The "Undervolting" Guide

How to know to which Vcc to connect?
What's the ideal/safe Vcc voltage?

 

Dono....Just lucky....For me only the 2 are accessible. Dono if the pin configuration is the same for T2500.

 

What I'm trying to find is do all Vcc have the same voltage? is the Vcc output of AC7 and E7 for instance the same?

For the layout I don't know if the T2500 and the X9000 are the same. And it does't really matter since I'm trying to understand the concept. Once I'm clear about it, and I'm sure I won't fry my mobo. I'll just dig the datasheet for my CPU and do the pinmod/undervolt.

I have a spare celeron M420 that I could use to test the underolt via pin mod, but I'm too scared to fry my motherboard... I don't hae another laptop so, I'll have to wait.

 

hi all! i have undervolted my T7250 XPS M1530 from a default 1.750 to 1.0750 highest voltage with IDA set to 1.0875v.

then i enabled auto-adjust intermediate states VIDs to compensate. lower than 1.750 makes my laptop throttle and when i choose to turn off my computer, it goes to BSOD.

well, so far so good....... is there anything i can do to lower the voltage or merom-2m procs have this setting as one of the lowest undervolt settings?

 

there is something wrong here. The default voltage of Core Duo / Core 2 Duo laptop CPUs is about 1.2v. 1.75v is high even for an OC'ed system with watercooling

What did you use to read the default voltage?

to solve the BSOD you need to up the voltage a bit.

 

Has anyone bothered asking the programmers of RMClock to release the software as open source if they do not plan to make any more updates to it?

After checking the forums for Notebook Hardware Control, I have little faith it will ever be updated as someone earlier suggested. It's been a year and a half since the software was updated and the author refuses to release it as open source.

It seems the issue is the same for RMClock.

 

You're wrong. Manni the guy who makes NHC is working in a new 32bit version. He redesigned the entire software in such a way that it won't need update to support new platforms.

The release date for new NHC is any day soon. Check this out:
http://forum.notebookreview.com/showthread.php?t=297590

 

For RMclock I don't have any news and you're right it doesn't look like they are developping a new version.

 

I will try contact Manni and see if i can beta test the new NHC

 

His own forum moderator was betting against a new version ever materializing in one of the threads in the English forums. The end of September is near so I guess we'll see in a few days. Hope it does materialize.

 

what's the lowest your temps go while ur on battery, surfing net / ms office etc...

i'm hovering at 25*c

 

Are you sure you're reading the correct sensors (case, maybe) 25C is 77F, a warmish room temperature. Try other utilities like Everest, Speedfan etc. to
make sure you're reading core0 and/or core1 sensors.

 

I usually run at about 30*C during light use according to HWmonitor.

 

Wow. Just wow. Using this guide dropped my system temps (including my GPU) about 14 deg C. That's more than I would haved dropped them if I had bought a notebook cooler.

On my P-6860fx with a T5550, I have to put my hand over the area that houses the CPU to type. Previously it got uncomfortably hot during gaming sessions, but now it's barely warm!

Thank you flipper!

 

If I tick the Auto Adjust Intermediate States VID's box, should I actually see different voltages in the boxes, or does it just calculate it on the fly?

 

Good question...I am curious also..Didn't seem to change for me. I am only using 6 and 12 though. Idle and fast only...I did away with the intermediate. Someone stated that they went with just 12 but I noticed my idle temps went from 29 to 35 so I switched back to both....

 

I'm so confused...done this many times, but new computer with a T5750...
6x-10x .95v
11x .975v
12x 1.075v ??

I can't figure out why the 12x will not go lower...and why Intel would lock their processors. My AMD runs .75v to .95v...=/

 

DV9700T Penryn 9300 chip.

Set CpuTempOFfset in Reg Tweaks to 05,05 That is the proper offset for the Penryn chips. RMClock reads 5c below what it should read. I tested with HWMonitor and a couple others, all read 5c higher than RMClock.

