
This great trick saves power, makes the notebook 20 °C cooler and (best of all) prevents hardware thermal shutdown (happens at ~100°C).
Just one thing: if you try it on your PC, you need to adjust the "cooler_vid" variable to find the lowest working voltage. Run burnK7 or other CPU-intensive apps. If the voltage is too low, your PC will crash, so you should *not* have your home directories mounted, and you should do this in runlevel 1.
In my case, I tested for stable operation at 2 GHz, and found the minimal voltage to be 1.250V. So I put this value in the cooler_vid variable. Please note that in my case, the new voltage is quite low, even lower than the original 1.8 GHz & 1.6 GHz modes. These lower frequencies *WILL* also run at the lower voltage, and they *WILL* work as stable as the top frequency. So obviously, you need to test the highest frequency only.
Please report your experiences with this
Last edited by numlock on Sat Aug 16, 2008 12:44 pm; edited 1 time in total |
|
Back to top |
|
|
Monkeh Veteran
Joined: 06 Aug 2005 Posts: 1656 Location: England
|
Posted: Sat Aug 16, 2008 12:18 pm Post subject: |
|
|
If I were you, I'd get your heatsink fixed. A 35W CPU shouldn't hit 95C. |
|
Back to top |
|
|
numlock n00b
Joined: 17 Sep 2004 Posts: 58
|
Posted: Sat Aug 16, 2008 12:59 pm Post subject: |
|
|
Monkeh wrote: | If I were you, I'd get your heatsink fixed. A 35W CPU shouldn't hit 95C. |
Thanks for your response.
I agree totally with you, but my dissipator is well seated. The problem is the very fine dust that deposits on the dissipator. In the first years, removing the keyboard and blowing air on the dissipator did improve things A LOT (ie: things went back to normal, ie: 35-45°C).
But after 4 years of very intensive use (almost 25'000 hours of fan blowing air), I think most of the dust can't be removed anymore (at least by me blowing air), and I'm easily at 70-80 °C (with the patch).
I've disassembled the laptop, but getting access to the dissipator for cleaning, is difficult.
Maybe I should try to "brush" the dissipator from the outside, using something long and thin ? Or use an air compressor instead of blowing ? |
|
Back to top |
|
|
termite Guru
Joined: 06 May 2007 Posts: 466
|
Posted: Sat Aug 16, 2008 1:10 pm Post subject: |
|
|
Yes, an air compressor will help. |
|
Back to top |
|
|
jcat Veteran
Joined: 26 May 2006 Posts: 1337
|
Posted: Sat Aug 16, 2008 6:24 pm Post subject: |
|
|
Nice. Using less power can only be a good thing anyway. If we can make hardware run more efficiently _and_ remain stable, then why not
Cheers,
jcat |
|
Back to top |
|
|
Mad Merlin Veteran
Joined: 09 May 2005 Posts: 1155
|
Posted: Sat Aug 16, 2008 9:36 pm Post subject: |
|
|
Try a vacuum cleaner on one of the external vents, that way you actually get rid of the dust instead of just moving it around inside the case. This did wonders for one of my older laptops. _________________ Game! - Where the stick is mightier than the sword! |
|
Back to top |
|
|
SeaTiger l33t
Joined: 22 Nov 2007 Posts: 603 Location: Toronto, Ontario, Canada
|
Posted: Sun Aug 17, 2008 6:26 am Post subject: |
|
|
When using vacuum cleaner or air compressor to clean heat sink and especially a fan, try stabilize(stop the fan from spanning) the fan first. Reason is high speed air flow may damage the wings of the fan and second, the fan become a generator and create current to the motherboard and speed sensor. Also try to ground the laptop (keep the power adapter connected) as high speed air flow can create static along circuit board and plastic parts. |
|
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
|
|