Undervolting!

[Frank M]

Huh, a mobile subforum and no word about undervolting?
Ok, you can make it Xtreme Undervolting! to better suit XS :D

For mobile, I think undervolting is more relevant than overclocking; maybe even
underclocking, too, to stay mobile and not to have to rely on chargers for a
longer time.

I got my first adjustable-voltage and -multiplier cpu yesterday in the form of
a C2D T5500 in a FuSi Lifebook T4215 tablet. Has a neat case with which it
can be used as a tablet even without taking it outside of the case; however,
it gets a bit hot in the case. And as always: battery life. It's quite good on
that front, just a 12.1" screen, 945GM, and a good 56Wh 6-cell, so with
wifi&bluetooth off, windows predicts 5hr. But that's idle..

So let's see the voltages!
10*166, and RMClock says VIDs equal to 0.95V-1.25V.
Cpu-Z however says that the Vcore is 1.063-1.363 - I assume the board
overvolts by 0.1125V.

I've got down to an Orthos stable 1.66G with just 1.025V (cpu-z: 1.138) :D
Unfortunately the 0.95 can't be set any lower, but this is already LV territory.
A 0.225V drop is fine too :D And I still didn't reach a bsod, maybe it can go
even lower :yepp:

Load temps also decreased significantly; I can't tell how much exactly, as
the stock at 75+C was too high for me and I quickly stopped Orthos (it was
still climbing); and also the ambient temps changed as it's nighttime; but
now it stabilizes at 67-68C. :cool:


So anyone else want to share undervolting experiences? Pinmods? :D

---

Update 2009.11.06:

I just came upon a small hack for RMClock to allow lower VIDs:
http://forum.notebookreview.com/showpost.php?p=5312927&postcount=87


Unlock hidden VIDs:

-> Start –> run -> regedit
-> HKEY_CURRENT_USER -> Software -> RightMark -> RMClock

Create new DWORD , rename it to UnlockVid, value 1.


I got an extra VID, now 0.9375V is the lowest instead of 0.95V.
I also tried a DWORD called UnlockFID, but that didn't have any effect.

[Savuti]

With a T5600, XPS M1210, via rmclock, was able to bring it down to 998.7 MHz @ 0.9500v. I get just a tad under 3 hours with a 6 cell, 20% brightness, Aero off, bare minimum services, disabled devices ect, this is with light browsing and word processing. With wifi off it'll be likely close to ~3.5 hours. Temps around 48c.

For some reason in Vista when Aero is on the lowest mode the 7400 switches to is 2d, with Aero off it goes to desktop mode @ 100/200 clocks. This gpu chugs, not sips, through battery power :)

[zanzabar]

u forgot the best part about underclocking/volting, the more power that u leave from psu the more power that other things can use like grafix cards

if u drop .1V with the new xps or gateways with a single gts u can oc the grafix card 35-40% were without droping th ecpu voltage u can only get like 10%

[Frank M]

With a T5600, XPS M1210, via rmclock, was able to bring it down to 998.7 MHz @ 0.9500v.

Isn't that the default setting though? ;)
At least for mine it is: 6*166=1G @ 0.950 (cpu-z: 1.063)
max by default: 10*166=1.66G @ 1.250V (cpu-z: 1.363)


I'm hitting new lows :D
I ran out of VIDs :( :eek: :D

So far 15mins orthos stable 1.66GHz @ 0.950V (cpu-z: 1.063V)!!! :eek::cool:

Cheers:
Frank

[Martijn]

You guys might be interested in trying NHC, notebook hardware control. It allows for voltages as low as 0.7v. One thing though: it doesn't support as many notebooks as RMClock, so your lappy might not be supported. I used to run .7v on a 1.86Ghz Sonoma @ 800Mhz.

[Repoman]

Pentium M 1.86GHz on my IBM Thinkpad -

800MHz@0.7v
1.86Ghz@ 1.02v

All Prime stable via rmclock!

[Malk4vi4n]

This is my undervolting experience with Asus G1S: http://lab.gamer.com.br/2008/03/25/o-notebook-g1s-se-sente-melhor-14%C2%B0c-mais-frio

[H2omg]

This is a little off topic, sorry. I was wondering if anybody knew why it is that Mac's have longer battery life than PC's. It's not out of the question for me to get 4.5-5 hours battery life when most PC's seem to die after 2.5-3 hours. I'm assuming it's simply better power management by the OS?

