M1530 Undervolting

I just received my M1530 today and I am going to try and undervolt it tomorrow. I was simply going to follow the undervolting guide available here ( http://forum.notebookreview.com/showthread.php?t=235824)

However, I was wondering if anyone has already done this to their M1530 and if so, if they could possibly post their voltages for each multiplier so I have a ballpark idea of how low I can go.

If its relevant I got the T8300, GeForce 8600, 4 gb ram, and the 320gb 5400.

Any help would be appreciated!

 

I believe undervolting works for any computer. I got the T8300 as well, and while running Orthos my system crashed at 1.05 volts. 1.0625 volts was stable for me, and the lowest multiplier was at 0.95 volts. A good voltage to start would be around 1.1 volts.

 

It's so funny...in the desktop world we overclock and overvolt the crap out of CPUs. But when it comes to laptops, people are trying to undervolt and underclock.

Start at the stock voltages and go down in small increments until its unstable. If it BSODs then you know you went too low. Try running Prime95 after undervolting to test for stability as it will max out all cores.

 

That's what I am using and they have been tested for 3h30mins each with Orthos. But remember these voltages may not be the same for everyone.
SUPERLFM 8x - 0.9500V
Normal 6x to 10x - 0.9500V
Normal 11x - 0.9875V
Normal 12x - 1.0375V
IDA 13x - 1.0750V

 

From my experience, 10 minutes of testing is plenty. All errors that are to occur would have in the first 5 minutes or so; if you're completely error-free at the 10 minute mark, you should be fine. Remember, in the real world, rarely are you ever going keep the processor at 100% 2.4 GHz for 16 hours or whatever. Regardless, with a T9300:

SuperLFM 0.9250
6x 0.9250
7x 0.9250
8x 0.9250
9x 0.9250
10x 0.9250
11x 0.9375
12x 0.9750
IDA 1.0000

Keep in mind that mine is a bit lower than usual though.

 

I did get an error at 45 mins during one of my test. BTW how did you get 0.9250, the minimal voltage on my RMClock list is 0.9500. Does the T9300 use lower voltage that the T8300?

 

That's rather rare. And be honest, under real world usage, how often are you going to keep your CPU at 100% load for 45 straight minutes?

I'm not positive of the answer myself, since my friend's T9300 only goes down to 0.9500 as well. But I can tell you it has nothing to do with the RMClock version, that's for sure.

 

Very nice values. Are you running into any problems with SuperLFM enabled, such as random CPU spikes?

BTW, I think on most newer 1530's with T9300 the lowest you can go is 0.9500V and unfortunately unlocking lower values in RMClock has no effect.

 

Is it suppose to? I've never heard of such problems when having SuperLFM enabled. Do you have links to anyone that has?

With the 11x multiplier, 0.9250 crashed in 3 minutes whereas 0.9375 was stable. Apparently lower values do have an effect. Again, where are you getting this information?

 

Personal experience. I only really noticed it while listening to music, which started skipping periodically. Disabling SuperLFM fixed it.

I meant to say that in your case setting a multiplier to 0.925V is enabled by default, whereas the lowest voltage I can apply is 0.95V. The registry tweaks that are provided with RMclock do unlock other voltages (i.e. 0.925V etc.) in the GUI, but when applied they still go only as low as 0.95V.

 

Everytime I eject a disc, the music skips, but other than that, I don't notice any skipping. I'll keep that in mind though, I do recall vaguely that there has been two or three times where it skipped.

EDIT: What is your SuperLFM multiplier set as?

 

If it was only a few times you don't have to worry about it. In my case it was pretty systematic (every few minutes if I remember correctly). And it was set to 0.95V, same as 6x multiplier. Hmmm, maybe I should try to set it higher and see if that has any effect effect at all, although logically it shouldn't since the lowest multiplier works at this voltage just fine and I doubt there would be major benefits in terms of further temperature decrease. Will see when I get off work.

 

No I mean it's probably a mistake on my part. Setting it at the 8x multiplier means SuperLFM will pretty much never be enabled... which totally defeats the purpose.

 

Sorry I misread that. It was set to 6x.

 

Alright, it's at 6x now and I'm not experiencing any skipping. I know SuperLFM is actually in effect, since it's showing my core clock at 624 MHz.

How was was the skipping for you?

 

By skipping I mean that there was a ~1 second pause during playback.

 

Was it ~1 second of silence or ~1 second playing the same loop?

