Workload-Static Optical Circuit Switching For Scalable LLM Inference Decode
cs.DC, cs.AI, cs.AR, cs.NI
Submitted: 2026-11-03
Updated: 2026-11-03
Code: https://github.com/deepseek-ai/DeepEP
Terminology
Sources
- When Light Bends to the Collective Will: A Theory and Vision for Adaptive Photonic Scale-up Domains
- Helix Parallelism: Rethinking Sharding Strategies for Interactive Multi-Million-Token LLM Decoding
- MPK: A Compiler and Runtime for Mega-Kernelizing Tensor Programs
- DeepSeek-V3 Technical Report
- DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
- Photonic Rails in ML Datacenters
- Short-circuiting Rings for Low-Latency AllReduce
- MixNet: A Runtime Reconfigurable Optical-Electrical Fabric for Distributed Mixture-of-Experts Training
- PipeWeave: Synergizing Analytical and Learning Models for Unified GPU Performance Prediction
- MegaScale-Infer: Serving Mixture-of-Experts at Scale with Disaggregated Expert Parallelism
- Serving Large Language Models on Huawei CloudMatrix384
Related papers
- Programmable Probabilistic Computer with 1,000,000 p-bits
- BatchGen: An Architecture for Scalable and Efficient Batch Inference
- Seamless Reconfiguration for DAG-BFT (Technical Report)
- ELDR: Expert-Locality-Aware Decode Routing for PD-Disaggregated MoE Serving
- iScheduler: Reinforcement Learning-Driven Continual Optimization for Large-Scale Resource Investment Problems
- TIRx: A Unified Tile Primitive ML Compiler Across Heterogeneous Backends