1: \begin{abstract}

2: In the present study,

3: the efficiency of preconditioners for solving linear systems associated with

4: the discretized variable-density incompressible Navier--Stokes equations

5: with semi-implicit second-order accuracy in time

6: and spectral accuracy in space

7: is investigated.

8: The method,

9: in which the inverse operator for the constant-density flow system

10: acts as preconditioner,

11: is implemented for three iterative solvers:

12: the General Minimal Residual,

13: the Conjugate Gradient and

14: the Richardson Minimal Residual.

15: We discuss the method,

16: first, in the context of the one-dimensional flow case

17: where a top-hat like profile for the density is used.

18: Numerical evidence shows that the convergence is significantly improved

19: due to the notable decrease in the condition number of the operators.

20: Most importantly,

21: we then validate the robustness and

22: convergence properties of the method on two more realistic problems:

23: the two-dimensional Rayleigh--Taylor instability problem

24: and the three-dimensional variable-density swirling jet.

25: \end{abstract}

26: