The "Undervolting" Guide

1 thing. hardware monitor is showing me idle temp cpu at around 45-47C while RMClock is showing me 30c. which one should i follow?

 

cpugenie is kinda free... if u know about torrenting. i have been using rmclock which is great, and im experimenting on cpugenie right now.
hardware:t9400 6x - 9.5x IDA available

comparison data:
rmclock:
use whole integer multipliers only (6-9)
IDA option not available on my processor
undervolts less than cpugenie could in 2 multipliers
more user friendly interface
better display option on system icons (temp, battery, frequency, load)
uses less memory (5xxx k)

cpugenie
uses more memory (24xxx k)
no system icon display option
its not easy to be free of cost
undervolts better in 8x (.938 to .9v) and 8.5x (unavailable to .925v)
can undervolt .5 multipliers (6-9.5)
can access and undervolt IDA (10x)
can self undervolt for you and has an included tester so it does not require second party programs like ortho

 

that's true only for version 1.0 which is full of bugs and consumes a lot of RAM.

 

Guys, why when i change voltages in cpugenie i have a loss of performance and not in RMclock ???? i put exactly the same values : 1.050V at 2 ghz.

cpugenie 3dmark score : 1599 (Grrrrrr !)
and RMclock : 1719 (same score than with no undervolting).

 

That looks very weird indeed - can you check what happens with the CPU speed during the test?

CPUgenie has couple of "profiles" for CPU management - I think "Balanced" is default, but there are high-performance profiles as well... maybe that helps.

 

in some of those modes, CPUgenie rises the clock of the CPU on 50% of load. that means that one core has to be completely loaded (100%) to triger freq change. in some conditions, that can be "too late" and there are performance loses. maybe setting CPUgenie to rise freq at 40% of CPU load can make some better results.

 

thanks a lot guys you were totally right i was using the wrong mode. Now with cpugenie i can go down until 1.00V at 2ghz while i was limited to 1.05 with RMclock.

3dmarks score : 1722 with "extreme high performance" mode

 

Well, seems like it's totally stable with 1.00V at full speed....temps are incredibly low, i didn't try to unlock voltages to go under 1,00V... Though, 1,00V isn't bad for a 35W T core. (It was 1,250V in the first place)

What are the lower voltages you've seen here on P cores ????

 

0.925V in all states for my P7350

it runs stable with 0.950V at 2.0GHz

 

not bad maybe i'll put a P9600 on my lap in a few months ;-)
Cpugenie is a great software, easy to use....Seems we have no choice but to buy it, at least concerning the v1.1.

 

I successfully underclocked my AMD Turion x2 TL-56 from 1.150 Volts to 0.900 Volts on 9.0x Multiplier (Fully Stable now, anything under will be unstable or BSOD), but my temps are still pretty high... Still 70 degrees Celsius though. I stress tested without RMClock and I still have 70 degrees. My fan turns on at 65 degrees. The temperature range on any setting is always 60-70 degress Celsius.

 

Try "custom" and set the threshold to 10 or 15%.
Works like a charm on my T8100 and only goes
"turbo" about 10% of total CPU time. I'm not a
game player so your mileage may vary, as the
commercials say.

 

Anyone tried or interested in a project to undervolt via Pinmod?
http://forum.rightmark.org/topic.cgi?id=6:1455

Smaller structures should need less voltage,
but Intel limited the minimal voltages
(maybe to keep their prices up for the marketing of ULV-products, ULV=ultra low voltage but same physical silicon)

Dothan-90nm is limited to 0,700 Volt (ok, near the physical minimum)
Meron-60nm is limited to 0,850 Volt
Penryn-45nm is even more limited, far away from the phisical possibilities.

The results at idle (no load) are

+ more power drain,
+ less battery time,
+ 100 million notebooks are producing more CO2 at idle!

Lower voltages should be physical possible.

 

Well, firstly, great guide. Stumbled across it when I was searching for laptop stands as I was looking for something to help with the fact I usualy use my laptop on my lap or bed, where the circulation is usualy hindered.

I have a Dell XPS 1530, couple months old, T9300 inside. Got it underclocked down to 0.9625V from 1.1375. Pretty good drop, It may even run fine at the lowest of 0.9500V but i havent stress tested yet (Also havent stress tested any multipliers other than higher, planning to do later but should be fine from what ive seen with others on this chip). Temperatures have dropped and has been a great help.

One problem though, if you look at the attached image in the first screen shot you can see that my voltage is jumping up and down between 0.9500 and as high as the max 1.1375V at times, even though nowhere in my settings should it ever go above 0.9625. Just wondering why as by the looks of it Im not getting the full rewards from this process.

If any more information is needed please just ask and I will respond asap.

Thanks again

EDIT: Thought I better say what I was running when these were taken just for reference. In these I am simply running Firefox with multiple tabs, and IE running quite a graphic intensive Java Game, flashget, msn and various others running in background.

 

turn off thermal monitors, turn on all the enhanced power states except the first one. most importantly.... go to windows vista and make the max and min processor speed to be the same whether it be 100% or 50%.

correct me if im wrong guys

 

what threshold are you talking about ??? P-state increase threshold ???? it's stuck to 40% as a minimum for me...

I had again a weak cpu score in 3dmarks...even with high performance settings so either this soft is buggy or there's something i misunderstood....

 

I am having issues with RM Clock 2.35 and Windows XP x64, RM Clock will not close or exit the only way is to Kill from Task Manager, any suggestions ?, thanks

 

Hi there..

really thanks for this gr8 topic

I have [email protected].. I have undervolted it ..

6X 0.9250V

7X 0.9250V

8X 0.9375V

and after 2 hours and 30 mins with stab test ..my max temp is 52 core0 50core1 ... after 72 70 before undervolting ...