Also, I've never heard anybody refer to Fujitsu as FuSi before O_o

[Newblar]

i think intel locked the vcore to above .9 on these c2d's
but the old p-m dothans and the like are underclockable to .7v

[Cheffy]

been doing this for 3 years with my lappy, rmclock 1.5 always worked best for me.

my (old) laptop does well it seems,

clawhammer 3400+ 89w dtr, stock is 2200mhz @1.5v.
runs 1.175v @ 2000mhz and 0.8v at upto 1200mhz (lowest is 800mhz)

sadly the Kt800 chipset and no-powerplay-functonality 9700 mobility and 17" screen still chew power madly - so 2hrs is about all the somewhat abused battery can provide (~55 of the 68w/hr capacity remains)

[ghost101]

Did this also when I first got my laptop 2 years back.

http://img.techpowerup.org/080830/untitled.jpg

Normally its 1.262v at 1.66ghz.

The chip is a t2300. These are the minimum settable voltages and everything is stable.

After 2 years of use I still get 45whr of the original 85whr capacity.

[Xcel]

I'm Running my Pentium M @ 800MHz 0.7v, 2GHz 1.068v

[4Qman]

My Laptop was Load at 1.305v. now load at 1.063v Prime stable with 12oC off the Temps. :D

I need lower Vids.

nice thread.

[Frank M]

Nice to see the thread got moving while I was away


You guys might be interested in trying NHC, notebook hardware control. It allows for voltages as low as 0.7v.

Tried it, but it only gave the same VIDs.
I think it's chipset/bios/factory settings-dependent.


This is a little off topic, sorry. I was wondering if anybody knew why it is that Mac's have longer battery life than PC's. It's not out of the question for me to get 4.5-5 hours battery life when most PC's seem to die after 2.5-3 hours. I'm assuming it's simply better power management by the OS?

Also, I've never heard anybody refer to Fujitsu as FuSi before O_o

It's also slightly better power management, but for the bigger part, it's other
things. windows like to access the cpu and hdd occasionally, therefore
preventing them from powering down. Os x is based on unix, which handles
this better. But my guesstimate is that this accounts for less than 10%
difference.
The main difference is: power supply, power conversion, voltage regulation,
chip voltage settings. Lower power loss and power consumption. next up is
higher capacity battery. Then you also have a smaller screen (13.3 vs 15.4),
which also has less consumption.

Around here, we regularly abbreviate Fujitsu-Siemens as Fu-Si :)


i think intel locked the vcore to above .9 on these c2d's
but the old p-m dothans and the like are underclockable to .7v

I think they locked the vcore minimum on 94xG not to hurt their LV and
ULV chip sales. Those can access the lower VIDs, even down to 0.7V.
But I suspect they need a bios refresh, so it's more on the sw level.
In 965 chipsets, though, there's a 0.875V or something similar low-voltage
idle setting for T7xxx and above cpus.



After 2 years of use I still get 45whr of the original 85whr capacity.

Let me guess: you had your battery in all the time, even when using it from
the mains for prolonged sessions? ;) My old lappy that this replaced, also 2
years old, still manages about 80-85% of it's original running time. All I did
was to take the battery out when it was fully charged if I was using it
plugged in for a longer time. Then when I needed it mobile again, I just
put the battery back.

[afireinside]

I undervolted my T5600 to 0.95v as well. CPUz reads the same 1.063v here. CPUz has never been the most accurate in the world with voltages though. I get 4 hours of battery life on my XPS M1210 with wifi on and brightness at 5/7 :D 7400GO underclocked all the way in riva tuner, aero disabled, 9 cell battery.

[fragmasterMax]

This is a little off topic, sorry. I was wondering if anybody knew why it is that Mac's have longer battery life than PC's. It's not out of the question for me to get 4.5-5 hours battery life when most PC's seem to die after 2.5-3 hours. I'm assuming it's simply better power management by the OS?

Also, I've never heard anybody refer to Fujitsu as FuSi before O_o

mac's have bigger batteries
(i could totally incite a mac vs pc battle here but wont, except my 2cents is that macs suck)

[acidpython]

mac's have bigger batteries
(i could totally incite a mac vs pc battle here but wont, except my 2cents is that macs suck)

Hmm that could be interesting, how would the OS power management differ between say linux and windows? I noticed myself that xubuntu on an old laptop of mine had a slightly shorter battery life. Nothing notable really though just wondered if anyone experienced anything substantial?

