FTWeave: Composing Logical Execution, Real-Time Decoding, and Magic-State Supply in Tightly Integrated Quantum-Classical Systems

Abstract

Fault-tolerant quantum applications depend on both decoded measurement information and a continuing supply of magic states. While an operation waits for either, its logical data require protection that produces more syndrome data and decoding work. Architects must provision the quantum and classical parts of the machine together, before complete systems are available to measure. FTWeave is a system-level simulator for this architectural exploration. It accepts OpenQASM 2/3 circuits, CUDA-Q Logical compiler artifacts, Stim circuits, and provider-derived physical inputs through adapters that preserve operation and measurement dependencies. Users bind these inputs to quantum resources, classical processors, memory, communication, and factory configurations. One event timeline determines application progress, resource ownership, and continuing protection. This supplies the system model needed to connect circuit and decoder evidence to application performance. Controlled demand studies and complete multiplier and Ising circuits show why the coupling changes provisioning decisions. A fixed-work approximation predicts a 3.55x multiplier speedup from additional factories where the coupled model gives almost none. Explicit ownership also changes useful capacity: increasing decoder workers alone has no benefit in a storage-constrained case, while increasing workers and feedback entries together gives 4.93x speedup. Selected color and hypergraph-product code mappings take 53-90% longer under decoder service variability, despite an unchanged mean. These comparisons identify useful resource combinations and distinguish protection work from the elapsed lifetime of quantum data.