Does the Amount of RAM Have an Impact on Laptop Battery Life in Sleep Mode?

[Jason Dunn]

When I wrote my <a href="http://www.digitalmediathoughts.com/articles.php?action=expand,12619">review of the Kingston 4 GB RAM upgrade</a> for my Dell XPS M1330, one of the questions I wondered about was what type of impact the amount of installed RAM would have on battery life. When a laptop goes into sleep (suspend) mode, it has to provide a small amount of power to the RAM in order to maintain what's in memory. When a laptop goes into hibernate mode it writes everything to the hard drive, and thus no power is used while it sits there unused. I tend to prefer sleep mode myself, because it's usually faster to go into sleep mode, and faster to come out of it.<br /><br />To test whether the amount of RAM has an impact on battery life while in sleep mode, I devised a simple test: I took my XPS M1330 with 4 GB of Kingston RAM, charged up the battery to 100%, then put it into sleep mode for 12 hours. I disabled the function where Vista will go into hibernate mode after "x" hours of being in sleep mode. While seemingly a good feature, on the XPS M1330 it would wake up the laptop (lights blazing) which results in the slot-load DVD drive making a "cha-chunk" sound. Having this happen in the middle of the night when the laptop is on a nightstand right next to your head is disturbing to say the least.<br /><!><br />At the end of the 12 hours, I opened the lid on the laptop, waited about 20 seconds for the system to stabilize, and took a battery reading. The results? With 4 GB of RAM installed (2 x 2 GB sticks), and with the 9-cell M1330 battery, after 12 hours the battery level had dropped from 100% down to 91%. With 1 GB of RAM installed (1 x 1 GB stick) after 12 hours the battery level had dropped to 92%. Because this test required me to not use the notebook for 12 hours, I only did each test once.<br /><br />My conclusion? The 1% variance is within the margin of error, so I believe that having 4 GB of RAM poses no significant added loss of battery life in sleep mode over having 1 GB of RAM. I might have seen a more noticeable difference if I had conducted this test over a 72 hour period, but I don't have a spare XPS M1330 sitting around.<br /><br /><i>Jason Dunn owns and operates <a href="http://www.thoughtsmedia.com">Thoughts Media Inc.</a>, a company dedicated to creating the best in online communities. He enjoys mobile devices, digital media content creation/editing, and pretty much all technology. He lives in Calgary, Alberta, Canada with his lovely wife, and his sometimes obedient dog. He's mostly impressed with the XPS M1330.</i>

[MitchellO]

Jason, were your tests done with the 4GB ram largely full? Would that make a difference?

[Chris Gohlke]

I'm guessing it would not matter. I would think the entire 4 GB would need to be kept in the proper state. The RAM does not know what is data and what is not, it just know each space holds either a 0 or a 1 and I doubt power management is smart enough to know which is data is which is just space or that there is the ability to smartly cut power to part of the RAM.

[Jason Dunn]

Jason, were your tests done with the 4GB ram largely full? Would that make a difference?

Yeah, Chris is right, I'm not aware of any RAM that can be partially powered. I'm quite sure 4 GB of RAM draws the same amount of power whether it's full or empty. And remember, Vista aggressively fills the RAM with pre-cache data, so it was "full".

[Phronetix]

I am not following Jason's findings here.

If I understand this, he is saying that an 8% loss of power is roughly equal to a 9% loss. This is not really the case. It may be only off by 1% total, which seems insignificant. But the real comparison is with the baseline of 9% lost in the 4 GB of RAM The 1 GB of RAM used 1/9 less power or 8/9 of the power used by the 4GB, an 11% difference in power use. This indicates a potentially significant difference in power usage and battery life. It is only potentially significant because, as Jason points out, there is a +/- .5% measurement error. The other reason is that it is only one trial, and as such, is either 100% right or 100% wrong. So, statistically it doesn't matter... the reaults are not significant.

Of course, Jason's point is that over 12 hours in sleep mode, this is a small real world usage issue. I guess that it is hard to disagree with the conclusion. So I won't. Maybe I am just splitting hairs tonight.

:wink:


Edit: 2007-09-26, 23:47 MDT. Rewording my argument. Sounded like an ass initially. :-)

[Jason Dunn]

But the real comparison is with the baseline of 9% lost in the 4 GB of RAM The 1 GB of RAM used 1/9 less power or 8/9 of the power used by the 4GB, an 11% difference in power use. This indicates a potentially significant difference in power usage and battery life.

True. The only reason I was trying to "rush" this test is because you want to buy the Dell RAM off me, but if you're willing to wait a while longer, I'll happily re-do this test with longer timeframes. :wink:

The other reason is that it is only one trial, and as such, is either 100% right or 100% wrong. So, statistically it doesn't matter... the reaults are not significant.

I've done a lot of benchmarking over the years, and have found computers to be remarkably consistent - they do not vary much with multiple tests. But if it's the only way you'll believe the test was "real" I'll do each one more than once.

[Phronetix]

True. The only reason I was trying to "rush" this test is because you want to buy the Dell RAM off me, but if you're willing to wait a while longer, I'll happily re-do this test with longer timeframes. :wink:



Heheh. No. That is quite alright. I am looking forward to the RAM, more then the results of a more comprehensive study. I also have an item for you to evaluate and hopefully share with the site's readers. Maybe that will pique the curiosity of the folks here. ;-)