Electronic – LTC4440 boosting in LTSpice, but not in reality

I have a simple LTC4440-5 circuit in both LTSpice:

(R1 is just a load resistor, which isn't present in reality)

And here in the matrix, there's:

And:

(Note that a few parts are slightly different from simulation to reality: The inductor is a 5uH part, the diodes are both DB2W40300L shottkys, and the mosfet is an AO3400A. However, I can plug any mosfet and any diodes into the simulation and I can't break it.)

I have tried running at both 1mhz, and 100khz input frequency. In LTSpice, everything's peachy; in reality, I get no change from a floating gate.

Now, in the LTSpice simulation, I get this on the boosted gate pin:

And in reality, I get this:

The gate floats at ~3v.

The mosfet seems to be half on, as the TS/inductor pin is at 4v, no load.

I've confirmed that all the rails are as they should be; 5v VCC, 12v on the mosfet, etc. I've tried swapping out 4440s, tested all the diodes, changed input frequency, etc. Nada. I've tried to break the LTSpice simulation to get it to approximate the behavior, but I can't get it to not charge pump.

What have I done wrong?

EDIT: Okay! Now we're getting somewhere. With a 100r resistor as a load in ltspice, I get a nice and stable charge pump. With a 130r resistor in reality, I get a sick 5khz oscillation in the charge pump:

(And zoomed)

Which… certainly wasn't the behavior I was expecting.