Segfault during circuit simulation with 2^19 controlled rotation gates

[bmhowe23]

Required prerequisites

• [X] 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.
• [X] Make sure you've read the documentation. Your issue may be addressed there.
• [X] 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

If you have too many controlled rotation gates (cr1, crx, etc.), the simulation will segfault. This happens with at least the following targets: nvidia, nvidia-fp64, and qpp-cpu, so the problem is likely outside of the backend-specific code.

The problem seems to happen with both Python and C++ CUDA-Q programs (when compiling with --enable-mlir). I was unable to reproduce the problem in C++ when NOT using --enable-mlir.

Steps to reproduce the bug

Create a test program like this:

import cudaq

@cudaq.kernel
def kernel():
    qvec = cudaq.qvector(2)
    for _ in range(2**19):               # the problem does not occur with 2**18
        r1.ctrl(1., qvec[0], qvec[1])

print(cudaq.sample(kernel))

And then run it:

$ python3 test.py
Segmentation fault

Expected behavior

The program should run successfully without a segfault.

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

Not a regression

Environment

• CUDA Quantum version: problem exists in latest, 0.7.1, and 0.7. I did not test earlier.
• Python version:
• C++ compiler:
• Operating system:

Suggestions

No response