🛰️ Daily AI Frontier
‹ back to 2026-08-17

GraniKV: Asymmetric Granularity KV-Cache Paging for Multi-Agent Systems with Long Shared Prefix

Research Efficiency & Systems

Ranking

Overall 79
Content 95
Popularity 41

Observed public metrics from 1 member.

Representative image for GraniKV: Asymmetric Granularity KV-Cache Paging for Multi-Agent Systems with Long Shared Prefix

Merged summary

TL;DR - GraniKV improves multi-agent LLM serving by storing long shared prefixes contiguously while allocating request-specific suffixes at token granularity. It achieves up to 2.16× higher output-token throughput than a production paged-serving baseline.

  • Uses separate HOT and COLD pools for shared prefixes and per-request suffixes.
  • Dispatches each decoding step across dual backends based on compute, memory, or communication bottlenecks.
  • Delivers 1.57–2.16× throughput gains across tested Llama and Qwen configurations.
  • Sustains 1.95× throughput with heterogeneous prompts, where batch-global cascade attention falls to baseline performance.

Sources (1)

GraniKV: Asymmetric Granularity KV-Cache Paging for Multi-Agent Systems with Long Shared Prefix

arXiv cs.LG Jinhyun Jeon, Sungjoo Yoo 2026-08-16 arXiv:2608.15584
Public signals Semantic Scholar citations 0 · Semantic Scholar influential citations 0
Providers: Hugging Face · N/A OpenAlex · N/A Publisher · N/A Semantic Scholar · Citations 0 · Influential citations 0 X · N/A Fetched 2026-09-07 14:22:39.377430 UTC

TL;DR - GraniKV improves multi-agent LLM serving by storing long shared prefixes contiguously while allocating request-specific suffixes at token granularity. It achieves up to 2.16× higher output-token throughput than a production paged-serving baseline.

  • Uses separate HOT and COLD pools for shared prefixes and per-request suffixes.
  • Dispatches each decoding step across dual backends based on compute, memory, or communication bottlenecks.
  • Delivers 1.57–2.16× throughput gains across tested Llama and Qwen configurations.
  • Sustains 1.95× throughput with heterogeneous prompts, where batch-global cascade attention falls to baseline performance.
item →