Undervolting the AMD Turion 64 ML-34 processor (L2000)

Great results on the undervolting.
Any idea how this core temp works? Why doesn't Mobile Meter work, different method of obtaining the temp?

 

Thanks for the lead to Core temp. However the download site was probably down. Anyway, googling it I found alternative download site.
Core Temp does indicate my cpu temp.
On idle with all background programs working + interneting, max temp reading was 45C.
On max load with default voltage, max temp = 70C, max load at 0.900V max temp= 63C.
Unfortunately, the fan is spinning all the time evem when temp was below 40C (very slow rpm). I think a BIOS upgrade will address this issue.


According to Prime 95, it is better for dual core to stress one core only which means I have to make 2 test for my specific cpu.(prime95 windows - advanced - affinity)

While retesting cpu 1, I found something strange about my cpuid setting. I already set cpuid to run on every startup. I point my cursor to cpuid icon on system tray, pop up said that multiplier= 8x, voltage=0.900v. But amd dashboard indicated that my voltage= 1.125v (pre undervolting setting)
I openned cpuid main windows, click function and AMD K6/K7/K8 multiplier, it also said that my current voltage = 1.125v
Openning multiplier setting + clicking applying button will solve this (amd dashboard showed 0.9v) but restarting windows, rerun prime 95, reopen dashboard, I am back at 1.125v again.

Is there anything that I missed in saving my new multiplier setting??

 

I just purchased an L2000 Special Edition and would like to know the steps to undervolt this laptop. I would like to reduce the heat primarily.

Thank you.

 

Look the beginning of the thread.

Lonthong, look for an option in CPUID that says "remember settings" or something of the sort. I use RMClock and it has these two settings: 1. Start at start-up 2. Remember settings. I think CPUID should have them too

 

Looked everywhere but could not find 'REMEBER SETTINGS" or alike in CPUID.
I also discovered that at multiplier = 8.0 voltqage = 1.075V ( not 1.125v as previously informed).
Probably my cpu or Turion duo in general is locked somehow by AMD. This probably explain why RMClock has settings for 1.125, 1.100, 1.075v only.

 

Yesterday I loaded prime95 on one core and hot cpu tester on the other.
AMD Dashboard did indicate that both core were fully loaded.
However few seconds later my laptop shut down by itself
I was on default voltage (1.075v) when doing this

 

Sounds like CPUID an't doing good for new Turion X2?

 

Someone posted in RM clock forum that he was succesfully using version 2.05 to undervolt his compaq turionx2.
I am glad to confirm that statement also.
I managed to let it autorun my customized setting on every startup.
FID 4x, VID 0.800
FID 8x, VID 0.900
I passed prime95 in 2-3 hrs w/o any single issue.
temp is mostly at 60-62C. In very rare instance, it went upto 64 for a second or so and immediately dropped back again (without undervolting / at VID 1.075 I noticed it once hit 70C)

However I am still not convinced with this method of stressing only one core.
As mentioned earlier it failed badly when I loaded (at FID 8x, VID 1.075) prime95 on one core and Hot CPU tester on the other

 

If your result is stable under prime95 testing for 4 hrs, then there will not be any problem in most case. The temp response is just like Turion 64, as it rarely goes to 59~62 C and sharply dropped.
Bad thing is the new RMClock V2.1 sounds trouble some.

 

Just passed 4.5 hrs with Prime95 so I guess I will stick to this setting for now.
By the way, Power schemes showed "RMClock Power Management" after I set it to run on every startup

 

The Power schemes used to be a problem for people who couldn't get it that the Power management of Windows needs to be disabled. This is why, v2.05 introduced this new scheme.

RMClock does not detect correctly the CPU, right? Could you post a screenshot of the undervolted values with RMClock? I'm sure there will be other people who want to undervolt their X2s and this will be of great help.

 

Some screen shots of my current setting:







Edit:
As you may notice, I did not underclock my max performance profile

 

The improvement in power consumption amazes me. I have Turion ML34 with 0.925v at 4x...
The X2 model is not supposed to have a low voltage twin, is it?

