Metraipolis › archive › dc_f22139bf8ac8
Scaling Inference Prefill with High-Radix Photonic Interconnects — arXiv:2609.01821
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.01821
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T04:19:19.057Z
- SHA-256 of the archived text
a53e6decc4d3d1858c4d185e6c211c71ecc0f05b155fd263bb1c7c427c25c7a9- Stable URL of this page
https://metraipolis.com/documento/dc_f22139bf8ac8
How to cite
Scaling Inference Prefill with High-Radix Photonic Interconnects — arXiv:2609.01821 — arXiv (en), https://arxiv.org/abs/2609.01821 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f22139bf8ac8 (Metraipolis archive dc_f22139bf8ac8, sha256 a53e6decc4d3d1858c4d185e6c211c71ecc0f05b155fd263bb1c7c427c25c7a9)
BibTeX
@misc{dcf22139bf8ac8,
title = {Scaling Inference Prefill with High-Radix Photonic Interconnects — arXiv:2609.01821},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.01821},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f22139bf8ac8; sha256 a53e6decc4d3d1858c4d185e6c211c71ecc0f05b155fd263bb1c7c427c25c7a9},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01821. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
With the rise of inference as today's dominant AI workload, the industry is transitioning to high-bandwidth photonic interconnects to meet the large scale-up requirements of increasingly complex Mixture-of-Experts (MoE) models. This paper quantifies the benefits of 3D-integrated photonic interconnects for inference prefill by analyzing tradeoffs between high-concurrency throughput for Large Language Model (LLM) chat and the large context windows typically required for reasoning and agentic AI. We simulate three MoE models: short context (1K--8K tokens), medium context (128K tokens), and long context (1M tokens). We evaluate this workload across existing copper-based GPU systems and one with high bandwidth integrated photonics. We show 2.1--3.2x latency improvements in the stressed high-batch regimes and 2.8--5.8x improvements over baselines in communication-limited configurations. 3D photonics enable the 1152-GPU footprint required to lower time-to-first-token, yielding 2.2--4.5x speedups across production-grade platforms when electrical systems cross their inherent scale-up-pod limits. Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-03 in cs.DC). Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC) ; Hardware Architecture (cs.AR). Pagina del lavoro: https://arxiv.org/abs/2609.01821 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T03:27:00.759Z dalla pagina https://arxiv.org/catchup/cs.DC/2026-09-03?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).