Applying a gate to the cutensors wavefunction is slow

[ermalrrapaj]

Hi, I followed the instructions in the closed issues section and also the example provided in gate_evolution.jl and I get the following warning after line 46 from gate_evolution.jl (ψ = apply(gates, ψ0)): Warning: Performing scalar operations on GPU arrays: This is very slow, consider disallowing these operations with allowscalar(false)

Since for a simulation of a real time evolution of the wavefunction one needs to repeatedly apply a series of gates to the wavefunction, I am worried there is no actual speed up from the GPU version of ITensors. Am I wrong in being worried? Is there something that can be done to convert the gate application to matrix multiplication (not a scalar operation)? Thank you in advance.