 

Some more shots while running Prime 95 on performance on demand:

 

edit
double post, sorry

 

The analogy is probably 2 brains with only one mouth, thinks twice - consumes less......

What I don't understand, how come the temperature between core 1 & 2 is not the same. Is core temp really that accurate or just rough figure based on workload????

You probably can underclock yr lowest setting.
Mine is locked at 4x = 0.800v

 

Is there anyway to underclock an integrated GPU?
In my case is ATI X1100 which probably is identical to X200

 

Check this out

It is used for overclock ATI, you can work on underclock.

 

hi all,
I recently bought a compaq presario v2000z with a Turion ML-37. I like the laptop except the high heat it generates. Before I bought this laptop I had the impression that the problem is not so severe. After using it for an hour, the left side of my keypad and touchpad are getting warm in addition to the left palm rest area. I am not sure if this is normal or just my laptop is generating more than average heat.

In any case I am considering different options:
1)call HP and ask them for help.
2)undervolting
3)replace the hard drive with a faster and cooler one.
4) return the laptop and buy a thinkpad.

Any advice to solve the problem would really helpful. How effective is replacing HD for reducing the heat?

thanks
Raj.

 

Your problem is the HDD obviously. Undervolting will help with the heat generally (highly recommended!) but the heat from the hard drive is a different thing.
If you upgrade to 2gb (or even 1.5gb) RAM, turn off virutal memory as this has very noticeable effect on hdd temps. For example, I recently sold my 256mb stick and left with only 512mb (for a short time I hope). The temperature of the HDD increased by from 40 to 43 avr. plus the fact that it's not that hot outside anymore.

The other alternative (the most effective) is to get a cooling pad, preferably one which blows air upwards and not one that sucks air from the laptop. This will significantly decrease the heat.
A new HDD will not solve your problem because the design of the laptop is kind of faulty as far as heat dissipation is concerned.
Buying a Thinkpad is by far the most expensive solution . You need to decide on the trade off between cost and build quailty. The Thinkpad will offer superior quailty but at the logical high price.

 

Hi,
Thanks for the information. So I will try undervolting and a new HD. HD more for the reasons of speed than heat.

The concerns I have about cooling pad are that it can add upto 0.5lbs weight and sap some more power. I got the 12 cell battery. So weight is already a bit more than 6lbs.

I realize Think pad is more expensive for comparable performance. I got this for $676. (ML-37, 256MB, 40GB, 12-cell, no dvd burner). But I will upgrade to 1GB, 5400 RPM 40GB. I think this will add up to $900. So if I can't get the heat under control and if I can get a comparable thinkpad for 1000-1100, I will consider switching. Any suggestions?
thanks!

 

ThinkPad won't buy you the HDD heat problem. It is common for all 4200 RPM HDD. My friend has a T42, and make him sweat even the A/C blow directly from his head. That is 7200 RPM do.

Yes, ThinkPad has superior building quality, but what you get is generally looks nice, and high performance on dedicate GPU. Of course, you pay for it.

 

I have the V2000 with the Sempron 1.8Ghz 3000+ running on the ATI Radeon 200M. With the latest AMD Sempron Powernow driver, I monitor my CPU through AMD PowerMon being run at between [email protected] and [email protected]. Is there any more leeway where I can still undervolt this processor? Right now I am getting 2.5hrs with Powernow enabled.

 

If I remember correctly, the Semprons' lowest voltage is locked so you may not be able to go lower than that. Yet, you can undervolt at the greater multipliers and actually this is where most of the heat savings come from.
Try RMClock and see what you can do.

 

I have an Acer Aspire 3003WLCi with a Mobile Sempron 3000+ (1.8GHz) that I've undervolted successfully and this is what I got:

4X - .900 (default was 1.075)
8X - 1.125 (default was 1.300 I believe)
9X - 1.200 (default was 1.400)

So I'm really happy with my results. I've experienced no instability and I've also tested the lowest and highest steppings for 3 hours with Prime95 and I've gotten no errors. I am using RMClock 2.15 and it's by far the best. It makes it so much easier to lower the voltages and there is some new cool stuff like a battery metter and it displays important info about the battery.

