Manuscripts & Research Notes
Q-Psi prepares formal scientific manuscripts documenting quantum compilation architectures, physical QPU benchmark results, and oracle query-complexity advantages.
Compiler-Enabled Quantum Query Advantage for Candidate-State Search on Physical Quantum Hardware
Reports an empirical demonstration of compiler-enabled quantum query advantage for software-repair candidate-state search on a 156-qubit Heron processor (ibm_marrakesh). Across 9 frozen instances from 7 language ecosystems at N=4, N=8, and N=16, physical quantum execution achieved lower effective verifier query complexity than the classical black-box baseline (9/9 cases, 3/3 problem sizes; N=16 95% CI upper bound 7.360 < 8.5).
Quantum-Assisted Synthetic Code Data Generation via Physical QPU State Selection and Executable Verification
Demonstrates an end-to-end physical-QPU-assisted coding data generation pipeline on IBM Heron r2 (127.58 quantum seconds, 332,768 shots across 4 jobs). QPU sampling of a 65,536-state universe supplied 65,138 candidate states converted by automated classical verifiers into 93 unique verified positive code repair records (1.593 MB, ~417k token-equivalent) with 100% QPU provenance.
Mid-Circuit Measurement and Qubit Reuse Compilation on Superconducting Quantum Hardware
Evaluates dynamic qubit reuse compiler passes on IBM Heron (ibm_marrakesh), compressing active physical register width by a median 78.1% (up to 87.5% on 16 logical qubits mapped to 2 physical slots) while satisfying pre-registered non-inferiority bounds across all 6 tested streaming circuits.
Hardware-Aware Compiler Routing on Superconducting Processors: Depth-Fidelity Tradeoffs and Latency Limits in Physical Dynamic Circuits
Reports physical QPU benchmark results comparing unitary SWAP routing against dynamic measurement and feed-forward routing on IBM Heron (ibm_marrakesh, 156 programmable qubits, Job da1t22mg52gs73cm31i0, 18,432 shots). While dynamic circuits maintained constant two-qubit depth (2 vs 13, 37, 73), unitary routing won Bell fidelity on SHORT (3 hops) and MEDIUM (7 hops), with LONG (13 hops) inconclusive. Pre-QPU model confirmed 2/2 conclusive predictions.
Cross-Compiler Benchmarking on IBM Heron: Q-Psi vs Qiskit O3 and TKET on the MQT Benchmark Suite
Evaluates 18 compilation cases and 6 physical execution cases on IBM Marrakesh against industrial standard compilers Qiskit O3 and TKET. Documents valid 18/18 compilation, workload-dependent physical competitiveness (-0.0067 mean delta vs Qiskit, +0.0398 mean delta vs TKET), and analyzes sensitivity to single-workload routing failures.
Experimental Demonstration of Algorithmic Query Advantage in Single-Shot Dynamic Bernstein-Vazirani
Reports an empirical demonstration of quantum query-complexity advantage for dynamic single-shot Bernstein-Vazirani on a 156-qubit Heron processor. Measured scaling exponent alpha_Q = 0.1532 vs classical alpha_C = 0.6963 (p = 3.47e-7).