When you really start increasing fuel delivery, it is easily possible to pull a vacuum at the intake port of the IP. I've done a lot of measurements with this, and was successful with several solutions, using various fuel pumps. A CIS pump will work fine, as will most FI pumps. Be advised, the gas fuel pump will draw more current when used on our cars.
As far as the 8 whp increase with the fuel pump on, I would suspect with the fuel pump off, it may simply be from a slight or strong vacuum being created in the fuel line, which decreases power. The process qualitatively behaves as follows: a noticeable, fairly linear decrease in power, injection timing starts to become retarded, then onto a massive, non-linear loss in power etc. The power loss will be most prevalent wherever maximum fuel delivery occurs, and often well before it if you are pumping a large quantity of fuel.
A good test would be to dyno the car a few times without the fuel pump producing full pressure, but with no vacuum, say at 1psi fuel pressure. A PWM power supply allows this, or a simple bypass. Then, dyno the car under the same starting conditions (engine temp, IC temp, hopefully same atm conditions, etc) again a few times with the fuel pump producing full pressure.