The "Undervolting" Guide

You're right, it is battery saver. Will undervolting work properly with that? I've never really played with my power options before... I wonder if this is affecting my laptop's performance.

 

yes, the battery saver scheme will affect the performance.
Set it to balance and do undervolting for comparison.
You also can use RMClock's Power saving profile to lock the speed at low multiplier

 

Not yet. Since your voltage curve is flat, there's no energy savings by running at less than X9. Battery time is not compromised because your CPU gets the work done more quickly and thus spends more time at idle. It's a push with any of the other FIDs..

 

I just reinstalled windows xp and got all my drivers installed. I just tried redoing the rm clock stuff and the voltages i tried using before arent even available. Anyone know why this is.

also its an asus g1s with 8600 gt i didnt change bios. so why the hell did my vid voltages get locked. i used to use 1.0 volt for 9x or 10x multiplier now im locked in at 1.05 and i cant get the lower multipliers to their old setting of .925 or whatever it was before.

 

ugh this is pissing me off. i tried hitting defaults and realized its giving me more voltages than stock. this program sucks. i have the t9300 slayy c0 stepping with default voltages of 1.1375 and it tried giving it 1.25.

 

Try re-selecting "Mobile" in Advanced CPU.

 

I followed your guide, which helped a lot! My temp doesn't go over 65-68 now.

But I have one issue... ever since I did this my fan's been getting "weird." My temperature would shoot up to the 50s from low 40s, the fan would start whirring and lowers it to the 40s. Fan stops. Temp starts going back up into the 50s/60s right after and the fan starts all over again to lower it. This goes on every 2-3 minutes and it's really annoying me to have to keep hearing the fan start. Before, my temp would never randomly skyrocket with nothing stressful running in the background unless I turned, say, a MMORPG on.

Is this because of the cpu stress test I did today and yesterday? So my laptop's all 'burned out' inside which is why the temp keeps rising after the fan stops?

Or is it because of the 'power on demand' setting I had to select? Or did I lower my V too much...? I went so far from 1.3250 to 1.18750 with no problems, the 45-minute long tests went well.

EDIT: I also just noticed for the "CPU Info" tab in RMClock, it says my current Multiplier is 6.0x and current Req.Vcore is 1.3250x. My max is 10.0x (lowest 6.0x) which is the one I changed to 1.18750 but it still says current is 6x/1.3250...

EDIT 2: Just noticed again that it sometimes switches between 6x/10x, but the 1.3250V part stays the same. D:

I really hope this can be fixed because I'm probably going to switch back to 1.3250 to prevent this since the noise is annoying.

 

Your voltages are set backwards. The highest FID is always the highest voltage. Set them all to 1.1875 for now and see if things things improve.

 

I did it, but the fan problem still proceeds. Although it does take a little longer for the temp to rise, it still happens. Should I try lowering each temp by .0025 or something as I go from bottom to top, like the example in the first post?

Thanks so much for trying to help!

 

1. Undervolting starts at the top FID and works its way down. The next step will always be of equal or (usually) a lower voltage.

2. Your fan is thermostatically controlled and has nothing to do with undervolting. Your CPU will always heat up and the fan will always click on to cool it back down to the low temp. setpoint. Undervolting just delays the time it takes for the CPU to reach the "I need cooling" point.

3. If you have a really noisy fan either return the computer or try to find if you can clean or muffle (caution) the fan

 

1) So should I start making them lower from top to bottom? Haha, sorry, didn't quite catch what you said - I'm not computer savvy at all. xD

2) That's odd... before, my fan has never clicked on to lower my temp unless I reached something like 85C, which rarely happened. Now it's clicking on at 50. Max temp is 65C for me thanks to undervolting, which is great! I'm still confused why it keeps skyrocketing up after the fan cools it down despite me not doing anything, though. =/

3) I've already cleaned the fan two weeks ago since my laptop was constantly overheating and shutting down. That stopped the overheating.