[EneergE]

I've been undervolting all day on my new ASUS G50V. It's a centrino 2 c2d T9400 @ 2.56ghz, 1066fsb, 4gigs ddr2-800 memory, 9700gs, 2 sata2 drives raid 0.

At 100% load with prime95, I'm able to still bring the voltage down to 0.9625v and remain stable for at least two hours. I have not tested it further than that, yet, but that's pretty stable. I hope the battery life improves significantly as well, but I have not tested it yet.

I'm trying to figure out a way to undervolt the video card, but I haven't had no luck, yet.

[ziddey]

3.5 years ago I got a dothan 715. stock was 15x100=1500 1.325v iirc. pinmodded up to 15x133=2000 (760 equivalent, or highest end at the time). 6x133=800 @ 0.7v, 15x133=2000 @ 1.163v.

now, I've got a compaq with a t2370 and a 47Wh 6cell. Doing 6x133=800 @ 0.95v, and 13x133=1733 @ 0.95v as well. So I've hit lowest vid at highest fid. Next up would be a pinmod for lower voltages. Hopefully I'll be able to run in 0.5v to 0.9v. As is, I'm already seeing around 3 hours with fairly high brightness. Stock, I got around 2 hours with brightness reduced a fair amount. And with this model, I can use higher capacity batteries from other hp's/compaq's. So once this battery dies, it's 12 cell time! Hopefully I'll be in the 6 hour zone. Never had a laptop with that much life before. Excited :)

[Frank M]

Next up would be a pinmod for lower voltages.

If you do try that, tell us how it went.
I once tried to pinmod a Yonah-core CelM440, and to isolate a pin, I tried to
put a piece of tubing on it (I didn't want to break the leg off, not to ruin it) --
but the pins and the socket were so small that I could not get it to fit in when
I put it back -- and that was the thinnest piece of wire wrapping that I could
find :/
Moreover, I'm not sure that this would work without a bios-update/change.

Dunno about over there, but here, batteries are bloody expensive. I thought
about getting the second mediabay/multibay battery for my lappy, but I
could not find one, and a replacement primary battery was $200 :eek:
I'd rather buy a netbook with a high capacity battery, they can manage
over 7 hrs.

[ST4]

Dell Latitude D630

Intel Core 2 Duo T7300 - RM Clock controlled
Low: 1197MHz @.963V
High: 2000MHz @ 1.15V

Standard: 2000MHz @ 1.3375V

[omga14]

Very interesting. I have never owned a Laptop till not after seeing how useful my wife made hers. I was especially impressed with the temperature drops Malk4vi4n experienced. I'm looking for more time and lower temps. We'll see just how well I do. Also never thought to take the battery out if I was running off the wall. Didn't know if that was ok for the Laptop or not.

[luie]

Using NHC on Motion LE1600 tablet,

CPU = Pentium M 758 LV (1.5Ghz~600Mhz)
Specs:
Coreclock = 1.5Ghz
L2 cache = 2Mb
FSB = 400Mhz
Die = 90nm
Power (TDP) = 10W
Voltage = 1.116V
Core = Sonoma

Lowest stable voltage:

Multiplier---Voltage---Power (TDP)
6----0.700---1.6
8----0.732---2.3
10---0.796---3.4
12---0.876---4.9
15---0.988---7.8
15---1.116---10 (stock setting)

[STEvil]

Using NHC on Motion LE1600 tablet,

CPU = Pentium M 758 LV (1.5Ghz~600Mhz)
Specs:
Coreclock = 1.5Ghz
L2 cache = 2Mb
FSB = 400Mhz
Die = 90nm
Power (TDP) = 10W
Voltage = 1.116V
Core = Sonoma (2nd generation of Intel's Centrino Mobile Technology)

Lowest stable voltage:

Multiplier---Voltage---Power (TDP)
6----0.700---1.6
8----0.732---2.3
10---0.796---3.4
12---0.876---4.9
15---0.988---7.8
15---1.116---10 (stock setting)


Centrino is a platform (chipset, wireless, integrated GPU iirc), not a CPU core.

[luie]

oops. sorry. I copied that on Wikipedia while looking at additional info on the Sonoma. :-P

[cc-tuner]

Well got my lappy a few days ago, deside to undervolt it because of the "short" battery life...
IDLE is like 2,5 hours and @ LOAD is like 1,5 hours :( according to battery indicator.

Anyways, doing orthos stress test as we speak, from 1.363v to 1.100v on a c2d T2390 core name= Merom.

