flash-comm: Self-Synchronizing Collective Communication for LLM Decoding
cs.DC
Submitted: 2026-11-04
Updated: 2026-11-04
Code: https://github.com/sgl-project/sglang
Terminology
Sources
- Demystifying NCCL: An In-depth Analysis of GPU Communication Protocols and Algorithms
- Every Microsecond Matters: Achieving Near Speed-of-Light Latency in GPU Collectives
- SiFAR: Synchronization-Free All-Reduce for Low-Latency LLM Inference
- The Cost of Synchronization in GPU Scale-Up Domains
- TokenWeave: Efficient Compute-Communication Overlap for Distributed LLM Inference
- Comet: Fine-grained Computation-communication Overlapping for Mixture-of-Experts
- Fine-grained Computation-Communication Overlap via Tile-level Signaling and Scheduling for Mixture-of-Experts
- Domino: Eliminating Communication in LLM Training via Generic Tensor Slicing and Overlapping
- X-Stage: Modeling Post-Issue Backpressure in GPU Communication--Computation Fusion
- Collective Communication for 100k+ GPUs
- Understanding and Improving Communication Performance in Multi-node LLM Inference
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