Turn off IDA in Advanced CPU Settings tab, turn it on in your performance on demand profile. This will supposedly give you the 12.5 multiplier and get you running at full clock speeds.

Under load I'm seeing 45-47c now on both cores. I used to see 59-60c before undervolting. Additional 30-35 minutes of battery life as well. I've always idled at 24-25c in a 71 degree room.

 

Stupid question, but gona throw it out there anyways. I undervolted, and def noticed -10 C or so at full load.
I was just wondering, that for the part
Go to the Main Profile page > On the drop down box for startup choose "Performance on Demand" > Hit Apply I did this only for AC and left battery as no management, so this means that undervolting will work for AC and not battery correct?

 

You are correct about the Penryn being 5c low with RMC.

I have never used the Reg Tweaks. After changing the value in CpuTempOFfset do I just apply the entire reg tweak or maybe I should delete the other stuff to only get the offset I'm guessing?

Well hot dog, my RMC temps are right on now. Thanks.

 

No problems, glad to hear of the outcome.

Tick the Auto adjust box first, then change the highest multiplier voltage. Watch the middle voltages change accordingly too.

They locked it after the Pentium M chips to ensure idle stability and stop tweaks being done to it. My pentium M runs at .700v-.988v, nearly -.300v off stock voltage.

Yes that is correct.

 

Awesome thanks, I just left it that way to be safe till I have done couple weeks of loading and what not.

 

You can just double click it after making the change. All of the other settings are set to defaults and won't affect RMClock's settings.

 

LOL thanks, I did the delete trick but good to know.

 

Your setting is about the same as my T2500. All my multipliers are at 0.95v except 11x and 12x. I think that Core2Duo can go easily lower that 0.8v but they have been locked somehow by intel. This way all your multipliers between 6x and 10x have the same voltage.

Some of members at the forum using the Core2Duo P series such as P8300 and have all their multipliers set to 0.95v.

 

Odd. I have a T5750 and all the voltages are unlocked down to 0.950, incl.
11X and 12X. Aside from the Merom-2M ID, mine has Revision M0 and a Signature of 6FD. Does that ring a bell anywhere?

 

Hello everybody,

I have a Dell XPS M1730 and I do not know whether, in order to overclock my X9000 to 3.0 or to 3.2 GHz I shall do it via bios and then undervolt it via RM according to flipfire guide, or everything will be directly done via RM. May you please advise?
Thank you very much

 

OC with UV.....

 

Well I have managed to undervolt all my multipliers and am pleased with the result, I still have a couple of questions however.

1. Does it matter if I am not running "RMClock Power Management" profile in Windows Vista? It should still undervolt the processor regardless of what profile is running, as long as RMClock is running?

2. Is there anyway I can make UAC recognise RMClock as a trusted program so I don't have to allow it everytime my computer starts up.

Thanks & Regards

Stewart

 

Not all CPUs undervolt and overclock equaly even those belonging to the same revision. If the were equal Intel would have underclock them by itself

 

Not sure about Vista but in WinXP when I run RmClock, the power profil selected by the OS in the "Power options" is "RMClock Power Management".
Though it doesn't matter what profile is selected as RMclock will override it.

 

do you need to try and undervolt any other multiplier but your max one?
my max one is 12.0x and i just started undervolting that one, do i need to do the same for 11.0x and lower?

 

yes, you need to undervolt each of the multipliers. You should unclick the "auto adjust intermediate VID" in the profile tab.
You should undervolt each multplier and check for stability.
If your peak temperatures are all you are worried about, than only undervolt say 10-12 of the multipliers.

K-TRON

 

I dont know much about undervolating, but when you run a stress test prime95 ect, will you not only be testing the highest multiplier and it`s voltage, what about the rest!

 

In the active RMClock profile, uncheck the highest multiplier (PST), to stress the next lower multiplier, and so on.

