{"paper":{"title":"NeuroRing: Scaling Spiking Neural Networks via Multi-FPGA Bidirectional Ring Topologies and Stream-Dataflow Architectures","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A bidirectional ring topology and stream-dataflow architecture on FPGAs enables scalable faster-than-real-time execution of large spiking neural networks while preserving activity statistics.","cross_cats":["cs.DC","cs.NE"],"primary_cat":"cs.AR","authors_text":"Artur Podobas, Muhammad Ihsan Al Hafiz","submitted_at":"2026-04-30T16:04:26Z","abstract_excerpt":"Spiking neural networks (SNNs) are a promising paradigm for energy-efficient event-driven computation, but large-scale SNN execution remains challenging because sparse spike communication and synchronization can dominate runtime. Existing solutions across CPU, GPU, ASIC, and FPGA platforms offer different trade-offs between programmability, efficiency, and scalability. To address this gap, we present NeuroRing, a modular and scalable SNN accelerator based on a stream-dataflow architecture and a bidirectional ring topology, implemented in High-Level Synthesis (HLS) on FPGAs. NeuroRing supports "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"NeuroRing preserves the key activity statistics of the NEST reference model, achieves faster-than-real-time execution of the full-scale cortical microcircuit with a real-time factor (RTF) of 0.83, exhibits meaningful strong and weak scaling, and provides competitive energy efficiency on two programmable FPGAs.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the bidirectional ring topology and stream-dataflow architecture will continue to avoid communication bottlenecks and synchronization overhead when scaled beyond the evaluated two-FPGA configuration or to networks with different spike sparsity patterns.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"NeuroRing delivers a modular multi-FPGA accelerator for spiking neural networks that achieves real-time factor 0.83 on the full cortical microcircuit while preserving NEST activity statistics and showing scaling behavior.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A bidirectional ring topology and stream-dataflow architecture on FPGAs enables scalable faster-than-real-time execution of large spiking neural networks while preserving activity statistics.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f567976e5d03b2f5848b8234fc3266eab52246fe195f2dc8b2228761c7dee70a"},"source":{"id":"2604.28059","kind":"arxiv","version":2},"verdict":{"id":"4dfc6176-102a-494f-a3e9-0c54d2691442","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-07T06:13:08.794098Z","strongest_claim":"NeuroRing preserves the key activity statistics of the NEST reference model, achieves faster-than-real-time execution of the full-scale cortical microcircuit with a real-time factor (RTF) of 0.83, exhibits meaningful strong and weak scaling, and provides competitive energy efficiency on two programmable FPGAs.","one_line_summary":"NeuroRing delivers a modular multi-FPGA accelerator for spiking neural networks that achieves real-time factor 0.83 on the full cortical microcircuit while preserving NEST activity statistics and showing scaling behavior.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the bidirectional ring topology and stream-dataflow architecture will continue to avoid communication bottlenecks and synchronization overhead when scaled beyond the evaluated two-FPGA configuration or to networks with different spike sparsity patterns.","pith_extraction_headline":"A bidirectional ring topology and stream-dataflow architecture on FPGAs enables scalable faster-than-real-time execution of large spiking neural networks while preserving activity statistics."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.28059/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T20:41:49.167808Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T18:38:29.755547Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"bca8c97db538a947e92cfc7fce05ea93467aed11173544129fae75fda6c82f26"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}