Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Artagra

macrumors regular
Original poster
Sep 6, 2007
122
42
So I've been a big fan of undervolting on PC notebooks for a few years, so when I got my mac I immediately hunted down and bought Coolbook ( http://www.coolbook.se/CoolBook.html ). I've spent a couple of hours tonight playing with it, and it seems I can run with stability at 1837MHz / 0.95v and 2004 MHz / 0.975v (stock 2004 / 1.2125 and 1837 / 1.175).

Now, such a big drop should give an *impressive* battery increase (will test tomorrow), and has given a definite decrease in temperature.

Has anyone else played around with undervolting?
 
Why are you trolling him? If he wants to do it, it's good that he posted about it. There may be other people interested.
 
What a bunch of *ssholes. I use CoolBook for cooling my MacBook down. You don't necessarily have to slow down the CPU, just drop the voltage a little. It actually puts less stress on the computer because its running cooler at the same speed. An added benefit is that the fans work less.

I have found that 1.100V works for me a 2GHz and it takes my temps down be a few degrees and the fan speed goes down as well as a result (by about 700RPM).
 
What a bunch of *ssholes. I use CoolBook for cooling my MacBook down. You don't necessarily have to slow down the CPU, just drop the voltage a little. It actually puts less stress on the computer because its running cooler at the same speed. An added benefit is that the fans work less.

I have found that 1.100V works for me a 2GHz and it takes my temps down be a few degrees and the fan speed goes down as well as a result (by about 700RPM).

By the same token, can this overclock?
 
It says that its only partially workable on SR MBPs. Anyone have any experience with this? Sounds just like what I would love for when Im at school. I have no need for CPU just battery life.
 
I use RMclock on my windows machine to get the same result. Primarily I use it to lower the power consumption of the CPU and extend battery life, but I keeps the laptop cooler too.

Sadly it looks like coolbook is a no-go on my SR MBP.
 
Lol... that's hilarious. I have to wonder if the first three responders could read :p

1. It doesn't affect performance at all. At 2.0 ghz the machine still runs at 2.0ghz.

2. As mentioned by others, it places less stress on your system. Your whole computer will run cooler, lasting longer and being more reliable. (Caveat - obviously, if you undervolt it too much you will experience instability).

3. The improvements in battery life and operating temperature are pretty sizable.


To the guys asking about Parallels / RM Clock - I tried in parallels on my macbook and I didn't have much success, but YMMV. Apparently the author is writing a newer version that should work on the SR Macbook Pro's - may be worth waiting for that. I did get RM Clock working in Bootcamp (it's an awesome app btw, if you can get past the wierd user interface :D), and I've read of some people coming right with RM Clock using VMWare.

Battery life has gone up by about 25% when running with Wifi on and mid brightness (about 5 hours, compared with just over 4 before the changes). I use this in conjunction with SMC fan controller to have a long running, quiet, cool mac - love these apps :D
 
I don't get it ...

if there are only advantages to undervolting, why doesn't Apple build this right into the system? Surely they would want the extra praise their machines would get by the press & users, if they ran so much cooler and longer ...

Something has got to be wrong. Any EE major on this forum who can comment?
 
if there are only advantages to undervolting, why doesn't Apple build this right into the system?

I know on the G4s at least, "Better Battery Life" vs. "Better Performance" would vary the voltage slightly in order to change the computer's performance and battery life. I'm assuming the same thing happens on Intel Macs.
 
I have wanted to try it out for a long time but the $10 gets to me every time.

I mean, this is free on the Windows side... :(
 
if there are only advantages to undervolting, why doesn't Apple build this right into the system? Surely they would want the extra praise their machines would get by the press & users, if they ran so much cooler and longer ...

Something has got to be wrong. Any EE major on this forum who can comment?

I agree, when first reading about this it seems like a "that can't be possible - if it is, the manufacturer would do it" kind of situation, almost like chipping your cars ECU to make it more fuel efficient AND more powerful! (Note: I don't think ECU chipping and undervolting are directly or even remotely comparable :p) However, when you understand the logic behind it you will realise that undervolting can and does give between moderate and amazing gains if you have the patience and knowledge needed.

Apple doesn't do this in the factory for the same reason why every Intel CPU isn't "overclocked" from the factory. Basically, not all CPUs are created equal. Some are better, some are worse. Once Intel has manufactured a CPU, they test it to see how well the CPU performs (what clock speeds it can hit at what voltage and what temperature) and then set default voltages, multipliers and FSB speeds based on that. Therefore, if two CPUs come off the production line one after the other having gone through the same process, one may end up as a E6750 while the next may be an E6550.

Now, obviously, Intel has split up their product line into a finite number of CPUs (E6550, E6750, E6850, following from my earlier example). All have the same core, same design, etc - just different clock speeds set based on the individual CPUs performance. Now, a couple of things lead to the ability to overclock / undervolt:

1. The CPU may fall somewhere in between - IE, it's just not good enough to be an E6750, so they mark it an E6550. Someone who buys this CPU can then comfortably overclock it beyond the E6550 rating. The same applies with overvolting. You may find that a 1.83ghz 'normal' Core 2 Duo was just too power hungry to be a 1.6 Ghz LV Core 2 Duo, so it was marked a normal CPU - but that there's still a lot of headroom for dropping the volts.

2. Intel's fabrication plant gets so good and reliable that they are consistently producing CPUs in the top bracket. Now, they don't only want to sell E6850's - hence some of the CPUs, even though they are more capable, are marked and sold as lower speed CPUs depending on demand, their marketing strategy, etc. That seems to be the case at the moment - just about every E6550 or E6750 OC I've read about hit 3GHz +. A similar thing happens with the mobile CPUs - Intel wants to be able to sell more expensive LV and ULV chips to their customers, so they limit the minimum lowest voltage that the CPU runs at by default to a higher level than it could safely run at.

Answering your original question, Apple doesn't do it themselves because of the cost and difficulty that would be involved. Not all CPUs at the same rating will have the same ability to undervolt - so it's very hard (if not impossible) for Apple to go through and test every CPU and bin the chips themselves.

(Some boutique PC builders have done this, on the overclocking front. They buy 50 high end chips, test each to see the overclock, and then sell the best performers for more in highly overclocked rigs).

However, just because it's too hard for Apple to do when they are millions of laptops a year, it's not too hard for me to do! Your results will of course vary (for all the reasons listed above) but there's no harm in testing out at what volts your CPU will run and deciding if the benefit is worth it for you.
 
I have wanted to try it out for a long time but the $10 gets to me every time.

I mean, this is free on the Windows side... :(

Personally I wouldn't have tried it if it cost me $10 the first time - I had previous (free) experience from RM Clock on a windows machine that made me realise it was worth it.

I recommend you play with RM Clock in Bootcamp for a bit to see how well it works and what kind of gains you can expect - then you can decide if it's worth it.

For me, $10 to have a 20 to 25% increase in battery life, between 5 and 10 degree celcius lower temperatures at all CPU speeds, and a slower running fan at some CPU speeds was definitely worth it, no doubt about it.

Now, from my experience, this is an unusually good result - I don't think everyone will be able to get their normal "minimum" voltage (ie, my voltage at 1ghz, 0.95v) stable at one stepping below max (1.86ghz), and to get their max speed stable at only 2 voltage steppings above minimum (9.75v). But I think that it will be worth it for most users, even if you are pretty unlucky. Intel's current fabrication process is so good and so fine tuned, and the Core series designs so robust and effective, that just about *every* chip has some headroom.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.