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

try CrystalCPUID it can change voltage and the muti to thats how i been undervolting my dv5000z got shortcuts for low power and stock. For me low power is .800vcore 800 mhz stock is 1.375 2200mhz.

 

Hi. I'm new both in this forum and at the undervolting business.

My laptop is a HP Pavilion ze2430: Turion 1.8Ghz (Lancaster), 2x1Gb Corsair Value Select DDR400 and so on...

In my experiences, I archieved the following results:

At max speed (9x)
Original: 1.35v, 97º full load
Undervolted: 1.1v, 69º full load

At min speed (4x)
Original: 0.9v, 58º full load
Undervolted: 0.8v, 50º full load

The stability testes have been done with Prime95 in the three main stress modes for 24 hours for each voltages (yes, I'm quite neurotic when it comes to stability).

I guess this results are very nice, but I wonder why I wasn't able to get better undervolting settings for the 4x speed. Any thoughts?

Thanks.

 

To add some information:

The chip is a ML-32 Turion. I was able to put it running at 1.075v at max speed, but Prime95 returned an error after 9 hours of Blend stress test.

At min speed, 0.8v was really a cap. As soon as I change to 0.775v, the computer crashes.

 

Good result

 

the result is very good! Yet, an error after 9 hours...?? I suggest you check your memory with Memtest86+ because it's vey unusual to have Cpu error so late.

 

Vassil, my RAM passed at all Memtest86+ tests several times (I tested as soon as I get it, about two mounths ago). I don't think there's nothing wrong with it...

 

Hi guys,
I just reinstalled Xp on BenQ P41 (Turion X2 TL50 - Ati X1100). The new catalyst driver screwed up my previous install. Now with the new install, Iam using Omega latest driver (VGA indicated as X200M which is the same thing basically).
Yesterday I played around with latest version of RmClock again.
I used it in the past & was able to bring down Max VID at 8.0 to 0.900V.
At that time I was not able to undervolt VID at 4.0 (stock VID at 4.0 = 0.800V)
However, new version can do this with no problem.
Yesterday I managed to run 0.725V stable for more than 2 hours.
I even tested at 0.712 for about one hour with no problem but while doing torture test with Prime 95, My windows auto update started to install ie 7 & I got a crash.
For now, I will use 4.0 at 0.725V & 8.0 at 0.900V.
Still, it is a big improvement. Core temperature is also lower now.
So far so good.........

 

Hey hi people!

I have an NX 6115 with a Turion64 ML-32 chip. From what i know, i seem to be getting a little more than an hour worth of battery life. Looking at the max battery profile, i checked the box 4X and 1.4 volts. The thing is, the whole list of multipliers till 12X are all 1.4 volts. I have checked the forums with similar Turion chips (i.e ML-37) and they are able to undervolt to 0.8v. Am i missing something here? Thanks!

 

Even default settings will give you max 1v for 4x, so if you are set up at 1.4v, you are wasting power. Enable P-state transitions in RMCLock and set the voltage to a lower value.

 

I did enable direct P-State transitions, but it's still stuck at 1.4 volts.

 

By default RM Clock sets the max voltage for all multipliers. You should be able to change it using the drop down box as indicated by the arrow...

 

I'll try once i get my hands on the ML32. Will report the status later..thanks!

 

This was what i had, the VID select blanked out. Enabling/disabling the auto adjust intermediate VID yeilded no results.

Any advice?

Edit : After some testing i managed to get it to 0.9 volts, 4X multiplier. But this is via the CPU Default Settings in the management section. Also after charging, i have to restart the application just to get it back to 0.9volts. Can't it automatically switch the voltage too?

 

Change the AC/Profile to Power on Demand - this will allow you to change the voltage

 

For some odd reasons all multipliers are stuck at 1.4volts. Changing it to performance on demand for both profile did not work at all. The monitoring shows 8X and 1.3volts at startup.

Btw, which is more power efficient? The Direct P-State transition or the AMD P-State Transition rules?

 

My best current result:

0.775V at 4x
1.075V at 9x

Stable as a rock .

 

Forgot to mention its a Turion ML-32 .

 

I've been reading up on how to undervolt and such for awhile now..
Had been playing through RMClock for the longest time but to no avail could I get the volts to be changed.

Eventually I noticed there was a [+] beside profiles and I went into Power on Demand and changed that one manually..now my volts are lowered..

I tried for hours yesterday but everything kept staying at 9x and 1.450V on max load...Ran it a few times thinking I had changed the VID but was frustrated that the temps hovered around 70C
However thats the temp it showed in NHC

I uninstalled NHC and I kept working with RMClock
Now with the changes made directly in the profile...
I have

4X @ 0.875 V
9X @ 1.100 V

I've been running it for 1 hour straight using Prime95 and running a full screen DivX on VLC...

