Optimize kernel mode for hybrid QNN test case

[bmhowe23]

This PR optimizes for potential future modifications to examples/python/tutorials/hybrid_qnns.ipynb.

Enumerating and looping over all elements in inspect.stack() was incredibly slow when running in a Jupyter notebook. (Interestingly enough, the slowness was not apparent when running a Python script with the python3 command.)

Here are some sample timings before the change:

# Kernel mode
qubit_count = 1
@cudaq.kernel
def kernel():
    qubits = cudaq.qvector(qubit_count)
cudaq.sample(kernel)
%timeit cudaq.sample(kernel)
# 4.73 ms ± 156 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

# Builder mode
qubit_count = 1
kernel = cudaq.make_kernel()
qubits = kernel.qalloc(qubit_count)
cudaq.sample(kernel)
%timeit cudaq.sample(kernel)
# 220 µs ± 26 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)

And after the change:

# Kernel mode
qubit_count = 1
@cudaq.kernel
def kernel():
    qubits = cudaq.qvector(qubit_count)
cudaq.sample(kernel)
%timeit cudaq.sample(kernel)
# 209 µs ± 6.2 µs per loop (mean ± std. dev. of 7 runs, 10,000 loops each) <------- large reduction compared to 4.73 ms

# Builder mode
qubit_count = 1
kernel = cudaq.make_kernel()
qubits = kernel.qalloc(qubit_count)
cudaq.sample(kernel)
%timeit cudaq.sample(kernel)
# 200 µs ± 16.3 µs per loop (mean ± std. dev. of 7 runs, 1,000 loops each)

This is a >20x speedup for this test case, which brings the kernel mode in line with the builder mode.

[github-actions]

CUDA Quantum Docs Bot: A preview of the documentation can be found here.

[github-actions]

CUDA Quantum Docs Bot: A preview of the documentation can be found here.

[amccaskey]

+1

[github-actions]

CUDA Quantum Docs Bot: A preview of the documentation can be found here.