I am a physicist at heart, and so I think One-Way Quantum Computing is brilliant. In particular, Graph State Measurement-based Quantum Computing (MBQC) has been a really nice development in Quantum Computing research as originated by Raussendorf & Briegel. One just needs to prepare a multi-partite entangled state as described by a graph, and then perform sequential measurements on each node or qubit (adaptive measurements for deterministic computations).
Another excellent aspect of this approach is that Clifford circuits can be implemented in a single-round of measurements as shown by Raussendorf, Browne and Briegel. These circuits can be classically simulated (efficiently) as shown by Gottesman and Knill so it is an interesting connection between classical simulation and temporal resources.
However, not all temporally flat Graph State MBQC circuits (consisting of one round of measurements) are believed to be simulatable classically. For example, families of circuits in the quantum circuit model consisting of commuting gates called IQP circuits as introduced by Shepherd and Bremner can be implemented in single time-step in MBQC. These IQP circuits are believed not to be classically simulatable (in computational complexity terms, it would lead to a collapse of the polynomial hierarchy).
See also a nice description of a class of circuits implemented in one time-step here. Given that commuting/diagonal unitaries can have some interesting behaviour but non-commuting circuits be classically simulatable. It would be interesting if there were non-commuting circuits that can be implemented but not yet shown to be classically simulatable.
Anyway, my question is:
Are there other interesting circuits that can be implemented in a single time-step in MBQC?
Though I would prefer relations to computational complexity or classical simulation, I would find anything interesting.
Edit: After Joe's excellent answer below, I should clarify a couple of things. As Joe said (and somewhat embarrassingly I have said in one of my own papers), single measurement-round MBQC circuits are in IQP. To be more precise, I am interested in interesting circuits in the problems in IQP that can be implemented in one round of measurements in MBQC. Clifford circuits are an interesting example. If there any other examples which are classically simulatable that would be extremely interesting. Since simulating IQP circuits is believed to be unlikely classically, it would be interesting to find instances of circuits that are.