QuteFuzz: Fuzzing Quantum Compilers Using Randomly Generated Circuits with Control Flow and Subcircuits

This conference talk presents QuteFuzz, a novel tool for detecting bugs in quantum compilers by generating random quantum programs with advanced features. Learn how researchers from Imperial College London developed a fuzzing approach that incorporates higher-level abstractions like subroutines and complex control flows (if-else, switch statements) with varying nesting depths and gate varieties. The presentation reveals how this methodology uncovered seventeen bugs across popular quantum software stacks including Pytket, Qiskit, and Cirq - ranging from compiler crashes to silent miscompilations. Discover the importance of compiler correctness in the Noisy Intermediate-Scale Quantum (NISQ) era, where optimized circuits are crucial for reliable results on quantum processors with limited logical qubits susceptible to noise. The researchers demonstrate how their approach generates more diverse test cases than previous tools like QDiff, exposing bugs in previously unexplored areas of quantum compilation systems.