I also think this is an interesting topic so I did another google search.
This IEEE paper was most interesting.
An empirical study of performance and power scaling of low voltage DDR3
This paper appears in: Electrical Performance of Electronic Packaging and Systems (EPEPS), 2010 IEEE 19th Conference on
Issue Date: 25-27 Oct. 2010
E-ISBN: 978-1-4244-6866-9
As some of you probably have no IEEE access I will summ up the most important parts.
The biggest power savings are actually not at the DRAM side (about 20%) but at the MCH (around 40%). Still it is not a whole lot when not active.
DRAM+MCH / DDR3/DDRL/delta%
avg power mW 690/450/21%
self refresh mW 570/390/32%
Under full load difference was about 4-5%, which is around a 2W saving of power with 45W load power consumption of the tested notebook. Not a whole lot but about as much as taking away the entire power consumption of an HDD or 3 idle sitting HDDs.
Idle savings are much lower. 690-450 = 240mW. That is not very much with the idle power of a MBP it translates into about 10min more.
With Standyby almost only the DRAM is still active and the lower refresh power translates into a much bigger saving.
It was 1.3 vs 1.1 W with their notebook (7.5h added Standby). With MBPs low 0.6W or so it might be 0.4 W and yield you 30% longer standby time. That is just a guess though, one would need to check that out.
There are more tables but I am not entirely sure what they mean. It seems as some show only the DDR active power consumption without the total system power. And they tested different configurations. In Short. DRAM seems to be responsible for about 5-8W of consumed power under load and 3.5 (ddr3l) and 4.4 (ddr3) under light load. They tested different vendor modules and 8/4 GB configs. Active power between 8 and 4 GB is only negligible difference. There is more difference between different vendors. MCH IO is 30% less power. Another 1.7-2.3W high workload.
All in all it looks like the memory sub system is responsible for 7-10W of total system power under heavy load.
In conclusio the difference is about as big as cutting away the HDD or SSD power consumption completely. It is not so much that it really justifies significantly higher prices but 5% here and 5% there helps in the end.
Also keep in mind these numbers stay about the same so the benefit increases the less power a system needs. A netbook with an avg power of 4-5 W will benefit much more.
It seems that SSDs do not have a notable impact on power consumption. The SSD guide says that user reports differ in whether SSDs reduce or increase power consumption.
No but they help by increasing the run to idle tremendously. Launching my Eclipse dev suite on an SSD is done in about 1/4 the time and it consumes the same amount of power. The system sucks say 20W for 4 seconds as opposed to 16 seconds. That makes quite a huge practical difference. Also app launch speeds are so quick that it is less inconvenient to shut down apps one is currently not using and cut down on power. I used to have stuff like Eclipse running always because it took so long to launch with an SSD I can shut down everything but the essentials and get 7h battery life in stead of 4.
DRAM is still the same performance but with SSDs the speed helps a great deal.