Diversity Methods for Improving Convergence and Accuracy of Quantum Error Correction Decoders Through Hardware Emulation

As quantum computing moves toward fault-tolerant architectures, quantum error correction (QEC) decoder performance is increasingly critical for scalability. Understanding the impact of transitioning from floating-point software to finite-precision hardware is essential, as hardware decoder performance affects code distance, qubit requirements, and connectivity between quantum and classical control units. This paper introduces a hardware emulator to evaluate QEC decoders using real hardware instead of software models. The emulator can explore <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mn>10</mml:mn> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mn>13</mml:mn> </mml:mrow> </mml:msup> </mml:math> different error patterns in 20 days with a single FPGA device running at 150 MHz, guaranteeing the decoder's performance at logical rates of <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:msup> <mml:mn>10</mml:mn> <mml:mrow class="MJX-TeXAtom-ORD"> <mml:mo>&amp;#x2212;</mml:mo> <mml:mn>12</mml:mn> </mml:mrow> </mml:msup> </mml:math> , the requirement for most quantum algorithms. In contrast, an optimized C++ software on an Intel Core i9 with 128 GB RAM would take over a year to achieve similar results. The emulator also enables the storage of uncorrectable error patterns that generate logical errors, allowing for offline analysis and the design of new decoders. Using results from the emulator, we propose a method that combines several belief propagation (BP) decoders with different quantization levels, which we define as a diversity-based decoder. Individually, these decoders may show subpar error correction, but together they outperform the floating-point version of BP for quantum low-density parity-check (QLDPC) codes like hypergraph or lifted product. Preliminary results with circuit-level noise and bivariate bicycle codes suggest that hardware insights can also improve software. Our diversity-based proposal achieves a similar logical error rate as the well-known approach, BP with ordered statistics (BP+OSD) decoding, with average speed improvements ranging from 30% to 80%, and 10% to 120% in worst-case scenarios, while reducing post-processing algorithm activation from 47% to 96.93%, maintaining the same accuracy.