NoiseOpAll, NoiseOp are slow. Needs kernel fusion.

[Shayan-P]

In the current implementation, in order to add noise on multiple qubits, we manually run a separate Noise kernel for each qubit (so length(noiseop.indices) are being run.

In order to improve this, one could write a kernel that gets indices as a CuArray parameter and fuse all of those in one kernel.