Stable for 48 minutes at the moment and whily i'm typing this it looks like it's going to be a very long day :comp10: .

[jcool]

Nice thread ;)

T2400 here (1,83Ghz, FSB 166, Yonah Core, 2MB L2) in a well-aging IBM X60. Stock range 0,95-1,26V, does 1,83Ghz primestable at 0,975V, FID 10x and below all run on 0,95V. Still runs 6 hours on the 8 cell battery, even though it's down to 56wH capacity and I replaced the stock hdd with a 7200 rpm drive. Awaiting new SSDs, should save a watt or two.

[WhiteFireDragon]

is it possible to undervolt the intel atom? i tried but it looks like RMclock does not support the atom yet :(

[luie]

maybe when the next RMclock update. :-)

[motopen1s]

Hey guys, is there anything else available other then RMclock and NHC to play with voltage/cpu speed on the laptop?

I just received a T9500 ES for my Acer Aspire 5920 (upgraded from T7300). It seems like the multi is unblocked to x14 (instead of default x13). Basically, it speeds up to 2.8ghz, which is not a bad thing. :)

However, the voltage is all over the place (monitoring by Everest v5.01.1700 and CPU-Z): jumps from 1.00v up to 1.33v. While running OCCT, the temps get to 80°C and laptops turns into a blower.

Currently I want to find the minimum vcore and set limit for it as well as throttle parameters/cpu speed. Now running Windows 7 Beta but updating to RC within next couple of days.
I hope that RMclock will allow me to do this, since NHC seems to have no support for Vista/7. I haven't read the manual yet, but the 10 minutes of messing around with RMclock didnt seem to have any effect (will update on this soon).


Update #1:

1st results: laptop passes OCCT @2.6ghz with VID @1.1, temps reached 74°C with open back plate and a laptop couple cm lifted above the table; 1.0 VID @2.6ghz - no go. Will try to lower further. Also, haven't figured the way yet to use anything lower then 1.0 for VID in RMclock.

Will continue testing and looks like will be purchasing RMclock Pro later on.


Update #2:
2.6ghz OCCT stable at 1.05v. Max temps reached 71°C couple of times, at average 68°C. Not bad at all I would say. Will try lowevering voltage even more then testing other multies.


To complete the laptop madness - Intel X-25M is on the way.

[motopen1s]

Alright, after series of stability tests the VID/Frequency is as following: @2.6ghz - 1.05v, @2.4 & 2.2ghz - 1.025v, @2.0ghz and below - 1.00v.

Unfortunately I havent found the way to use anything lower then 1.00 VID in RMclock. I set the laptop to run @2.0-2.6ghz when powered and 1.2-2.0ghz when on battery.

With Intel X-25M (80gb version) the laptop feels as snappy as my main rig with VelociRaptor and quad@3.6. In fact, in many cases (like opening applications) - as expected, it is faster. Laptop is used mainly for different office applications, some graphical work, music reproduction & a bit of encoding/decoding and browsing. I am running now Windows 7 RC x64 with 2gb of RAM, but 4gb are on the way.

Due to SSD I have to get used to very small time required for reboot. It used to be enough to go get a cup of tea, and now it's almost not enough time to lift my ass from the chair. Loving it. :up:

Do I need such a beefy laptop? Maybe not entirely. Do I enjoy it? Hell yes.
Now I will definitely wait upgrading q9450 and better get an SSD when new Intel drives will come up (also mtrons and others). Then the aging GTX280 will go... and only then, I believe it will make sense to look for CPU upgrade (as far as I see it). Time will show how things progress.


If anybody found a way to opt for less then 1.00 VID, please drop a line. ;)

[BlindFreddie]

P4 530 Prescott 3.0GHz, Gigabyte GA-8IPE775 Pro (LGA775), 512Mb DDR400 3-4-4-8, ATI Radeon 9250 AGP 128MB gfx.

I just bought this old desktop machine for a mate who just wants something to type up his education course assignments, and have been giving it a bit of TLC before handing it over. I didn't have the opportunity to check out much before buying it, but when I got it home I found it was running at about 65C idle! This is the first P4 I've ever driven, but I know they have a rep for running hot and thirsty.
I removed the Intel HSF and CPU, scraped off the grey clay, put a bit of 2000-grit wet'n'dry on the glass-topped dining table, added a bit of kerosene, and gave the CPU a good rub. Whoops - voided the warranty :( Reassembled with a thin thin smear of AS5 on each surface. Now about 45C at idle. The only forced ventilation of the case was via the feeble PSU exhaust, and my only easy addition was a 60mm fan exhausting from the back of the case, giving about 2C temp reduction under load.

