I make it 63gpm, assuming cooling load is the same as the power output. That is also a rather small delta.

35gpm free flow for a 500hp motor is definitely inadequate, though.

Math time!

Get back in your seats, boys and girl!

I don't like elvish, so I'm going to be using the black speech. Stop cringing at the back!

35gpm is 2.2 l/s. We don't know thew flow characteristics, so lets be generous and assume no flow drop in actual use.

Plucking 20kPa out of my backside (3psi for you heathens) as the pressure drop, the power to pump the water is an amazing 44 Watts. the 8A 12V pump would therefore be 46% efficient. Pulling 80% out of my backside as the motor efficiency, that's an impeller efficency of 57%. That's about right for a reasonably well designed centrifugal pump.

But, what about the mechanical pump?

Pay attention Smithers, strange as it may seem, maths may actually be useful to you one day!

It supposedly ate 8hp, or 6000 Watts. To do 44 Watts of work? 0.7% efficient? That's truly appalling. Something is seriously wrong there.

I have no idea how to even start to design a centrifugal pump with less than 1% efficiency, maybe one of the water pump gurus out there could clue me in? The least efficient water pump that I have designed and built managed 22% efficiency. That one had no moving parts and a construction error, I was hoping for high 20's...