(1) I only know people commonly use Monte Carlo, but not exactly as the complex analysis I saw last week in the Smoke Analysis enquiry post which can performed in PSpice. This is really not a topic I familiar.
e.g. Monte Carlo analysis - QSPICE - Qorvo Tech Forum
(2) I think everybody hate “timestep too small” in SPICE, including myself. For Qspice, I will suggest you try following
.option fastmath=0: disable fastmath and use Qspice80.exe, this will allow 80bit math calculation, more precise math, but a bit slower in simulation.option trtol=7 method=gear: if the model is PSpice model, you may consider that. PSpice use different trtol and integration method. But you may be careful gear integration added damper into circuit, which suppress oscillation and introduce more error.option gshunt=<value>or.option cshunt=<value>, these are to add conductance or capacitance from every node to ground. Sometime, they can help simulation to converge as they add a path for current to flow to ground. Setting value to 1e-12 or 1e-11 for a try..option maxstep: force a maximum timestep!!! Check your switching frequency or highest operating frequency, set a maxstep at least 1/10/fmax (i.e. 10 calculation samples per period)- If everything fail, post the circuit to forum to see if someone can help.
.option ITL4=<value>: this increase the number of transient iteration limit. May be set to 100 or 1000 to see if any improvement by allow more iteration.