The KV Cache Is the New Memory Wall

arXiv:2609.30854 · cs.DC, cs.LG, cs.PF · Submitted 2026-09-25 · Read on arXiv

cs.DC, cs.LG, cs.PF

Submitted: 2026-09-25

Updated: 2026-09-25

Comments: 28 pages, 12 figures, 12 tables

Code: https://github.com/LMCache/LMCache

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Decode — 74× in this paper · explained in 2 episodes
• KV cache — 53× in this paper · explained in 3 episodes
• Quantization — 30× in this paper · explained in 12 episodes
• Benchmark — 15× in this paper · explained in 2 episodes
• Transformer — 8× in this paper · explained in 2 episodes
• Core — 6× in this paper · explained in 2 episodes
• Perplexity — 5× in this paper · explained in 3 episodes
• Accuracy — 3× in this paper · explained in 1 episode
• Embeddings — 3× in this paper · explained in 2 episodes
• Trace — 3× in this paper · explained in 2 episodes
• Conformal Prediction — 1× in this paper · explained in 5 episodes
• Consistency — 1× in this paper · explained in 3 episodes
• Generative Model — 1× in this paper · explained in 2 episodes
• Generative Models — 1× in this paper · explained in 2 episodes
• High Bandwidth Memory — 1× in this paper
• RAG — 1× in this paper · explained in 3 episodes
• Sparse Attention — 1× in this paper · explained in 5 episodes
• Sparsity — 1× in this paper · explained in 4 episodes
• Transformer Architecture — 1× in this paper · explained in 2 episodes
• Transformer Models — 1× in this paper · explained in 2 episodes

Sources

• The Llama 3 Herd of Models
• Optimizing AI Inference Across the Deployment Stack
• Fast Transformer Decoding: One Write-Head is All You Need
• DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
• KV Cache Optimization Strategies for Scalable and Efficient LLM Inference
• Llama 2: Open Foundation and Fine-Tuned Chat Models
• Mixtral of Experts
• PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funneling
• AI Benchmark Democratization and Carpentry

Related papers

• iScheduler: Reinforcement Learning-Driven Continual Optimization for Large-Scale Resource Investment Problems
• SAMM: Sharded Automated Market Maker
• InferScale: GPU-Native KV Injection for Personalized LLM Serving
• Vigil: Accountable Liveness against Selective Silence
• Steelhead: Interleaving Partially Synchronous and Asynchronous Commit Rules on a Shared DAG
• Lightweight Scheduling Policies for Continuously Batched LLM Serving Across GPU Architectures