CacheBlend: Fast Large Language Model Serving for RAG with Cached Knowledge Fusion
Reviewed by Pith T0 review T1 audit T2 compute T3 formal T4 kernel pith:RSBPFGHHrecord.jsonopen to challenge →
read the original abstract
Large language models (LLMs) often incorporate multiple text chunks in their inputs to provide the necessary contexts. To speed up the prefill of the long LLM inputs, one can pre-compute the KV cache of a text and re-use the KV cache when the context is reused as the prefix of another LLM input. However, the reused text chunks are not always the input prefix, which makes precomputed KV caches not directly usable since they ignore the text's cross-attention with the preceding texts. Thus, the benefits of reusing KV caches remain largely unrealized. This paper tackles just one challenge: when an LLM input contains multiple text chunks, how to quickly combine their precomputed KV caches in order to achieve the same generation quality as the expensive full prefill (i.e., without reusing KV cache)? This challenge naturally arises in retrieval-augmented generation (RAG) where the input is supplemented with multiple retrieved texts as the context. We present CacheBlend, a scheme that reuses the precomputed KV caches, regardless prefix or not, and selectively recomputes the KV values of a small subset of tokens to partially update each reused KV cache. In the meantime, the small extra delay for recomputing some tokens can be pipelined with the retrieval of KV caches within the same job, allowing CacheBlend to store KV caches in slower devices with more storage capacity while retrieving them without increasing the inference delay. By comparing CacheBlend with the state-of-the-art KV cache reusing schemes on three open-source LLMs of various sizes and four popular benchmark datasets of different tasks, we show that CacheBlend reduces time-to-first-token (TTFT) by 2.2-3.3x and increases the inference throughput by 2.8-5x from full KV recompute without compromising generation quality. The code is available at https://github.com/LMCache/LMCache.
This paper has not been read by Pith yet.
Forward citations
Cited by 13 Pith papers
-
Leyline: KV Cache Directives for Agentic Inference
Leyline adds a policy-directed KV cache edit primitive with closed-form RoPE correction for agentic inference, reporting +11.2 pp cache-hit lift and +14.3 pp solve-rate gain.
-
Sparse Prefix Caching for Hybrid and Recurrent LLM Serving
Sparse prefix caching via dynamic programming for optimal checkpoint placement under overlap distributions improves the Pareto frontier for recurrent and hybrid LLM serving on shared-prefix data.
-
MIST: A Co-Design Framework for Heterogeneous, Multi-Stage LLM Inference
MIST is a new simulator for heterogeneous multi-stage LLM inference that combines hardware traces with analytical models to explore configuration trade-offs in hybrid CPU-accelerator systems.
-
Semantic Integrity Matters: Benchmarking and Preserving High-Density Reasoning in KV Cache Compression
KV cache compression causes task-dependent degradation in high-density reasoning due to disrupted CoT links; ShotKV mitigates this by preserving few-shot examples as indivisible semantic units through phase separation...
-
Can I Buy Your KV Cache?
Proposes an agent-native prefill CDN where precomputed KV caches are hosted and sold to agents, delivering 9-50x compute savings with exact token and logit matching on Qwen3-4B.
-
SpectrumKV: Per-Token Mixed-Precision KV Cache Transfer for Prefill-Decode Disaggregated LLM Serving
SpectrumKV applies per-token mixed-precision KV cache transfer (FP16/INT8/INT4) with a model-specific probe for INT4 tolerance, achieving better perplexity and retrieval than PDTrim at equivalent budgets on Qwen2.5-7B...
-
Grounded Cache Routing for Retrieval-Augmented Generation: When Is It Safe to Reuse an Answer?
GroundedCache reduces unsafe-served rate in RAG answer caching to 0-1.5% (vs 15-51.5% naive) via four validation gates while keeping p50 latency within 1.07x of no-cache baseline.
-
CacheClip: Accelerating RAG with Effective KV Cache Reuse
CacheClip accelerates RAG prefill by up to 3.33x via auxiliary-model-guided selective KV recomputation while retaining 85-91% of full-attention quality on NIAH and LongBench.
-
MiniPIC: Flexible Position-Independent Caching in <100LOC
MiniPIC enables multiple position-independent caching methods inside vLLM via unrotated KV storage, per-request RoPE application, and three primitives, delivering 49% prefill throughput gains and up to 100x lower cach...
-
SIFT: Selective-Index For Fast Compute of RAG Prefill by Exploiting Attention Invariance
SIFT precomputes selective attention indices via local and cross-attention invariance to speed RAG prefill 1.71x while keeping accuracy within 1% of full recompute, storing only bit vectors 24,000x smaller than KV tensors.
-
Irminsul: MLA-Native Position-Independent Caching for Agentic LLM Serving
Irminsul recovers up to 83% of prompt tokens above exact-prefix matching and delivers 63% prefill energy savings per cache hit on MLA-MoE models by content-hashing CDC chunks and applying closed-form kr correction.
-
VLMaxxing through FrameMogging Training-Free Anti-Recomputation for Video Vision-Language Models
Training-free adaptive reuse of stable visual state in video VLMs reduces follow-up latency by 15-36x on Qwen2.5-VL while preserving correctness on VideoMME, with smaller first-query speedups via pruning.
-
HieraSparse: Hierarchical Semi-Structured Sparse KV Attention
HieraSparse delivers a hierarchical semi-structured sparse KV attention system that achieves 1.2x KV compression and 4.57x decode attention speedup versus prior unstructured sparsity methods at equivalent sparsity, pl...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.