Circuit-level decoders are essential for the realisation of low-overhead fault-tolerant quantum computing. However, they rely on complex hypergraphs that are traditionally compiled ahead-of-time. This static approach introduces a significant bottleneck for an emerging class of adaptive circuits, where the structure is modified during execution based on mid-circuit measurement outcomes. Pre-compiling hypergraphs for all possible circuit branches would incur an exponential memory cost, rendering current tools impractical for these workloads. Hence, we introduce GreenPeas, a C++/CUDA toolchain for the high-speed, just-in-time compilation of decoding hypergraphs. By lowering the circuit to a space-time error propagation graph, we show how Stim's backtracking algorithm can be mapped efficiently onto massively parallel GPU architectures, decomposing the O(nl) workload for a circuit with n qubits and l gate layers across thousands of concurrent threads. Our implementation achieves a greater than 10x average speedup over the Stim baseline across two of the leading fault-tolerant architectures: the surface and bivariate bicycle codes. As a key use case, we demonstrate that this speedup enables circuit-level decoding of adaptive syndrome measurement circuits, unlocking a regime previously restricted to less accurate phenomenological decoders. We aim to open-source GreenPeas to support the research of future adaptive circuit protocols.
Circuit-level decoders are essential for the realisation of low-overhead fault-tolerant quantum computing. However, they rely on complex hypergraphs that are traditionally compiled ahead-of-time. This static approach introduces a significant bottleneck for an emerging class of adaptive circuits, where the structure is modified during execution based on mid-circuit measurement outcomes. Pre-compiling hypergraphs for all possible circuit branches would incur an exponential memory cost, rendering current tools impractical for these workloads. Hence, we introduce GreenPeas, a C++/CUDA toolchain for the high-speed, just-in-time compilation of decoding hypergraphs. By lowering the circuit to a space-time error propagation graph, we show how Stim's backtracking algorithm can be mapped efficiently onto massively parallel GPU architectures, decomposing the O(nl) workload for a circuit with n qubits and l gate layers across thousands of concurrent threads. Our implementation achieves a greater than 10x average speedup over the Stim baseline across two of the leading fault-tolerant architectures: the surface and bivariate bicycle codes. As a key use case, we demonstrate that this speedup enables circuit-level decoding of adaptive syndrome measurement circuits, unlocking a regime previously restricted to less accurate phenomenological decoders. We aim to open-source GreenPeas to support the research of future adaptive circuit protocols.
Quantum computers can solve selected problems more efficiently than classical computers, but current devices are limited by high physical error rates. Quantum Error Correction (QEC) mitigates this by encoding many physical qubits into a logical qubit, with the surface code among the most widely studied approaches. Since syndrome measurements are produced continuously, the decoder must process them fast enough to avoid becoming a system bottleneck, making real-time decoding essential for fault-tolerant quantum computing. While most state-of-the-art real-time decoders rely on Minimum-Weight Perfect Matching (MWPM), we instead use a high-accuracy Graph Neural Network (GNN) that trades higher computational cost for lower logical error rates. To make this GNN practical for real-time decoding, we apply algorithm-hardware co-design. We reduce complexity through hardware-guided pruning and retraining, producing two hardware-friendly models that reduce parameter count by 3.1× and 6.5×. These target, respectively, an average decoding latency of one syndrome cycle and a worst-case latency within one syndrome cycle. We further reduce hardware cost using input-graph filtering and post-training quantization. Based on these optimized models, we propose an FPGA-based architecture for low-latency inference and real-time decoding. Evaluated on surface codes up to distance 7 under circuit-level noise at physical error rate p=10−3, our decoder outperforms MWPM in decoding accuracy for both average-latency and max-latency settings. It reduces logical error rate by 40% at 1μs average latency, and by 13% under a strict 1μs deadline.
Real-time decoding is a major bottleneck in scaling quantum error correction (QEC) from noisy intermediate-scale quantum (NISQ) devices to fault-tolerant quantum computing. We present an adaptive confidence-gated decoding framework for the rotated surface code that treats decoding as a two-stage inference problem. A lightweight feed-forward neural network performs fast-path decoding for the majority of syndrome measurements, while only low-confidence predictions are escalated to a minimum-weight perfect matching (MWPM) refinement stage. We benchmark the framework on rotated surface codes with distances d∈{3,5,7,9,11} under circuit-level depolarising noise using the Stim stabiliser simulator. The evaluation characterises logical accuracy, confidence-controlled accuracy-latency trade-offs, decoding throughput, per-shot latency, and decoding-graph resource scaling. Routing only 3.3%-6.2% of syndromes to the refinement stage improves logical accuracy from 99.21% for the neural-only baseline to 99.81% at a confidence threshold of 0.95 while incurring only a bounded increase in average decoding cost. Neural-decoder throughput saturates near 4.6×105 samples s−1 at batch size 512 on commodity CPU hardware, indicating that the neural fast path is not the dominant throughput bottleneck beyond code distance d=7. We release the complete benchmarking pipeline, trained models, raw benchmark data, and source code, and explicitly distinguish the experimentally validated contributions from the broader hardware-aware QEC co-design roadmap, including hardware-constrained code discovery, GPU-accelerated inference, and multi-noise optimisation, which remain directions for future work.
