Take a look at this power management page where you can find several white papers on this topic and even find out how exactly you can save in energy costs using Power Management ROI calculator.

Underclocking is basically just the opposite of overclocking – instead of looking to increase the CPU’s frequency, you look to lower it. Undervolting involves lowering the voltage used by the CPU to just above the threshold where it would prove unstable – this option is probably more ‘for the experts’.

As far as using these techniques, my BIOS does not allow me to attempt undervolting. I was, however, able to attempt some underclocking in the form of limiting my CPU cores to lower frequencies.

Basically I have a dual core 1.5Ghz CPU. Essentially, they idle at 1Ghz, and by default will ramp up to 1.5 when necessitated by demand. Using software, you can change the policy by which it decides if and when to ‘ramp up’. I’m a linux user and so I can use cpufreq settings to select ‘powersave’, which is a simple way for me to accomplish this. The powersave option keeps my cores at 1Ghz always. I don’t mind the theoretical 30% loss of speed, because of two things: a) it’s not really a 30% loss, because the CPUs would typically only jump up to 1.5Ghz for a second or two at a time, under my normal computer usage, so it’s probably only a 5% (if that) loss of speed for me (typically), and b) keeping the core at the lower frequency also keeps them running slightly cooler, which means my laptop’s fan doesn’t need to come on as often – which I like.

Thanks for the link. I am astounded at the 1 million hours of life you quote. Back when, when I was looking at such figures I remember seeing figures more like 170,000 hours. And based on what S.M.A.R.T. reports on a SATA drive on a system I’ve been using a little over a year, I would say the manufacturer is projecting something more like 80,000 hours. Nonetheless, again based on what S.M.A.R.T. reports I’ve come to a conclusion similar to yours. Having had the computer on an average of about 10 hours a day and on average power cycling more than once a day, S.M.A.R.T. is estimating I’ve used up about 6% of the disk’s life from power on hours but only about 1% of its life based on power cycling. (Maybe thats a poor way to phrase it, but my interpretation is that at this rate, the disk is more likely to die from being on too long rather than from power cycles.

@Malaz

I believe the assertion is based on the wear and tear of powering down the drive for suspend. As I’ve indicated, I remain skeptical of the claim. But my mind is open if presented with data indicating otherwise.

I don’t see how that would work. Suspend does not perform any access to disk, as current states are backed up into RAM. Perhaps using hibernate several times a day would be hard on the HDD, since a large amount of write/read goes on…

Can you provide a link for that assertion please? I have read that sort of assertion before but have never found anything to back it up.