I remembered reading something about undervolting here on XS. Now, the high end chips of this P4 5xx series run at 3.8GHz, so the 530 core is way underclocked at 3.0, so I thought perhaps it will undervolt too. Auto volts in BIOS are 1.3875. I'm now at 1.2000 BIOS, 1.11 load, Prime95 stable, and have yet to hit bottom. P95 load temp is 47C, which drops to 41C if I take the lid off the case.

There must be literally millions of these old P4s out there, cooking away, pumping totally unnecessary tons of CO2 into the atmosphere. Let's encourage people to undervolt them (unless they want to increase their performance - but then 25% more nothing is still - you guessed it - nothing). Spread the word.

And report your min stable Vcore. BTW, the RAM runs at 2-3-3-7 (2.5V), set with memset as there's no adjustment in the BIOS.

[Oldguy932]

I'm working on my new MacBook Pro(quit the snickering) with Windows 7 RC x64. Right now the T9900 is priming at 3.06 with 1.2v. Minimum setting allowable is one above 1.0v so I can't really do much with RM Clock. If anyone knows of another program that's been recently updated I'd be grateful.

[FrAN_oC]

I already run my old Pentium M dothan 1.6Ghz @ 0,94v. The battery lasts almost 1 hour more than default settings.

[dennisjai215]

just picked up another laptop, T2390 @ 1.86ghz 1.25vid - 1.36v~ windows (cpu-z).

Undervolted to 1.0vid - 1.113v windows @ 1.86ghz, temp drop 14C @ load. (priming for 10 minutes before i undervolt some more)

However, the lowest VID setting in RMclock only allows for .9v, which shows up as 1.013v in windows, which is useless since that is the default. If anyone know of a program that can undervolt it more, please share, thanks.

[Smartidiot89]

is it possible to undervolt the intel atom? i tried but it looks like RMclock does not support the atom yet :(
Honestly you won't gain that much... Its the god awful useless piece of crap built on 110nm chipset in it thats the true powerhog that consumes x4-x5 as much as the Intel Atom. Is it possible to undervolt the 945 chipset?

[Bas3onAc1d]

Hm, I was able to undervolt my dell E4300's SP9300 down to 1.0V under load but had to underclock it by 100MHz as the program doesn't support .5x multipliers. I did notice better battery life but I'm curious to know if this is below average for this processor or not.

[Frank M]

Unfortunately I havent found the way to use anything lower then 1.00 VID in RMclock.
If anybody found a way to opt for less then 1.00 VID, please drop a line. ;)

Minimum setting allowable is one above 1.0v so I can't really do much with RM Clock. If anyone knows of another program that's been recently updated I'd be grateful.

Sorry, no can do. It's not a limitation of RMClock, but of the bios features /
chipset. On older Pentium M ULV machines, RMClock can go as low as 0.7x V.
The other programs I've tried show only the same VIDs.


just picked up another laptop, T2390 @ 1.86ghz 1.25vid - 1.36v~ windows (cpu-z).

However, the lowest VID setting in RMclock only allows for .9v, which shows up as 1.013v in windows, which is useless since that is the default.


Your two sentences seem to be contradicting each other :p:


Honestly you won't gain that much... Its the god awful useless piece of crap built on 110nm chipset in it thats the true powerhog that consumes x4-x5 as much as the Intel Atom. Is it possible to undervolt the 945 chipset?

Even worse, it's actually 130nm.
945G chipsets with lower voltage are the 945GS and 945GU versions, setting
the voltage most probably requires bios modifications if it can be set at all
and is not hardwired. Me too, I'm very frustrated that lower processor VIDs
and chipset, memory voltages cannot be set manually using an open,
accessible bios like on desktops.

[dennisjai215]

To clarify regarding my statement. When my computer is set to "power saver" it underclocks and undervolts to 800mhz & .9 VID (shows up as 1.013v in CPU-Z). RM Clock's lowest VID setting is .9VID, so that is why I say it's useless because at the absolute lowest setting on laptop & RM clock, I cannot get it lower than .9VID. So i'm wondering if there is anything out there to further undervolt? TIA.

[topdog5050]

ok.

[Frank M]

Read my post just above your reply...

