1: \begin{abstract}

2: %\noindent

3: In order to approximate solutions of stochastic partial differential equations (SPDEs) 

4: that do not possess commutative noise, one has to simulate the involved 

5: iterated stochastic integrals. Recently, two approximation methods for iterated 

6: stochastic integrals in infinite dimensions were 

7: introduced in~\cite{MR3949104}. As a result of this, it is now possible to apply 

8: the Milstein scheme by Jentzen and R\"ockner~\cite{MR3320928} to equations 

9: that need not fulfill the commutativity condition. We prove that the order 

10: of convergence of the 

11: Milstein scheme can be maintained when combined with one of the two approximation 

12: methods for iterated stochastic integrals. However, we also have to 

13: consider the computational cost and the corresponding effective order of 

14: convergence for a meaningful comparison with other schemes. An analysis 

15: of the computational cost shows that, in dependence on the equation, a 

16: combination of the Milstein scheme with both of the two methods may be 

17: the preferred choice. Further, the Milstein scheme is compared to the exponential 

18: Euler scheme and we show for different SPDEs depending on the parameters describing, 

19: e.g., the regularity of the equation, which one of the schemes achieves the 

20: highest effective order of convergence.

21: %

22: \end{abstract}

23: