Reenable parallel computation in the simulator

[arcondello]

Currently, looping over the basis states is not done in parallel, see

dwave-gate/dwave/gate/simulator/operation_generation.py

Line 331 in 10b70b6

# c.Pragma("omp parallel for"),

The reason is that supporting OMP in Python wheels is non-trivial because it requires the user to have omp installed. We should investigate using C++ parallel execution.