QECTOR Workbench Decoder Benchmark Master Report (v0.6.6)

This repository contains the comprehensive dataset and performance metrics for the QECTOR Workbench Decoder Benchmark Master Report (v0.6.6), executed on July 13, 2026. The dataset aggregates 1,858 individual decode-timing measurements drawn from 105 independent benchmark runs. All evaluations were conducted under a fixed physical error rate (p = 0.05), spanning code distances from d=3 to d=19. The benchmark evaluates six distinct quantum error-correcting (QEC) code topologies:- Heavy Hex- Repetition- Ring- Rotated Surface- Toric- Unrotated Surface Performance analysis captures the raw latency (µs), throughput (harmonic mean decodes/s), and peak memory consumption (KiB) across multiple decoding algorithms. It highlights the tight performance profile of the fastest runtime decoders (cpu_batch, fast_union_find, and union_find) while documenting the latency-accuracy trade-offs of exact matching and belief-propagation implementations (blossom, sparse_blossom, and bp_osd) across scaling distances.