And for those of you who want to know about how much battery life I get. Well, I just got a brand new 8-cell battery (4400 mAh) five days ago and without undervolting I would get about 2 hours and 15 minutes with the screen's brightness all the way up and with brightness fully dimmed about 2 hrs and 30 mins maximum. Now, after undervolting and setting RMClock to Power Saving I get 3 hours with screen fully bright, cooler on, and laptop's fan always on, but if I turn the brightness all the way down and turn off the cooler and the laptop fan I get 3 hrs and 30 minutes. Isn't that amazing?!?

 

Lucky you. 3 hrs is really a good score. How is the heat?

 

The heat is not bad at all. When I'm on batteries the hard drive never goes above 35C and the CPU never goes above 45C (I always have a notebook cooler on, though). When I'm on AC and not doing anything too taxing the hard drive stays pretty much at 35C and the CPU goes up to 48C when it needs to run at 9X, but as soon as it goes down to 4X the temperature drops to 39C and stays there because I have the laptop's fan always on and the cooler, too.

 

So, is it confirmed that for RMClock to work, you should set the power scheme to Always On? If I uninstall the AMD CPU driver, will you still be able to underclock/undervolt through the RMClock utility? So, is it one or the other, but not both? Thanks!

 

You have to switch to Always On with the old 1.8v of RMClock. The newer versions have their own powerscheme, which is activated automatically.
You don't need to uninstall AMD CPU driver, leave it as it is.

 

I first tried NHC in my attempt to undervolt my Mobile Sempron 1.8G 3000+. Well, it did recognized the AMD Powernow capabilities but sadly it was just using the same profile as the default AMD driver. The FID/VID checkbox is grayed out so I can't do anything by way of custom settings. I really liked the UI of the program and the fact that it can show HDD temp and somehow fiddle with the ATI Powerplay settings... right now being done by HD Tune and Powerstrip respectively.

So I uninstalled NHC and installed RMControl instead. I really have to go through each and every step from the readme file before I was comfortable with my settings. I ended up with [email protected] and [email protected] with RMControl setting the voltage automatically in between. Seems the voltage for the lowest multiplier is programmed at 0.975V so it stays locked in that setting. I used D3DRightmark to test for stability. My reason for that is I'm really not going to use the unit for really cpu/graphics intensive apps so if it could at least pass that test I'm fine with it.

I don't think I could get much better battery life from that mod though. Using the default AMD driver, I am almost at 4x most of the time where it will also be staying using RMClock. If only I can get it to go down to at least 0.900V though. Anyway, I will be testing for increased battery life if there's some improvement gained from the exercise.

 

It's too bad that you were not able to decrease the voltage a bit more at 4X. BTW, the voltage at 9X is impressive but how long did you test your computer with D3DRightmark? Are you sure is stable? Did you experience any instability (lock ups, BSOD)?

Also, do you have the latest driver for your processor? I believe the latest version is 1.3.2 and if you do have it installed on your computer this could be the reason why your lowest voltage is locked at .975. I have version. 1.1.0 and the lowest voltage I can choose from at 4X is .700v. So you could try uninstalling version1.3.2 and installing 1.1.0 and I recommend that you get RMClock 2.15 which is the latest version and certainly the best one!

 

I think the lowest voltage of the Sempron (on HP laptops at least) is locked at 0.975. It will be superb if the latest AMD drivers allow for lower voltage at 4x but it may also be a BIOS issue.

 

Yeah, the old Sempron doesn't have good undervolting score. Will see if this will be changed in new AM1 Sempron.

