When Shards Fly: Cross-Tensor In-Flight Conversion for Low-Latency, Low-Memory Weight Updates in Disaggregated RL
cs.DC
Submitted: 2026-10-31
Updated: 2026-10-31
Terminology
Sources
- ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL
- GPU-Initiated Networking for NCCL
- AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Training
- SparseRL-Sync: Lossless Weight Synchronization with ~100x Less Communication
- DORA: A Scalable Asynchronous Reinforcement Learning System for Language Model Training
- WeightBridge: An Efficient Weight Transfer Library for Reinforcement Learning
- StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Training
- UCCL-EP: Portable Expert-Parallel Communication
- Understanding and Exploiting Weight Update Sparsity for Communication-Efficient Distributed RL
- AReaL-DTE: Sparse Policy-Weight Transfer for Online Agentic Reinforcement Learning
- Miles v0.1: Production-Level Post-Training
- Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Library
- LlamaRL: A Distributed Asynchronous Reinforcement Learning Framework for Efficient Large-scale LLM Training
- DeepSpeed-Chat: Easy, Fast and Affordable RLHF Training of ChatGPT-like Models at All Scales
- TensorHub: Scalable and Elastic Weight Transfer for LLM RL Training
- {\lambda}Scale: Enabling Fast Scaling for Serverless Large Language Model Inference
- StreamRL: Scalable, Heterogeneous, and Elastic RL for LLMs with Disaggregated Stream Generation
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