Error not thrown when broadcasting custom unitary to a qvector.

[justinlietz]

Required prerequisites

• Consult the security policy. If reporting a security vulnerability, do not report the bug using this form. Use the process described in the policy to report the issue.
• Make sure you've read the documentation. Your issue may be addressed there.
• Search the issue tracker to verify that this hasn't already been reported. +1 or comment there if it has.
• If possible, make a PR with a failing test to give us a starting point to work on!

Describe the bug

When attempting to broadcast a custom unitary to a qvector, an error is not thrown. In the reproducer below, the kernel produces an incorrect result when sampled.

Steps to reproduce the bug

cudaq.register_operation("custom_i", np.array([1, 0, 0, 1]))                                                                                                                                                                                                                       
                                                                                                                                                                                                                                                                                   
@cudaq.kernel                                                                                                                                                                                                                                                                      
def bell():                                                                                                                                                                                                                                                                        
    qubits = cudaq.qvector(2)                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               
    custom_i(qubits) 

Expected behavior

An error is thrown when attempting to broadcast a custom unitary to a qvector.

Is this a regression? If it is, put the last known working version (or commit) here.

Not a regression

Environment

• CUDA-Q version:
• Python version:
• C++ compiler:
• Operating system:

Suggestions

No response

[schweitzpgi]

@amccaskey is the desired meaning to broadcast the custom unitary across each qubit? (That's the interpretation for the builtins.)

[Squirtle007]

+1
Accepting cudaq.qvector can provide flexibility and convenience for multi-qubit custom op within a kernel.

error: invalid number of arguments occurs when the code below is executed with cudaq.draw, cudaq.sample, etc:

num_qubits = 2
cudaq.register_operation("custom_op", np.identity(2**num_qubits))

@cudaq.kernel
def kernel(num_qubits: int):
    qubits = cudaq.qvector(num_qubits)
    custom_op(qubits)  # Doesn't work

Then we need to manually pass the qubits (or cudaq.qubit) one by one:

@cudaq.kernel
def kernel(num_qubits: int):
    qubits = cudaq.qvector(num_qubits)
    custom_op(qubits[0], qubits[1])