dynamic range - dxomarks and dpreview do not give he same resuts

Dynamic range is the number of stops from the noise floor to the saturation limit, and is dependent upon the area over which the DR is measured.

For example, the DR / pixel will be less than the DR / µphoto, where a µphoto is a millionth of a photo (e.g. 24 pixels for a 24 MP photo).

While the choice of the noise floor is arbitrary (DxOMark uses the 100% NSR, sensorgen uses the read noise, and Bill Claff uses a 20% NSR), comparing the DR between systems on the basis of equal area on the photo (as opposed to equal are on the sensor) is the relevant measure in terms of the photo (DxOMark's screen measure is per-pixel, and their print measure is per-pixel of a photo normalized to 8 MP).

In any case, the utility of DR is simply that more DR allows more lattitude in terms of the tone curves that can be applied to the photo.  So, for those that decry DR measurements as being bogus because they don't see any evidence of greater DR from sensors that have greater DR, that's because they've not seen photos which apply tone curves to the captured files that make use of the sensor's extended DR.