GPU-Initiated Discrete Simulated Bifurcation: Low-Latency Requests and Streaming Dense Couplings
Abstract
GPU-based optimization faces two communication bottlenecks: coordinating frequent requests and delivering dense models that exceed device memory. We present a discrete simulated bifurcation (dSB) architecture that addresses both through NVIDIA DOCA GPUNetIO. For resident models, a persistent service receives field updates, executes each solve within one GPU thread block, and returns the result. Exact integer coupling sums, GPU work queues, and batched transmission keep the receive--solve--reply path on the device without a dedicated CPU data-path core. In comparisons with socket-based servers using the same solver, the largest latency gains occur under concurrent load. As the offered load increases from 400 to 800 thousand requests per second, median round-trip latency rises by only 6\%. At the highest tested load, median and 99th-percentile latencies are 189 and 218~μs, compared with 288 and 609~μs for the tuned persistent CPU proxy across repeated runs. For models larger than device memory, a streaming solver retains dynamical state on the GPU and reuses incoming coupling tiles across replicas. It evaluates ten-million-variable dense binary matrices at approximately 307~Gb/s, consuming a 12.5-TB logical matrix through a 64-MiB packet buffer. Ground-state recovery on planted instances and agreement with reference executions verify the computation. Together, the two modes scale dSB to concurrent requests and dense models beyond GPU memory.
Related
- Constrained Optimization of Higher-Order Cluster-Expansion Hamiltonians for Alloys Using Simulated Bifurcation
- Mentor-Initiated Asymmetric Bidirectional Quantum Teleportation Protocol for Arbitrary Qubit States
- Two-loop QCD contribution to the quark-initiated production of a top-antitop quark pair and a Higgs boson at the LHC
- Two-photon photopolymerization directly initiated by spiropyran photochromic molecules
- Numerical Investigation of Acoustic Streaming-Induced Shear Stress in 3D-Printed Bone-Mimicking Scaffolds
BibTeX
@misc{chen2026gpuinitiated,
title = {{GPU-Initiated Discrete Simulated Bifurcation: Low-Latency Requests and Streaming Dense Couplings}},
author = {Yaocheng Chen},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.06622},
eprint = {2610.06622},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.06622},
}
RIS
TY - UNPB
TI - GPU-Initiated Discrete Simulated Bifurcation: Low-Latency Requests and Streaming Dense Couplings
AU - Yaocheng Chen
PY - 2026
DA - 2026/10/05
PB - arXiv
DO - 10.48550/arxiv.2610.06622
UR - https://arxiv.org/abs/2610.06622
AB - GPU-based optimization faces two communication bottlenecks: coordinating frequent requests and delivering dense models that exceed device memory. We present a discrete simulated bifurcation (dSB) architecture that addresses both through NVIDIA DOCA GPUNetIO. For resident models, a persistent service receives field updates, executes each solve within one GPU thread block, and returns the result. Exact integer coupling sums, GPU work queues, and batched transmission keep the receive--solve--reply path on the device without a dedicated CPU data-path core. In comparisons with socket-based servers using the same solver, the largest latency gains occur under concurrent load. As the offered load increases from 400 to 800 thousand requests per second, median round-trip latency rises by only 6\%. At the highest tested load, median and 99th-percentile latencies are 189 and 218~μs, compared with 288 and 609~μs for the tuned persistent CPU proxy across repeated runs. For models larger than device memory, a streaming solver retains dynamical state on the GPU and reuses incoming coupling tiles across replicas. It evaluates ten-million-variable dense binary matrices at approximately 307~Gb/s, consuming a 12.5-TB logical matrix through a 64-MiB packet buffer. Ground-state recovery on planted instances and agreement with reference executions verify the computation. Together, the two modes scale dSB to concurrent requests and dense models beyond GPU memory.
ER -