English

Zipage: Maintain High Request Concurrency for LLM Reasoning through Compressed PagedAttention

Distributed, Parallel, and Cluster Computing 2026-03-11 v1 Artificial Intelligence

Abstract

With reasoning becoming the generative paradigm for large language models (LLMs), the memory bottleneck caused by KV cache during the decoding phase has become a critical factor limiting high-concurrency service. Although existing KV cache eviction methods address the memory issue, most of them are impractical for industrial-grade applications. This paper introduces Compressed PagedAttention, a method that combines token-wise KV cache eviction with PagedAttention. We propose a comprehensive scheduling strategy and support prefix caching and asynchronous compression for Compressed PagedAttention. Based on this, we have developed a high-concurrency LLM inference engine, Zipage. On large-scale mathematical reasoning tasks, Zipage achieves around 95\% of the performance of Full KV inference engines while delivering over 2.1×\times speedup.

Keywords

Cite

@article{arxiv.2603.08743,
  title  = {Zipage: Maintain High Request Concurrency for LLM Reasoning through Compressed PagedAttention},
  author = {Mengqi Liao and Lu Wang and Chaoyun Zhang and Bo Qiao and Si Qin and Qingwei Lin and Saravan Rajmohan and Dongmei Zhang and Huaiyu Wan},
  journal= {arXiv preprint arXiv:2603.08743},
  year   = {2026}
}
R2 v1 2026-07-01T11:10:52.984Z