We measured the power consumption on the SDRAM on the EZKit lite recently, and got some odd results. We didn't do it "properly" with a 'scope on the probe point (we don't have one), just a multimeter, which may or may not have anything to do with it.
But the answer we got, with a particular software running, is that the power varied cyclically, with a period of about 5 seconds, a factor of about 2 (80->180 mW). The repeat period of our program is ~300 ms, so it isn't that directly at least. Does it remind anyone of anything that I should be suspicious of?
Also strange, most of the other measurements we took indicated that in general the DDR RAM power did vary systematically with what software was running (as expected), but not in a predictable way - i.e. not compared to which programs are using external RAM. In particular, the "semi-idle" program, which just basically adds 1 to a section of internal memory without touching external memory at all - that one used the MOST DDR RAM power.
Again - not having tried this before on any board, not sure what to expect, but it didn't make any sense to me.
We are pretty sure that the various variables are in the memory-space we expect, based on both the memory-map inspection, and also the execution-time - shoving stuff out multiplies the execution-time by a factor of ~10 as you expect
The core processor power measurements were all stable, and made a bit more predictable sense based on what software was running.
Many thanks if this sparks any comments along the lines of "obviously you've thought of X....."