And my idel temp is core0 38 Core1 23 there is 15C between the 2 cores is it normal ??or i did something wrong ??

Edit: I noticed something .. when i do the satb test i found that :

OS Load is 100%

Cpu load is 94.14% !!!

what is this ?

 

Can someone help me out? No matter the CPU type selection i changed, it cannot find my CPU. It always has unknown core for my p8700. However CPUZ1.5 can show the right infos

 

How did U get VID to 0.875 for 6 and 7 multipliers? The lowest a can choose for my P8400 in a Lenovo T400 is 0.925V

 

I have the same 94% load CPU max.
My max temp after undervolting to 8x @ 0.9875V is 46C.
When idle, i get 27C on both cores.

As a conclusion: I don't see any reason to use RMClock on AC power (loss of 0.5 multiplier), while on battery and full load I get a 27W consumption @0.9875V compared to [email protected] which is a GOOD result!

 

anyone know off-hand if the atom processors can be undervolted?

 

Nice to know that am not the only one who has the 94% load : )
But actually i am not impressed because of my 15C deference between my 2 cores ... I don't know if it right or normal thing or not ..

And the most wired thing the same issue was exsist be4 undervolting the same 15C deference ..

Is my processor dieing or what

here is a CS @ idle



So any help Guys ??

 

I wouldn't panic, though it's unpleasant to know.
I heard several such complains on CPU's with high temperature core differences, seems like it's a matter of hardware.

I am waiting for the person with the P8400 who managed to undervolt to 0.8xV his CPU instead of 0.925 which is the lowest option to choose.

 

I that means there is a hardware problem in my CPU ?

 

not necessarily a problem. It just means you didn't get lucky.

 

huh.. actually i didn't get ur point here ..

 

No, if you have huge difference between core temps, the first thing to check is if the cooling solution has been installed correctly.

Displaced heatsink can give such problems.

 

and how can i check the cooling solution ??

would u plz explain plz

 

If this is a notebook CPU, I'd advise doing that only if you have experience with opening the notebooks - depending on the model, CPU assembly might be well hidden and might require removing many components before you can reach it.

If you want to do that, usually what's done is that you clean the CPU and heatsink surface and then re-apply the thermal paste, preferably AS5.

You can also tighten the fittings that hold the heatsink and press it to CPU, etc...

 

First of all am not experienced @ all in oping the notebooks ..

Secondly .. the laptop is new i bought it for almost 3 months !!

and from fist day and am having this wired problem

 

Hi, for some reason i do not have the Advanced CPU Settings on my cpu clock utility .

any help on this ?

thanks

 

I've opened my Amilo Pi 3525 and found out that heat-pipe is connected not only to CPU, but to chipset and some electric components, too.

There was some gummy thing between the heat pipe and those components (except the CPU), so I've decided not to touch anything.

Looking forward some useable CPUgenie version. AS5 will rest on the shelf.

 

CPUGenie 1.2 alpha posted on their forums seems to fix the IDA problem and the dreadful "Hardware configuration has changed" message...

I asked CPUgenie devs about the release and got the feedback that it will be most likely out next Monday, along with further reductions in CPU and memory usage.

 

Can anyone help me? My [email protected] at default is 2.5000 volts at default, and even then when running the CPU stress test RMClock complains about a CPU overheat and HWMonitor gives me temperature readings of either high 70/low 80s or -40s. Without even installing RMClock, my CPUs max temp was at around 85 - am I just not able to undervolt my processor?

 

2.5Volts ?!?!?

There is just no way that the CPU would be running at 2.5V - it would be fried as soon as it gets powered on.

 

Wow, sorry, I haven't slept in a while I meant 1.25V

 

i cant undervolt my asus G50V's p8600...i've done everything from the first page...but it runs allways on 1.125 V

what am i doing wrong?

sry for my bad english

 

What is your maximum CPU temperature under full load, and when doing nothing?

As a suggestion to your problem, try to run the RMClock_WipeOut and then configure from the beginning.

 

max temp is 72°C
min temp is 30°C

i've uploaded some images from my rmclock

strange...cpu-z 1.5 shows me that my P8600 is running @ 1.313V

 

anything above cpu-z 1.41 shows 0.1 volt more on the CPU. Your processor is running at ~1.2V.

 

hmmm i think i found the problem...it was power 4 gear running in the background...

 

any idea about this ? i have rm cpu clock utility pro but no advanced cpu setting option

 

http://forum.notebookreview.com/showpost.php?p=4670439&postcount=2671

That is on my Lenovo T400 with P8400. CPU-Z shows incorrect values.

 

great guide mate

I stumbled across this becuase im trying to find a solution to games like
UT3, L4D, TF2 dropping from 70-100 fps down to 20-30 when in high carnage areas
ive played around with drivers for the video, and stumbled across this

Comes to my question

What are people getting stable for 10x for T7250? I'm trying 1.1 and its cruising along, how much lower have people got?

Full specs Vista 64bit, 4gb ram, dell xps1530, 256 8600m GT

 

It seems to me, or I get the same effect with my 4500MHD video in CS for example. Can someone test the video performance exactly with some benchmark software to see if there is any difference between normal and undervolted CPU? Thanx to submit the results

 

Somebody knows why RMCLock does not work properly in Windows x64 ?, I've tried to use it in Windows XP x64 and Windows 7 x64 and it will not exit properly, the only way is to KILL it, there is any workaround for this issue ?, thanks

 

You need a third party driver to get rmclock to work correctly in x64.

 

I did install the specific x64 driver (.sys file) for Vista x64, but RMCLock is still stuck, right now I am using XP x64, dod you know about another driver for x64 ?, thanks

 

Can you undervolt GPUs?