betaFeedback

WorksPhysics and Astronomy

QuFrame: Pauli Frame Tracking for Fault-Tolerant Quantum Control

Abstract

Fault-tolerant quantum computing requires continuous quantum error correction (QEC) to suppress physical noise. While Clifford-only workloads can defer corrections to software post-processing, universal computation breaks this paradigm: executing non-Clifford gates via magic-state gadgets introduces hard, real-time control-flow dependencies on error-corrected measurements. Consequently, the controller must extract syndromes, decode them, and update the Pauli frame in real time, closing the entire QEC loop within a strict 1 μs1\,μ\text{s} cycle. However, prior proposals on Pauli frame tracking remain largely theoretical, leaving the microarchitectural bottlenecks of tracking, resolving, and synchronizing frame updates across distributed control pipelines unaddressed. To this end, we present QuFrame, a control microarchitecture for real-time Pauli frame tracking that closes the QEC loop within 1 μs1\,μ\text{s} while sustaining QPU-rate throughput. By introducing QuFrameISA, we cleanly decouple instruction execution from continuous Pauli frame management. Execution is distributed across per-qubit Instruction Resolution Engines (IREs) that resolve instructions against the frame via a single table lookup, while a Pauli Frame Management Unit enables independent, lock-free concurrent commits. To strictly guarantee the 1 μs1\,μ\text{s} deadline, QuFrame pairs an asynchronous primary decoder with an on-chip fallback decoder that resolves in 144 ns144\,\text{ns}, natively reconciling divergent updates in hardware without degrading logical error rates.

Paper details

vlnc:W_yswnjd7h
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
arXiv nonexclusive license
Usage

views · 4

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…