arXiv:2610.18225
PPVM: A Circuit Simulation Framework for Digital Twins of Quantum Processors
Xiu-Zhe Luo, David Plankensteiner, Kai-Hsin Wu, Rob Patterson, Alexander Schuckert, Oriol Rubies-Bigorda, Refaat Ismail, Jonhas Colina, Neelay Fruitwala, Danylo Lykov, Tyson Jones, Stefan Krastanov, Rafael Haenel, Stefan Ostermann, Sheng-Tao Wang
Abstract
Quantum computers are approaching the early fault-tolerant regime. Verifying the next generation of quantum hardware requires joint emulation of the quantum processor, hardware noise, and classical control and decoding logic. We introduce the Programmable Pauli Virtual Machine (PPVM), an open-source framework for digital twins of quantum processors. PPVM unifies Pauli propagation and generalized tableau simulation under a common abstraction within a single virtual machine. Its bytecode supports arbitrary classical control flow, with measurement-conditioned feed-forward available in the generalized tableau backend. For neutral-atom hardware, we derive independent and correlated atom-loss channels and integrate them into both backends. We demonstrate loss-aware magic state distillation on over a thousand physical qubits and integration with classical decoding through Bloqade QEC Studio.
AI Summary
PPVM addresses joint simulation of near-Clifford quantum processors, hardware noise, and classical control for processor digital twins, with a focus on neutral atoms. It provides an open-source bytecode virtual machine with arithmetic, branching, loops, debugging, and interchangeable generalized-tableau and Pauli-propagation backends. Generalized tableaux support sampled measurements and feed-forward; Pauli propagation computes expectation values through backward observable evolution. The framework incorporates independent and correlated atom-loss channels, including different loss probabilities when a gate partner is already absent, and reset of lost atoms. These are implemented through classical loss flags in tableaux and an extended operator basis $\{I,X,Y,Z,L\}$ in Pauli propagation. Additional methods protect specified conserved-quantity components during truncation and approximate noisy-state sampling using truncated classical mixtures. A mixture-weight cutoff $C$ limits retained branches to at most $1/C$, with accuracy depending on discarded probability. Demonstrations include five-to-one magic-state distillation on up to $1435$ physical qubits, taking approximately $0.1$ seconds per shot on one Apple M4 thread at loss and depolarization probabilities of $10^{-4}$. Integration with classical decoding enables adaptive logical programs. A nine-qubit Heisenberg example checks conservation-preserving truncation against analytical survival scaling; simulator benchmarks provide performance comparisons. Limitations include exponential worst-case growth with non-Clifford operations, uncontrolled aggregate error from loss-branch truncation, and no measurement-conditioned feed-forward in Pauli propagation. Asymmetric loss in generalized tableaux and leakage in Pauli propagation remain future work. The decoder integration does not yet use loss reports.
AI agents iteratively propose and evaluate implementation changes using tests and benchmarks developed under human guidance.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 21 pages, 8 figures, 1 table. Code available at https://github.com/QuEraComputing/ppvm