 

are you sure your fan was off until 85c before?
because usually the fan has speed, like : off - slow - medium - high

I'm guessing your fan was usually in slow mode everytime,
and now bcause your temp is drop, your fan goes off until 50 and then
goes medium/high speed.. it's something that ACPI controls, not rmclock.

 

Yeah sorry, it probably wasn't off (was on slow speed, not off) but at 85 the noise got more noticeable so I guess the fan speed picked up. I suppose I'll get used to the noise but I'm just wondering why the temp starts spiking after the fan turns itself off.

Then again, if I'm remembering correctly, I think my fan was on at least slow mode 100% of the time before I tried this, so that's why the temp. wasn't spiking up... when it gets low now the fan turns off completely which is why it may be rising. Just what I think.

 

you can try to use speedfan or NHC to control your fan speed,
if it's not working, I think it's worth it to try this:
http://forum.notebookreview.com/showthread.php?t=413380

 

Sadly speedfan isn't compatible with Inspiron, which is what I have.

And I'm a bit worried about trying that one, especially seeing the word "hack" in the title lol..

 

then you're good to go with i8kfanGUI, I've been using in my vostro

 

I have 18kfanGUI, but I believe it doesn't work on Inspiron 1525 either.. I've tried messing with the fan settings there but nothing ever changed. D:

 

boomer.. i guess it's not working for newer models

 

I think you are corect. Good observation. Two things you can do:

1. Since your voltage curve seems flat, there's no energy savings by using the lower FIDs. So, undervolt the top one and uncheck the rest. The heat range leveling may keep the fan running in mid-range and not be so spastic.

2. A long "lag" period between fan speed changes strongly suggests poor heat transfer. Re-do the thermal paste job with great care and a good paste. Using too much is worse than none at all.

3. After trying 1 and 2, you may have to limit your top FID to control the max. load heat flux. Then, send a thank you note to HP. Your model has a rep for running hot.

 

I just read this guide last night and followed the instructions. My laptop is a Toshiba Satellite L355 with a T3400 CPU. At the maximum multiplier of 12x, I got it running at 0.975 volts, which dropped temperatures from a blistering 85(!) degrees at full load to a much cooler 62-63 degrees. However, my problem is that knowing the CPU is stable at full load at 0.975 volts, I'm sure I can go much lower with the lower multipliers. However, RMClock offers no options below 0.95 for any of them. Is there a way to get it to allow me to select lower voltages?

 

Possibly another program, I have never tried it, but its a start. Check out TevashSzat's post and scroll down to uv'ing: http://forum.notebookreview.com/showthread.php?t=401345

 

It's not RMC. Your bios/mobo is locked by mfr to prevent bsod returns under warranty.

Why are you bothering with other FIDs when you can't exceed 63C at flat out operation ?

 

I just thought it would help with battery life. If it's not going to make a difference, I won't bother.

 

If your voltage curve is flat, twice the work at top FID is done in half the time. It's a push vs. low speed and longer proc. time. Speedswitching for battery lfe is an Urban Myth if all your VIDs are the same. Older CPUs had exponential voltage curves so running slow paid off. But not with new CPUs, unless you venture into overclocking territory. But that's a different game.

 

I have an old HP dv5074 which has got an AMD turion ML-34. I have there these:

4x normal 0,9 | Undervolted to 0,75
5x I Don't use
8x I Don't use
9x normal 1,3 | Undervolted to 1,025

Temperatures went from idle 50+ and fan starting kicking in (low), to a pleasant ~39oC no fan. And from 65oC in heavy use and high fan rpm to somewhere of 55oC and fan on low.

 

Nicely done !

 

i just ordered a HP notebook with the new i5-440m processor. would rmclock work on this platform? it will be here in a few days...

 

I don't think rmclock supports it. You would wanna try cpuGenie though

 

any one try to undervolting a vaio? im using z11 so if you have done 1, please share your experience in these steps.
thx

 

Yeah, VAIO AW undervolted with CPUID (make sure you press F4 while it's running).

Cheers!