Device error rates on current quantum computers have improved enough to where demonstrations of error correction below break even are now possible. Still, the circuits required for quantum error correction introduce significant overhead and sometimes inject more errors than they correct. In this work, we introduce adaptive syndrome extraction as a scheme to improve code performance and reduce the quantum error-correction cycle time by measuring only the stabilizer generators that are likely to provide useful syndrome information. We provide a concrete example of the scheme through the [[4,2,2]] code concatenated with a hypergraph product code and a syndrome-extraction cycle that uses quantum error detection to modify the syndrome-extraction circuits in real time. Compared to nonconcatenated codes and nonadaptive syndrome extraction, we find that the adaptive scheme achieves over an order of magnitude lower logical error rates while requiring fewer gates and physical qubits. Furthermore, we show how to achieve fault-tolerant universal logical computation with [[4,2,2]]-concatenated hypergraph product codes.
Quantum Error Correction (QEC) is essential for building robust, fault-tolerant quantum computers; however, the decoding process often presents a significant computational bottleneck. Tesseract is a novel Most-Likely-Error (MLE) decoder for QEC that employs the A* search algorithm to explore an exponentially large graph of error hypotheses, achieving high decoding speed and accuracy. This paper presents a systematic approach to optimizing the Tesseract decoder through low-level performance enhancements. Based on extensive profiling, we implemented four targeted optimization strategies, including the replacement of inefficient data structures, reorganization of memory layouts to improve cache hit rates, and the use of hardware-accelerated bit-wise operations. We achieved significant decoding speedups across a wide range of code families and configurations, including Color Codes, Bivariate-Bicycle Codes, Surface Codes, and Transversal CNOT Protocols. Our results demonstrate consistent speedups of approximately 2x for most code families, often exceeding 2.5x. Notably, we achieved a peak performance gain of over 5x for the most computationally demanding configurations of Bivariate-Bicycle Codes. These improvements make the Tesseract decoder more efficient and scalable, serving as a practical case study that highlights the importance of high-performance software engineering in QEC and providing a strong foundation for future research.
Fast, scalable decoding architectures that operate in a block-wise parallel fashion across space and time are essential for real-time fault-tolerant quantum computing. We introduce a scalable AI-based pre-decoder for the surface code that performs local, parallel error correction with low decoding runtimes, removing the majority of physical errors before passing residual syndromes to a downstream global decoder. This modular architecture is backend-agnostic and composes with arbitrary global decoding algorithms designed for surface codes, and our implementation is completely open source. Integrated with uncorrelated PyMatching, the pipeline achieves end-to-end decoding runtimes of order O(1μs) per round at large code distances on NVIDIA GB300 GPUs while reducing logical error rates (LERs) relative to global decoding alone. In a block-wise parallel decoding scheme with access to multiple GPUs, the decoding runtime can be reduced to well below O(1μs) per round. We observe further LER improvements by training a larger model, outperforming correlated PyMatching up to distance-13. We additionally introduce a noise-learning architecture that infers decoding weights directly from experimentally accessible syndrome statistics without requiring an explicit circuit-level noise model. We show that purely data-driven graph weight estimation can nearly match uncorrelated PyMatching and exceed correlated PyMatching in certain regimes, enabling highly-optimized decoding when hardware noise models are unknown or time-varying, as well as training pre-decoders with realistic noise models. Together, these results establish a practical, modular, and high-throughput decoding framework suitable for large-distance surface-code implementations.
Realizing practical quantum computers requires quantum error correction, but the most widely implemented approach, the planar surface code, demands a substantial physical qubit overhead. Here, we demonstrate that partitioning quantum processors into manageable, flat modules with non-local inter-module connections provides a natural solution. By wiring sparse, static boundary connections between Euclidean planar modules, we construct modular hyperbolic surface and color codes, based on new families of semi-hyperbolic codes. Using modular syndrome extraction circuits and efficient neural network and matching decoders, our circuit-level simulations show that this modular memory design can reduce the physical qubit overhead by tenfold or more compared to the surface code, even under elevated inter-module seam error rates. Projecting to larger system sizes, we find modular codes with encoding rate k/n=1/16 and distance d≥22, exceeding both the rate and distance of the [[288,12,18]] two-gross bivariate bicycle code while using mostly nearest-neighbor gates within planar modules, and implying an over 30× overhead reduction relative to surface codes. We develop fault-tolerant walking circuits for implementing logic via code automorphisms, which we co-design with low-weight, highly symmetrical logical bases and a modular extractor system that is ≈4.5× smaller than the code itself. Combining these methods, our modular memory can be used as dense storage in a universal fault-tolerant architecture, unlocking the efficiency of high-rate codes without sacrificing the fabrication advantages of planar modules.