Awesome System Papers Wiki

Home

❯

proposals

文件夹: proposals

此文件夹下有5条笔记。

  • 2026年8月20日

    probes

    • 2026年6月28日

      StateBudget: Unified Weight/KV/Expert Residency for Heterogeneous Small-Cluster Multi-Model Agent Serving

      • multi-model-serving
      • heterogeneous-gpu
      • agent-serving
      • moe
      • kv-cache
      • model-switching
      • pcie
      • memory-hierarchy
    • 2026年5月06日

      Importance-Guided KV Cache Tiering: Joint Optimization of Sparse Attention Selection and Memory Placement

      • kv-cache
      • sparse-attention
      • llm-serving
      • memory-management
      • tiered-storage
      • cross-layer-optimization
    • 2026年4月30日

      When KV Cache Heuristics Break: Rethinking Tiering for Thinking-Model Inference

      • kv-cache
      • thinking-models
      • chain-of-thought
      • llm-inference
      • tiered-storage
      • page-migration
    • 2026年4月05日

      ElasticMoE: Expert-Level Elasticity for Multi-Node MoE Decode Serving via P2P RDMA

      • LLM-Serving
      • MoE
      • Expert-Parallelism
      • Elastic-Scaling
      • P2P-RDMA

    Created with Quartz v4.5.2 © 2026

    • GitHub
    • Discord Community