Simple Undervolting Tutorial - Pentium M Dothan - Page 52

The biggest benefit that one can get from undervolting is a cooler laptop. Overall, it probably gives about 15 min or so extra battery life too. If done properly, it in no way affects CPU performance (unless your CPU gets WAAAAYY too hot), and will make more efficient use of the laptop's battery. I think it's worth it as its free and doesnt require much time or effort

I tried to download the Prime95. It always stuck, 3%, 18%, 56%, 90% etc. Never was able to get a full download. Can anyone send the file to me, thanks.

OK, I finally get it downloaded at early hours of the day.

If all you get is 15 minutes of extra battery life, then its not worth the hassle IMHO.

One question: is it safe to assume that if I have 6x voltage working fine and a 17x working fine, too, I can create more P-States and the voltages will be ok?

In other words: is it fine to assume that the medium values between 6x and 17x will be safe, if 6x and 17x are already OK?

The intermediate values will run at stock voltage, so they will be more than ok

i think you didn't understand...

If i use only maximal and minimal P-States, the processor will run only at either 600Mhz or at 1700Mhz. However, I've put intermediate P-States (8x,10x,12x,14x) and they do not allow me to change the voltages. I was curious if it's possible to have minimal and maximal voltages OK with Prime95 and then, with the intermediate voltages generated by RMClock, errors in Prime95.

--> Is it safe to assume that if maximal and minimal P-States run OK with no errors whatsoever, the medium values will, too? Intermediate states generated by RMClock (also undervolted), not default voltages.

thanks for a great thread,my pentium m 1.86 now runs at idle at 31 degrees,flat out on full load at 52 degrees and is now running on .700x6 and .988x14
cooler and quieter than ever before and runs an extra half hour as well

gateway 8510gz 1.73
now at :
6x = 0.7v
13x = 0.972
thank you so much for this tutorial made a big difference in the temps.

What does "Run HLT command when OS is idle" do? I checked it...for no particular reason other than that guide said to.

I have a dell i9300, specs and stuff in signature.
I have a couple of questions about the tutorial
I stepped down to 1.084V 15x for maximal, next step crashes firefox and prime95 gives an error.
(i just got another firefox crash with even 1.084v but pirme95 is still going, dunno whats up with that, does anyone else have same the problem ?)
Minimal is .700v 6x
My question is what is the advantage of lowering the FID ? i am already seeing awesome changes in cooling, processor is not going over 44 celsius with Half Life 2 . what will lowering the FID do ?
Also what does the performance level in the management thing control (AC MODE and BATTERY MODE, it has options such as performancec level 1,2,3...., powersaving level 1,2,3...)?
Also is there a way to not have the program launch and still use the configuring for maximal and minimal speeds ?
thanks a bunch, great tutorial !

High for my newly bought NIB 735 Dothan doesnt go below 1.2V without errors. Everything else is fine though...including temperature.

Any guesses?

It idles at 28C without undervolting and maxes at 5xC on load without undervolting .

[ I use a laptop cooler set on high all the time ]

Thank you for this wonderful thread. I have not read the whole thread, so please accept my apologies if these questions have already been answered - I did try searching but could not find clear answers;

1. If I test an undervolt and my 9300 locks up and fails to boot up, how do I fix it ?

2. I have a stock 6800... will the standard underclocking affect it ?

3. If not, can the 6800 be underclocked ?

4. What would I test the 6800 with since I doubt that Prime95 will exactly stress the GPU ?

5. Are the instructions on Page 1 up-to-date ?

Whooaaa, sorry guys, I've been neglecting this thread . Hope I can give you some solutions:

DDDa- Using RMClock generagted VID values for intermediate states "should" be stable. The thing is, it's not worth it. The processor spends maybe 0.01% of it's life at intermediate states. However, I'll apply them in my RMClock and see what happens .

mark123can and canusa- More successful undervolts! Glad it's working well for you guys, those are some great numbers

nephilim- For some reason, HLT commands replaces the "System Idle Process" as a process that occupies idle CPU time. The Windows System Idle Process works fine, so you dont need run HLT commands, I'd uncheck it.