[DragoonXX]

anyone have any luck undervolting the chipset on an asus EEE 1000HE?
I know undervolting the proc is moot (heck, I can't even get a reliable reading for temps or voltages on mine...all I know is my N280 is on the higher end of the VID range)
and the only other real power savings I could do would be to go with an SSD, and they're too expensive atm to be worth it

[kiwi]

Merom 7400

idle no change, 6x166 0.95V
load, 13x166 @ 1.025V instead of 1.2V

Load temps down from up to 100 to ~85

BTW, it shuts down at 100 already, I should not run occt on laptop :/ Well, after lowering volts I can :)

[jcool]

85C at 1V? Wow, your cooling sucks :D

[kiwi]

Yea, it's Toshiba P100. I replaced stock thermal paste with AS5 and gained like 15C. Before it would not even pass OCCT at 1.1V. I suspect fan speed is the problem. Even at full load and temps approaching 90C fan is spinning at low rpm. Even though I set performance mode (maximum cooling efficiency in bios).

I am looking for a software to change RPM, dunno I got this laptop a few days ago only :)

[veseliburek]

I overclocked my sister's Sempron LE-1100 on a J&W RS-780UVD 128M motherboard from 1.9 to 2.1ghz and undervolted it from 1.35V to 1.1V, will dig out the screenshot soon...

[jcool]

Yea, it's Toshiba P100. I replaced stock thermal paste with AS5 and gained like 15C. Before it would not even pass OCCT at 1.1V. I suspect fan speed is the problem. Even at full load and temps approaching 90C fan is spinning at low rpm. Even though I set performance mode (maximum cooling efficiency in bios).

I am looking for a software to change RPM, dunno I got this laptop a few days ago only :)

Hm.. no idea about Toshibas, I use Thinkpad Fan Control for my X60... below 60C the fan is set to off and at 70 it's ramping up to almost full spead already, but it almost never reaches past 65C (fan level 2) even when running prime - thanks to 0,975V load Vcore :up:

.. and the huge-ass copper heatpipe cooler that makes up like 1/3rd of its weight :ROTF:

[WeeMaan]

Honestly you won't gain that much... Its the god awful useless piece of crap built on 110nm chipset in it thats the true powerhog that consumes x4-x5 as much as the Intel Atom. Is it possible to undervolt the 945 chipset?

True, but it would help the noiselevel.
On the Asus 1005HA the cpu fan can get quite noisy even at light load.
Or perhaps it's mainly because of the chipset the fan speeds up?

Anyway, overclocking the Atom without increasing the voltage would be nice, or perhaps even overclock and undervolt it. :up:

[Aimr]

I've just Undervolted my new Notebook with Intel T4300

It runs on 0.95vCore now, sadly i can't get any lower than this...
But it's quite cooler than on default vCore =D

[ziddey]

you can go lower but it'd require vid pinmodding

as for atom, anyone try bsel mods for possible 166/200fsb? or was the 945 the first chipset to have multiplier lock with tampered bsel

[Evilsizer]

as for atom, anyone try bsel mods for possible 166/200fsb? or was the 945 the first chipset to have multiplier lock with tampered bsel

atoms are pinless so not possible. otherwise the fsb we can use are limited to the chipset supports, i think it is 266mhz max.

[Frank M]

Good news everyone!


I just came upon a small hack for RMClock to allow lower VIDs:
http://forum.notebookreview.com/showpost.php?p=5312927&postcount=87


Unlock hidden VIDs:

-> Start –> run -> regedit
-> HKEY_CURRENT_USER -> Software -> RightMark -> RMClock

Create new DWORD , rename it to UnlockVid, value 1.


I got an extra VID, now 0.9375V is the lowest instead of 0.95V.
I also tried a DWORD called UnlockFID, but that didn't have any effect.
Might be useful for some of you :)

[ziddey]

showing the extra vid and actually being able to use the extra vid are different

edit: i seem to remember trying these and seeing that they have no effect. but I'll try again. i could have tried something different.

[Frank M]

Dunno, you can never trust sw measurements, but RMClock let me choose
and use that lower VID on my laptop. The fan spins a bit less, but that
could also be because of the colder weather.

[WhiteFireDragon]

that's good news and all, but i figured that out like 3 weeks ago, but it has no effect. true you can't trust software reading, but i verified with measuring wattage from the wall and indeed, it had no effect. the lower VID will only be shown, but it's not actually changed

[Frank M]

Bummer :-/