 

Silence. So I take it it works fine for you even on 6x?

 

I've never had "silence" skips. It shouldn't anyway. The CPU should move on to the next multiplier long before the initial one is overloaded with information: the beauty of SpeedStep.

 

Theoretically yes, at least that's what happens with the other multipliers. Might be just an issue with RMclock, which hopefully didn't die completely seeing as the last release was back in February.

 

SLFM 600mhz might not be enough for whatever programs your using causing it to skip or slowdown. Its supposed to throttle up, but for some reason it doesnt.

Try changing the SLFM value to 8x, this will change the idle to 800mhz. See if the issue still persists.

 

Since we're not able to lower the voltage lower than 0.925, is there any real benefit to use a 6x multiplier rather than something higher? Sure the core clock is lower, but the power consumed is the same, as well as the throttle.

 

Voltage does not directly equate to power draw. A cpu thats running on 400mhz will draw less than a CPU running 800mhz even at the same voltage.

 

So the lowest "stable" (aka non-music-skipping) multiplier is recommended?

 

It shouldnt even be skipping. I can run alot of things on SuperLFM 600mhz and my music doesnt skip at all.

 

I've played around with the settings yesterday. Setting SuperLFM to 8x as you suggested did improve it a little bit in terms of making it less frequent. Still quite noticeable though, so I just turned it off again.

 

One thing about the undervolting guide that confused me was that it was suggested to turn off LFM and IDA during the undervolting. Am I supposed to turn both of these back on? Also, what's wrong with undervolting IDA? It seems like you guys posted lowered voltages for IDA... (Sorry if it's been asked before but 100+ page thread is kind of hard to read)

 

Just thought I would share the results of my undervolting. I seemed to get mine really low compared to most:

SUPERLFM 8x 0.9250V
6X-10X 0.9375V (I stress tested at 10x for 2 hours, didn't bother with the rest)
11x 0.9500V (Stress tested for 2 hours)
12x 0.9625V (Stress tested for 3 hours)
IDA/13x 1.000V (BSOD'd at .9875 after 2:25)

All seems great so far and my laptop is only getting up to about 62C under stress at max multipliers.

Do these numbers seem low to anyone else? I only actually got BSOD'd that one time at 13x.

 

I'm curious about this as well.

I've been successfully testing the numbers from the 4th post for about 30 minutes now btw, don't think I'll go any lower.

EDIT: I am actually experiencing the sound skipping mentioned earlier

 

It turns out the sound issue was unrelated and I've since fixed it with newer sigmatel drivers.

An issue I am noticing with RMClock is that according to the tray icon occasionally it will report a VID that is higher than my 12x or 13x. It switches back to my defined values eventually, but for a few seconds it will be using what seem like the default values. Am I missing an option somewhere?

 

T8300...
SuperLFM - 0.9000v
6.0x --- 0.9000v
7.0x --- 0.9000v
8.0x --- 0.9125v
9.0x --- 0.9250v
10.0x -- 0.9375v
11.0x -- 0.9375v
12.0x -- 0.9500v

 

Hmmm... my T8300 won't even display voltage that low (and I ran the reg file)

 

jusr wondering how you can get 0.9000v. mine shows the lowest 0.9250. can u tell me how to get that low of 0.9000v? tks

 

Hi guys... i've also experienced the sound skipping issue, however, today i made a pretty big discovery...

sound skipping occurs only when using WINAMP.... if i listen to music with windows media player then there is no skipping whatsoever (using SLF and a 6x multiplier with a T7250).

Any idea why this may be happening?? I really like winamp but with the skipping issue WMP is starting to gain some pretty big points...

 

Hi guys... i've also experienced the sound skipping issue, however, today i made a pretty big discovery...

sound skipping occurs only when using WINAMP.... if i listen to music with windows media player then there is no skipping whatsoever (using SLF and a 6x multiplier with a T7250).

Any idea why this may be happening?? I really like winamp but with the skipping issue WMP is starting to gain some pretty big points...

 

even without sound skipping there are other things that will spike your cpu usage in an instant like for example sometimes firefox browser will freeze up on me when too many windows are opened up within the browser or pages I should say and then the task manager will report abnormal cpu usage on the process for firefox browser.

So with an under voltage cpu maybe this kind of scenario for me would result in a BSOD every time it happens ?? (Question ? because I have not tried this and was hoping someone else did)