How much difference does 1.35V DDR3L make?

I see a lot of marketing touting the power savings of low voltage DDR3L (1.35V vs 1.5V). Does any have experience that can confirm this? Do people feel like using DDR3L is actually extending your battery life?

Theoretically, I'd say the voltage drop corresponds to a %20 reduction in power. I don't know exactly how much power a dram chip draws, but I'd say two dimms would draw maybe as low as 5W and as high as 10W. So DDR3L could theoretically save 1-2W of power.

1-2W of power is a decent amount at idle, but not so much under heavy usage. If the savings only apply during peak usage then it doesn't seem all that noticeable.

 

I don't know about the benefit for desktops, but mobile haswell chipset REQUIRES 1.35v RAM and may work badly if one uses RAM that is 1.5v etc. So it's not really a point to fight over.

As to your question about power, I can't really say. Just wanted to point out that for many of us with systems bought in the last year and a half, it's not a choice.

 

D2 Ultima has said my main points however just to add, battery life wise, I don't buy the power savings... Ok let's do a simple calculation: Power = Current x Voltage.. You have a saving of 0.15V however the current is probably so small (miliamps) that on the whole, it makes no difference at all.. You might save juice for maybe 1 min or so.. It's literally negligible...

 

I think you are underestimating the difference. Power = V squared / R so there's a possible power reduction of 19% (although Crucial here mention 10%). The power requirement when actively reading / writing (I would assume that writing needs most power) but the power required to refresh the RAM is much lower. The graphs here are for a system with SODIMMs so the results should be fairly representative of notebooks. Visual interpretation of the graphs suggests about 0.1W per SODIMM under idle conditions but around 0.8W per module under load.

Hence, under idle conditions, saving 0.2W on, say, 5W is 4% or 14 minutes on 6 hours. The proportational difference is likely to be similar under load because CPU activity will be much increased when the RAM is being read / written.

John

 

I think you are greatly overestimating the power savings. The RAM accounts for very little of the overall consumption, and you are reducing a nearly negligible amount by only 19%. If you have two sticks of RAM in your computer, pull one of them out. You just reduced RAM power consumption by 50% and you will be hard pressed to notice any difference in battery life even with that.

 

Would the RAM power consumption necessarily drop by 50% if one module is removed? The refresh power requirement should do so, but all the memory access will be concentrated on the one remaining module.

Anyway, my extimated 2.5 minutes per hour is getting into the range of difficult to measure. Turning down the display brightness gives a much more measurable impact.

John

 

I've also doubted your estimates.. Compared to the amount of power used by major components like the display, RAM hardly uses a fraction... I would say again, the effect will be a matter of 1-2 minutes... The power used by RAM is really small in the overall picture..

 

I think John's estimates are sound. They match mine based on dram consumption numbers that I can find on the internet, but may be an upper limit. 2 minutes per hour isn't negligible, but one would be pretty hard to pressed to notice.