 

You'll find a lot of different opinions here about your particular question. After I did all the experimenting on my T5750 as suggested by many, I took the simple route. I lowered 12X (2GHz) to 0.9625 volts and set RMC to run at Max all the time. AC and Battery. I bought this notebook because (party) it's faster and I want the Core2-Duo to run at top speed to finish my applications more quickly without overheating. Since, at no load, the CPU consumes about the same watts at ANY speed, not runing applications is the best power saving scheme. I had to get over the mental hurdle that 6 speeds was beter than 1 and that reduced speeds saved any energy at no-load conditions. Turn off the screen and disk to save batery.

 

how long should i test in order to consider it stable?
so far ive come down to 1.0750V on 12.0x multi with 1 hr 45 min of ortho running and still going.

 

oh, i see, i tried putting my 12.0x multiplier earlier on .950V and hit apply and it bsod'd right away lol

 

.9625....Wow that's great...I have .9875 and only run the 12X....I bsod at .9625 so ran it up 2 notches....

When running the 6 and 12, I could be imagining things but it seemed to have almost a barely perceptible delay going from 6 to 12 thus I only use the 12X max all the time and I am happy with the performance, battery and temps..

 

shin, stick with the 1.075V on your 12 x rail. It is running stable, if orthos has run for over 1hour
You should work onthe lowest multiplier now, and than you can kindof fill in the rest later with voltages between the lowest voltage and the highest one. It helps to do the extreme's first, so you know the range the other multipliers must be in, which speeds up the process of becoming stable at a lower voltage

K-TRON

 

Better sell it......

 

just did some undervolting, didnt stress test it for 3 hours but did it long enough to se where the temps max out
cpu t5750

before undervolting
idle 34 C
stress 1.25v 83 C

after undervolting
idle 31 C
stress 1.00v 59 C

pretty big improvement i'd say
thanks for the guide

 

What's your notebook make and model. I got similar results with my T5750 in a Fujitsu A6120. I'm collecting info for NB Hardware Control Program so I can run the fans too.

 

gateway m-6873h. i think you can find the specs on the gateway website.

 

why?
im getting 60C @ 1.075 stress tested for 4+ hours...

 

Another success story. Top multiplier went from 1.225V and 86 degrees at full load to 1.025V and 64, 4+ hours ORTHOS Stable plus 30 min 3DMark06 plus 2 hours TF2.
My system doesn't get thermal downclocking any more! I can play games and encode video without putting my room fan on my laptop!!

 

Those temps are great, its fine. Not all processors will reach the lowest voltage possible.

Great to hear. Though you might wanna look into cleaning your fans or change the thermal paste.

 

This was just AFTER cleaning my fans. I'll probably update the thermal paste if/when I upgrade my CPU to a Penryn.

 

Hi i wanted to know if OS has anything to do with voltage settings as i was using xp till now and lowest i could get ideal temp was 43-44 C and in vista with same setting it goes below 34 C

Also is it better to use Vista power management or RM clocks and if yes to what settings
Please suggest
Thanks

 

Ok. I undervolted my XPS m1330 (T7250) using this guide. Changed voltige for the highest multiplier from 1.175v to 1.025v, also changed the other multipliers and gained sufficient temperature loss.

I noticed in RMC, that occasionaly the voltage reaches to 1.175v, which was my default highest voltage. This peak occours for 6x multiplier as well as for 10x. Is this normal or not? Under stress test, it runs correctly ([email protected]).

Also noticed that when I enabled Thermal monitor 1 and Thermal monitor 2 (by default, they both were disabled), CPU used only 6.0x multiplier but didn't get any 1.175v peaks).
When I enabled only thermal monitor 2, the same results occured. But when I enabled only thermal monitor 1, CPU was able to use all multipliers but tended to use more 10x (more on full power), but no 1.175v peaks. Also the throttle line in the monitor graph was much lower than the core clock line (whatever this indicates).

So, shoud I be considered about these 1.175v peaks and what are the benefits of these thermal monitor options.