the power-per-dimm numbers puzzle me too. the problem seems to be that the text is lifted from:
which is a slightly confused reuse of an example ("greener with ddr3") from:
the example is actually a power comparison for two 48 GB systems, one with 1.8V 1Gb ddr2 versus 1.35V 4G ddr3. rather extreme worst-case versus best-case, but it does illustrate the point. (which has to be interpreted as "minimize part count and voltage" really!)
Actually, I've often heard that in servers, nothing uses more power than the DRAM. Remember that each module here has 64 chips, compared to 16 or less for a consumer grade module. So I am not surprised at the power numbers. Notice that Samsung's 4GB module based on 16x2 gigabit, 30nm parts uses 3 watts, corresponding to 24 watts for a 32 GB module at this smaller etch size.
A Book For All Reasons Bernard Cole1 Comment Robert Oshana's recent book "Software Engineering for Embedded Systems (Newnes/Elsevier)," written and edited with Mark Kraeling, is a 'book for all reasons.' At almost 1,200 pages, it ...