Caboose101- An immediate crash in prime95 means that your CPU is VERY unstable. I think an everyday value of 1.116v should be great for your CPU. Most CPU's can do .700v at 800MHz, so that sounds good. The FID is the multiplier on which your CPU runs. The 9300 has the Intel 915 chipset which has an external clock of 133MHz. Your CPU runs off this clock. At idle, FID 6x, the CPU runs at 133MHz x 6, or 800MHz. At full load, the CPU will run at 133MHz x 15, or 2000MHz (2.0GHz). Lowering the FID will make your CPU run slower, which is a bad thing . The performance level indicator I think governs how sticky the CPU is to a certain speed. At higher performance levels, I'd imagine it makes your CPU very likely to jump to full speed. At higher battery levels, I'd imagine it makes your CPU very sticky to 800MHz in order to save power and will only amp up to full speed when it really has to. Overall, there's not much difference between all of them since most of the time the CPU runs at very low usage (<10%) or full load (100%), so leave them on balanced. Finally, RMClock requests voltage in real-time, but I guess you could uncheck the "restore defaults upon exit" box and see if your temps stay low after you exit. Try it out!

futureaudio- Is this 735 pin-modded to 2.26GHz? Or is it running at 1.7GHz? I'm taking that NIB is new-in-box and that you put this CPU in yourself? Please give me more info and I'll see what I can do .

Siggy-

1. If your 9300 locks up, simply hold the power button until it turns off, then press the power button again to turn your computer back on. As soon as RMClock exits, your CPU goes back to default. Just make sure that while you're testing you DONT select the "run automatically at startup" box and DO select the "restore CPU defaults on exit" box until you have found stable voltages

2. Nope, CPU and the video card are totally separate.

3. Yes, the 6800 can, but it involves BIOS editing and flashing, and really doesn't give any lower temps (just a little more battery life). I can help you with that, but it gets pretty damn complicated if you're new to computers.

4. GPU can easily be tested with 3DMark05, newer games (FEAR, Splinter Cell: Chaos Theory, etc.).

5. The RMClock instructions on page 1 are for the older v.1.6 of RMClock. Here are the v.1.8 instructions: http://www.notebookforums.com/showpo...&postcount=976

Hope that helped guys, happy undervolting

I'm using Notebook Control Center instead of RMClock Utility to do my undervolting and it seems like my Pentium M 740 is a good one. I prime test 1.73GHZ (133x13) @ 1.1v for 1 hour and it seemed stable so I'm still testing right now. It's at 1.052v and still I have had any problems, I guess I got a good one . Further testing to come!

EDIT: This is my newest low, still think this chip has some more in it though..

Thank you for your quick response, but I do have one more question;

You said;

However, in the notes you say that HLT does need to be run for 9300's (which use the 915 chipset). I assume that when nephilim says that the guide told him to it was because he has a similar chipset.

How do we know whether we need that option set ? Is it purely down to chipset or does it vary by machine ?

Thank you.

I'm not a newbie, so can probably cope with undervolting the 6800, if this is possible... I'd love to keep it a little cooler and extend the battery life as much as possible... I play Civ IV and my deal with my wife is that I can play on the battery since this provides a solution to One-More-Turn syndrome. The longer the battery lasts, the better !

I was messing with my 1.73GHz PM last night as well, and I came to the same conclusions: Prime95 ran for 11 hours straight @ 1.004V, but just to be safe I bumped it up a notch to 1.020V (I think) for everyday use.

why if it ran for 8 hours or more its perfectly stable and you should have no issues, I'd run it at the lower voltage as less heat and more batt, congrats on the low volts.

Yeah, but Prime95 failed in 40 seconds when the voltage was at 0.988V (the next notch down from 1.004V).

I'm just trying to be safe.

I'm running mine at 1.1v @2.13 and it failed after about 40 seconds at 1.084v but is perfectly stable at 1.1 and have been running that way since I got it as it passed prime for 13h at 1.1v, it doesn't matter if it failed at the next notch down as long as it runs at the current voltage.

I've got a new low, it's in my signature... That's the lowest this thing can go.