 

After reading gearhead's post and seeing that he was able to undervolt his Sempron processor all the way down to 1.100v (9X) I decided to keep undervolting my Sempron and this is what I got:

4X - .850v (was .900)
8X - 1.100v (was 1.125)
9X - 1.150 (was 1.200)

Default

4X - 1.075v
8X - 1.300v
9X - 1.400v

I know that at 9X 1.150v is the lowest I can go because I tried to go as low as 1.125v but I got an error right after I started testing with Prime95. I might also try to undervolt a bit more at 4X and 8X but I will do it during the weekend because I need to test the new voltages for at least 6 hours and I don't have the time right now.

 

One other quick test is to run PIfast with super large digit. It really runs large amount of memory I/O

 

daniel306:

yup.... i have the latest amd driver, the v1.3.2. maybe that's why i can't go any lower. i will try to look for the old driver though and see if it will resolve the issue. i tested with rightmark for only like three consecutive runs i think. then i also tested with 3dmark2001 and aquamark v3. my reason for the short test being that i won't really be subjecting the unit to any prolonged intensive apps so if it can survive several runs of a benchmark then its good to go. the test is only to see if it won't give me errors for the short peaks when i will be using it in the real world. using that criteria, it really hasn't given me any bsod's so far.... unlike when i set the CAS Latency to 2.5 using A64Tweaker where it immediately gave me data and display corruption i have to force it to shutdown quickly.

 

HP L2105 (ML-32)

4x 0.775V
6.5x 0.900V
9x 1.000V

Each level has been tested Prime95 stable for at least 2 hours. 9x has been tested at least 4 hours.

I don't believe it myself... but I can't argue with it.

 

I have undervolted my L2000 ML-32 but every time i reboot the voltage is reseted to original.

 

These figures are simply amazing. could you upload a screen shot? Have you made any other adjustments except for the voltage?

 

here ya go... no other adjustments other than voltage.

 

What is your notebook and what driver do you use ? HP, Acer? These voltages should really prolong your battery life, don't they? Great, great undervolting

 

I guess AMD's 90nm process is getting very good yields since we are seeing lower voltages than when it was first released. I guess 1.0V is not bad considering they are selling TX2's stock at 1.075V.

 

Do you mean that there are technology innovations or improvements in the latest 90nm ML Turions? I was about to say it's pure luck but maybe it's more of an improvement.

BTW, any news on the 65nm Turions?

 

Ofcourse undervolting is still based on luck but if a large percent of Turions being manufactured undervolt better then it can be attributed to improvements in the manufacturing process. I heard they had developed a new process to manufacture the TX2's..maybe they have incorporated some of those into manufacturing the Turions.

As for the 65nm Turions, they are rumored to be released in a month or two, but nothing official yet.

The only other new release has been the Turion MK36 - a single core Turion based on Skt S1 running @ 2.0GHz, 512K cache.

 

I have an HP L2000 (L2105NR actually, one of the LiveStrong series), AMD driver 1.3.2, using RMClock 2.15.

 

Hi,
I have fujitsu-siemens A1650G ( ML-34 1.8Ghz ) and RMClock works perfect.I also used SysTool and have changed MEMCLK from 166 to 200MHz ( but only after processor is undervolted to 1.2 V at 1.8GHz ) .Memory tests shows almost 20% more. ( I have integrated X200 video )

 

has anyone undervolted the dv6000?

 

The process is the same. HP may have locked the voltage for the lowest multiplier, though.
You may want to see Lonthong's posts on pages 25-26.

 

Hi everyone,
New to the forum, but want to share my underclocking results. I have a Acer Aspire 5002wlmi with a turion ml-30. My current settings with rmclock are 0.850v@800mhz and 1.050v@1600mhz. Have been running for a week and has been stable, without any problems.

 

Hi everyone,
I used to use RMClock to undervolt my v2000z (Turion ML-34) which is nicely done at 4x @ 0.925v and 9x @ 1.15v stable. I have upgraded the OS to Vista 32-bit Ultimate and now RMClock no longer works (well, it does run and able to setup all voltage but Windows override the voltage to stock). CPU-z shows the undervolt doesn't kick in anymore.
Any thought to get around this? Vista is nice but my palms are burning.
BTW, no problem on driver support with Vista. I'm running RTM (build 6000) version.