Searcharxiv⌕ Search

arXiv subjects

Tristan Müller

Publications and source records attributed to Tristan Müller.

5 recordsLinked to original sources

Scaling Pêlagos: Multi-GPU Orchestration for Relay Belief-Propagation Decoding

Fault-tolerant quantum computing requires quantum error-correction (QEC) decoders that achieve low logical error rates under real-time operating constraints. These constraints motivate heuristic decoding algorithms whose error-correction performance must be evaluated through large-scale simulation. High-throughput decoding accelerates this evaluation and the exploration of decoder parameters. Using multiple graphics processing units (GPUs) is a natural way to scale these simulations, but naively distributing syndrome batches introduces host-side preprocessing and output-collection overheads that limit scaling. To address these limitations, we extend the Pêlagos relay belief-propagation (Relay-BP) decoder with process-isolated orchestration, parallel host preprocessing, and shared-memory output. We evaluate both one-gross ($[[144,12,12]]$) and two-gross ($[[288,12,18]]$) codes in a multi-GPU environment. Specifically, for the one-gross circuit at $R=50$, eight AMD Instinct MI300X GPUs and eight AMD Instinct MI355X GPUs achieve optimized end-to-end speedups of $5.68\times$ and $6.23\times$, respectively. For the two-gross circuit, the corresponding speedups are $6.92\times$ and $7.01\times$. When throughput is calculated using GPU decoding time, the corresponding two-gross speedups are $7.58\times$ and $7.45\times$, indicating near-linear GPU scaling. We deploy Pêlagos on Lux, Oak Ridge National Laboratory's AMD Instinct MI355X-based supercomputer, as a pathfinding application for high-throughput QEC simulation. This deployment provides a concrete quantum-computing use case for the AI + HPC infrastructure supporting the Genesis Mission's scientific-discovery goals. These end-to-end measurements demonstrate a practical path to accelerating large-scale QEC simulation through efficient multi-GPU decoding.

quant-ph↗

Real-time decoding of the gross code memory with FPGAs

We introduce a prototype FPGA decoder implementing the recently discovered Relay-BP algorithm and targeting memory experiments on the $[[144,12,12]]$ bivariate bicycle quantum low-density parity check code. The decoder is both fast and accurate, achieving a belief propagation iteration time of 24ns. It matches the logical error performance of a floating-point implementation despite using reduced precision arithmetic. This speed is sufficient for an average per cycle decoding time under $1\,\mathrm{μs}$ assuming circuit model error probabilities are less than $3 \times 10^{-3}$. This prototype decoder offers useful insights on the path toward decoding solutions for scalable fault-tolerant quantum computers.

quant-ph↗

Improved belief propagation is sufficient for real-time decoding of quantum memory

We introduce a new heuristic decoder, Relay-BP, targeting real-time quantum circuit decoding for large-scale quantum computers. Relay-BP achieves high accuracy across circuit-noise decoding problems: significantly outperforming BP+OSD+CS-10 for bivariate-bicycle codes and comparable to min-weight-matching for surface codes. As a lightweight message-passing decoder, Relay-BP is inherently parallel, enabling rapid low-footprint decoding with FPGA or ASIC real-time implementations, similar to standard BP. A core aspect of our decoder is its enhancement of the standard BP algorithm by incorporating disordered memory strengths. This dampens oscillations and breaks symmetries that trap traditional BP algorithms. By dynamically adjusting memory strengths in a relay approach, Relay-BP can consecutively encounter multiple valid corrections to improve decoding accuracy. We observe that a problem-dependent distribution of memory strengths that includes negative values is indispensable for good performance.

quant-ph↗

Extending solid-state calculations to ultra long-range length scales

We present a method which enables solid-state density functional theory calculations to be applied to systems of almost unlimited size. Computations of physical effects up to the micron length scale but which nevertheless depend on the microscopic details of the electronic structure, are made possible. Our approach is based on a generalization of the Bloch state which involves an additional sum over a finer grid in reciprocal space around each ${\bf k}$-point. We show that this allows for modulations in the density and magnetization of arbitrary length on top of a lattice-periodic solution. Based on this, we derive a set of ultra long-range Kohn-Sham equations. We demonstrate our method with a sample calculation of bulk LiF subjected to an arbitrary external potential containing nearly 3500 atoms. We also confirm the accuracy of the method by comparing the spin density wave state of bcc Cr against a direct supercell calculation starting from a random magnetization density. Furthermore, the spin spiral state of $γ$-Fe is correctly reproduced and the screening by the density of a saw-tooth potential over 20 unit cells of silicon is verified.

cond-mat.mtrl-sci↗

The Intrinsic Magnetization of Antiferromagnetic Textures

Antiferromagnets (AFMs) exhibit intrinsic magnetization when the order parameter spatially varies. This intrinsic spin is present even at equilibrium and can be interpreted as a twisting of the homogeneous AFM into a state with a finite spin. Because magnetic moments couple directly to external magnetic fields, the intrinsic magnetization can alter the dynamics of antiferromagnetic textures under such influence. Starting from the discrete Heisenberg model, we derive the continuum limit of the free energy of AFMs in the exchange approximation and explicitly rederive that the spatial variation of the antiferromagnetic order parameter is associated with an intrinsic magnetization density. We calculate the magnetization profile of a domain wall and discuss how the intrinsic magnetization reacts to external forces. We show conclusively, both analytically and numerically, that a spatially inhomogeneous magnetic field can move and control the position of domain walls in AFMs. By comparing our model to a commonly used alternative parametrization procedure for the continuum fields, we show that the physical interpretations of these fields depend critically on the choice of parametrization procedure for the discrete-to-continuous transition. This can explain why a significant amount of recent studies of the dynamics of AFMs, including effective models that describe the motion of antiferromagnetic domain walls, have neglected the intrinsic spin of the textured order parameter.

cond-mat.mes-hall↗