I suggest replacing the MOSFET with an ideal switch (voltage-controlled switch). This simplifies the simulation and avoids the need for an accurate MOSFET model. This also allows you to remove the boot-strapped gate driver, further simplifying the circuit and increasing the chances of the simulator "working as expected".
After the circuit simulates well (ie: the simulator runs to completion with no errors, and is sufficiently fast) then proceed to investigate overall stability, and ensure the circuit is well-behaved on start-up and fault conditions.
After that is done, replace the switch with the real-world MOSFET you plan to use, but only after confirming that the MOSFET model is accurate, and of course, restore its boot-strapped gate driver.
Simulation is a game of stacking the odds in your favour: keep things simple to start with, use "ideal" components wherever possible, and gain an understanding of what effect the imperfections of "real-world components" have on circuit behaviour. Improving how well the simulated circuit matches the real world should be done slowly, carefully, and methodically.
And always ask "What am I trying to learn from this simulation?" For example, are you trying to determine how sensitive the start-up behaviour is to the type of MOSFET used, perhaps to a particular MOSFET parameter such as the drain-to-gate capacitance? Or are you more concerned with how the inductance of L1 changes with the current it carries, which may affect the stability of the control loop?