My CPU core temp records as 47C right now...
I don't know if this is an error, that seems to be WAY less..

From 70C to 47C, a 23C drop..that seems like its wrong to me

Running a Acer Aspire 5040
Turion ML-34 @ 1.8ghz

 

23C is a bit of too much indeed. 10C is much possible. But somehow, I think your 70C is not that accurate. Also, 47C is not likely the temp in full load, rather it is in my idle mode.

 

I don't know how accurate the 70C is, however..whenever I run full screen apps or games..my laptop becomes extremely hot. Almost shockingly hot....almost to the extent that if I touch the bottom side it will practically burn me.

I'm running Prime95 right now, it claims full load, but its going at about 55C...i've done some tweaking...reupped the VIDs a bit

When its on idle, sometimes it drops to 28-30C

Are there any progs that show a more accurate temp?

I noticed in NHC for the CPU temp, it shows about 3-4 degrees higher than what I get in RMClock

 

Use Coretemp for accurate temperature levels although RMClock also give the temp but it takes it from a different place.
What is your room temperature? Do you use a cooler? 28-30C is too good to be true but possible (My Athlon 3500+ in Zv6000 idles at 30-35 max, heavily undervolted). But hey, the lower the better for your hardware!.
What is the temperature of your hard drive?

 

That doesn't sound right to me. If you feel extremely hot, then it must be 65C above. Though it is possible due to GPU temp, but your CPU should get high enough to hear loud fan noise. What is your fan condition in Prime95? Did 55C dropped during prime95? The 28-30C is only possible if your fan keep spinning. But this is not what mobile CPU designed for.

 

Feeling my hard drive I know it gets super hot. I had NHC running earlier however its not installed anymore so I don't have a reading for when the HD is hot. It is a 4200RPM though so it does get quite hot..

My room temps are about 23C and no I'm not using a cooler.
Before any undervolting, when the laptop was hot and the fan was on (and very loud) it blew out very hot air

How do you check fan condition in Prime95?
No the 55C didn't go any lower

Is it best to have several levels of P-states checked off?
Or is it better to have it at a minimum and a maximum only?
Ex...4x and 9x compared to 4x, 5x, 6x, 7x, 8x, 9x

Just loaded Coretemp for CPU and Dtemp for HD
I'll watch them a bit and report back.

Coretemp shows 29C right now
Laptops been on for about 30 mins now..

HD at 33C, but I know it gets MUCH hotter.
Dtemp says its critical is 41C, and I know I'm going to demolish that number....

 

So I've been running Pro Evolution Soccer 6 full screen for about the last hour and a half..usually by this time the computer is pretty hot.

Its warm but not burning hot.

The RMClock cpu temp is identical to the Coretemp number, they're both at 44C right now

Dtemp is showing my HD at 53C

Which is know is a lot cooler than it was running at before.

 

Hi was wondering is it normal to get these results from undervolting
4x @ 0.800V and 9x @ 1.000v with ML-34

 

Those are very good undervolting results for the single core Turion. If you read through this thread you will see most are getting 0.9V at 4x & 1.1V at max multi. 0.8V is the lowest possible VID for the Skt 754 Turion and and 1.0V at 9x is very good as well.

 

I've gotten 0.925v off an 8x ML-30... surprising, since the MT-30 I had was only able to do 1.1v...

 

Just received my $580 brand new dv6000 with Turion x2 56 and wanted to share my undervolting result:

4x - 0.675v
9x - 0.975v

Both are stable and keep laptop cool. I don't think I've seen a better result for the 4th multiplier, so I'm quite happy with it .

 

Quite impressive considering my TL50 freezes with anything other than the stock 0.8V .

 

Yeah, just like with the single core Turions, I guess AMD experienced a steep learing curve with the manufacturing process of the X2 cpus

 

I downloaded AMD Power monitor and Dashboard, and Rightmark 2.25

The Rightmark screens are different than those shown, and I am not sure what to do. Under profiles I have selected:
AC - Power on demand, Batt - Power saving, and for start-up, do I select these same options? Do I change any of the other tabs, such as management?

Then do I start adjusting voltages since they are all 1.4V now?

Also, the AMD shows temp at 53, and the RM shows 63 degrees, which would be the one to follow?

So then I would change the 4x setting, then run rmclock and see what happens? How do I know rmclcok is running at 4x?

A link to some sort of guide would be helpful, as the link in the first post just goes to some page now.
I have an HPl2000 1.8GHz processor, and i am just trying to find out where to start.

Thanks!