The "cylinder" or "indicated" horsepower (same thing) is measured by the use of an "indicator card" that measures the mean effective pressure in one cylinder of a locomotive. (Most tests used indicator cards on both sides of the engine). The MEP was measured to "point of cutoff", and the card would show this. Cards were taken at both ends of the cylinder. A photo of an Indicator card is on p. 60 of Staufer's "Thoroughbreds". The speed of the locomotive and the cutoff at the time that the card was taken was noted. Both ends of each cylinder times 2 cylinders is "cylinder horsepower" for a two cylinder locomotive.
Drawbar Horsepower is the amount of "pull", in pounds, at the coupler at the rear of the tender, and the speed at which this "pulling force" occurs. This is usually done with a dynamometer car. PRR was an exception. What PRR did at the Altoona test plant was to measure, with the engine driving wheels on rollers, the pull at the engine's drawbar, ignoring the weight and resistance of the tender and with no deduction for wind resistance since the engine was not moving.
Unless the track was both level and straight and the acceleration (but not the speed) was held to "zero", corrections were required to obtain true drawbar pull. That is why NYC used a "brake engine" behind their dynamometer car to hold the speed constant, and could easily replicate multiple readings at any speed.
So DBHP is calculated as Pulling force (in pounds) x speed (in MPH), divided by "375". (The "375" is called the "horsepower constant", and is derived by the conversion of "one HP = 33,000 pounds hoisted one foot in one second".) (To correctly calculate DBHP, acceleration has to be held to "zero" since it takes more horsepower to accelerate, just like your car.)
For example, if a locomotive can pull with a force of 16,000 lb. at 100 MPH, its drawbar HP (AT 100 MPH) is 4267. (By the way, to my knowledge there were only two steam locomotives that could exceed 4000 DBHP at 100 mph, a PRR T-1 and a NYC Niagara.)
I have several test reports from the Altoona Test Plant, and it appears that PRR was mostly interested in boiler performance and the coal and water rates. For example, the Plant started to record drawbar pull at 35 mph, believing that lower speeds would damage the plant's rollers, and "drew" the remainder of the pull curves to "zero speed" by "filling in" the curve below 35 mph. Since the "general" speed limit of PRR trains in the steam era was 40 mph, the curves were almost useless in determining over the road performance.
The weakness of a steam locomotive was that it produced its maximum HP at one speed. A diesel has an "almost constant" horsepower throughout its entire speed range, so its "average horsepower" during a trip is usually much higher than a similarly rated steam locomotive.