 

Just a question, in CPU info, why is it showing Unknown Core in CPU Core, Unknown in Revision ? Mine's P8600

 

Your CPU is newer than the last rev. of RMC. It still will work fine.

 

So, just watching my RMClock in the task bar (hovering over it to get details), and it keeps bouncing between the custom voltages I have set (1.0125 at 12x) and the default (1.2125 at 12x). Just started doing this today, as far as I have noticed, and I didn't change anything. Any clues/ideas?

 

W7 has stability problems with RMC which was completed 3 years ago. I did two things:

1. Uncheck every adv. cpu feature except TM1 and hook on TM1.

2. Disconnect RMC power management and modify a W7 power plan to your likng. RMC is now in charge of FID and voltage only.

 

I'm running Power Saving profile in RMclock with 6x running at the minimum voltage given 0.925v and my temp is 39. I don't see any difference of running this RMclock.

 

Of course u wont feel anything on the lowest FID and VID.. lol.. the difference only can be felt when FID is at max and VID a little lower than the max VID ..

 

Does CPUz report 0.925V and what CPU do you have?

 

So that means I won't experience any changes in the lowest F/VID ?

Yup, both CPUz and RMclock reported that. I have P8600

 

yes in some (but not always) the stock idle voltage is already the lowest VID

 

Yes you wont have any changes in the lowest FID.

But the max FID should be around 10C cooler.

 

I have a P8600 too Max FID 9.0x and VID at 0.9875V

@scott.ager

How to disconnect the RMClock power management? Im having almost the same problem as him.. I can see the voltage and clocks go over the Standard clocks and voltage..

P8600 2.4GHz 1.1125V

i can see it sometimes @ 2.53GHz 1.2000V ???

And upon restart/ startup pc RMClock would always be default lol.. weird..

 

Absolutely correct

Uncheck "Enable OS power management integration" box in Management window.

Then select and modify the Balanced plan in W7

I haven't seen anyone totally fix the random IDA voltage spikes that start happening after long waits at idle. But, they go away as soon as you start an application. It's RMC arguing with W7. Doesn't seem to hurt, just annoying.

 

I'm hopefully not trying to repeat any issues previously stated here but whenever I shutdown or let my computer enter sleep mode my undervolting settings are lost. I'm running Windows 7 64bit and the base multipliers are appearing as 8.5x at 1.350v. I keep lowering the values to 6.5 or 7.0 at 1.1750V, but they return to normal after restart. I've been using CPU-Z 1.51, RM Clock, and Crystal CPUID. I added task schedules to run the programs at startup, so I must be missing something. I'm not usually great with this stuff but so far the guide has been very helpful. Any ideas are well appreciated.

 

If you have worked around the UAC (User Account Control) issues with RMC so it loads correctly at startup, may I humbly suggest NOT auto-starting the others. Their drivers fight amongst themselves and with W7 too.

 

The UAC does stop the process until I tell it to stop doing such (I hate UAC). Any way to stop it from doing such? Is RM Clock or Crystal CPUID better and which should I kill?

 

There some excellent posts on this thread regarding RMC and UAC. Lots of people, like me, are the sole users of their laptops just turn UAC off. Otherwise, "Search this Thread".

I prefer RMC myself. The drivers do conflict so pick one or the other to load at startup.

 

Is it normal to idle around 40C with a C2D T8100? I normally just have explorer.exe, firefox and songbird open. Nothing really demanding.
CPU clock: 1256.82 MHz
FID: 6.0x, V: 0.9500V

Thanks in advance.

 

I'd say it's time to do cleaning job and changing paste

 

Not normal. My t8100 idles 1C over room temp. (right now it's at 27C). It's summer here in Australia.

 

Hi all.

For those who are interested in downvolting an DELL XPS M1330, just have a look here:

http://forum.notebookreview.com/showthread.php?t=453580

I have written this yesterday. These are my best values I could find to cool down the CPU and a little bit the GPU.

I hope that I can somewhat contribute ...

Cheers.