Laptop GPUs can simulate quantum circuits with a memory trick

State-vector quantum simulation struggles on integrated GPUs because accessing the full state vector thrashes memory bandwidth — performance can actually worsen compared to the CPU as qubit count grows. By reorganizing the state vector so that frequently accessed blocks fit in last-level cache, the authors cut costly main-memory fetches and recover consistent GPU advantage. Tested on Intel, AMD, and Apple M1 Pro hardware running Quantum Phase Estimation, the method flipped a net slowdown on an Intel Core i5 (0.95× → 1.89×) and boosted the M1 Pro from 3.71× to 5.88× at 28 qubits. The work makes high-qubit simulation practical on ordinary laptops without vendor-specific frameworks.