I'll pitch in one idea that Amarbir might have missed:
Technical fuel consumption is always measured in the laboratory by mass of fuel. This is true of wood, coal, diesel, gasoline, methane, etc. You burn so many grams or KG of fuel and get so much heat (or power) out. The reason for this is that the fuel's energy density is measured by unit weight - a gram of a particular diesel fuel makes so many calories of heat energy when burned.
But it's difficult to measure the weight of a container with a hose coming out of it, so the first thing the scientist does is take a sample of the fuel and find its density at standard temperature, and then calculate the energy density of a litre of fuel from that. It is very easy to measure the amount of ml of fuel consumed from a marked container, so this is the way it is usually done.
Fuels vary both in their density (grams/litre, usually about 850 for diesel fuel) and in their energy density (cal/litre, Mj/litre, diesel usually about 34 Mj/litre). Back when I worked in a lab, we had to measure a sample of fuel from each batch to determine just what it was so that we knew what its characteristics were. We would also run the fuel in an engine of known properties and verify that it behaved as the laboratory results and manufacturer numbers predicted.
It's important to compare fuel consumption results with known loads and known fuels, or you don't really know just what has changed to cause a different result.