I don't really believe in compromise of high MP and high ISO, I really like Nikon's approach with D3x (high MP) and D3s (high ISO)
Here is a little trick that anyone can do. Download the RAW photos taken with those 2 Nikons from:
Load them in your RAW editor, deactivate all processing, and export them as JPEG (100% quality) or TIFF; in LR you need to increase the exposure of the photo from the D3x with about 0.5 stops (to match the brightness). Then load the results in your photo viewer, in full screen (NOT at 100% view / pixel level, but full screen because that's how people see photos on displays or on paper).
Then see if the difference in the noise levels from the two photos, considering the years of extra research for D3s + only half the resolution, means more than diddly squat.
Disclaimer: I believe that the photos taken by IR are taken in different amounts of light, which affects a scientific test, but since here two Nikon cameras are compared, I see no problem (despite the huge difference in exposure compensation and post-processing exposure adjustment which doesn't favor the D3x).