All execution functions should be able to broadcast automatically.

Required prerequisites

• Search the issue tracker to check if your feature has already been mentioned or rejected in other issues.

Describe the feature

We have 3 execution functions: sample, observe, get_state.

We also have their async versions.

Currently observe and sample can broadcast across many parameters however get_state cant and neither can any of the async versions.

import cudaq
from cudaq import spin
import numpy as np

cudaq.set_target("nvidia")

qubit_count = 5
sample_count = 10
hamiltonian = spin.z(0)

parameters = np.random.default_rng(13).uniform(low=0,high=1,size=(sample_count, qubit_count))

        
@cudaq.kernel
def kernel(theta:list[float]):
    qubits = cudaq.qvector(qubit_count)
    for i in range(qubit_count):
        rx(theta[i], qubits)
        
result = cudaq.observe(kernel, hamiltonian, parameters)
result = cudaq.sample(kernel, parameters)

# result = cudaq.get_state(kernel, parameters)  

# result = cudaq.observe_async(kernel, hamiltonian, parameters, qpu_id= 0)

# result = cudaq.sample_async(kernel, parameters, qpu_id= 0)

# result = cudaq.get_state_async(kernel, parameters, qpu_id= 0)

RuntimeError: Cannot pass ndarray with shape != (N,).

Can we enable broadcasting for all execution functions please?

Thanks team.

Start by tracing the execution entry points for sample, observe, get_state, and their async variants, focusing on how parameter arrays are validated. Reproduce the reported ndarray shape error, then verify that all six functions accept broadcast parameter arrays and return results consistently across synchronous and asynchronous execution.