mlsanityData to 14 Oct 2026

Affix Cache for Diffusion Large Language Models

arXiv 2608.26140 ↗First posted 26 Jun 2026cs.CL cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Diffusion Large Language Models (DLLMs) enable non-autoregressive decoding, but efficient inference support remains immature: unlike autoregressive models, whose requests reuse a shared prefix key-value (KV) cache, DLLMs use bidirectional attention, so a shared context's KV states depend on the tokens still being decoded, leaving directly reused caches stale and full recomputation necessary. We present ACache, a cross-request cache reuse mechanism for shared spans, or affixes, at any position: prefix, infix, or suffix. ACache measures the influence of affix tokens on the masked generation region to identify a small request-specific subset as Anchor Tokens, and recomputes only their KV states while reusing the remaining affix cache. Built on state-of-the-art intra-request caching mechanisms, ACache recovers most of the accuracy lost to direct affix-cache reuse on average when recomputing around 20% of affix tokens, and at that budget preserves more accuracy than selection criteria adapted from prior cross-request cache-reuse systems. We co-design ACache with a modern inference engine, whose attention reads each request's recomputed Anchor KV states alongside one affix cache shared across concurrent requests. Against the same system with only intra-request caching, ACache cuts recompute latency by up to 56.7%, translating to as much as 1.71× end-to-end throughput, while reducing peak KV cache memory by up to 45.8%.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 87 of its 105 references that are in the map. First posted on arXiv 26 Jun 2026.

Builds on

87 of 105 references.

References · 87

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2022Training language models to follow instructions with human feedbackcited by 15,971
  3. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  4. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  5. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  6. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  7. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  8. 2022FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awarenesscited by 3,673
  9. 2021Program Synthesis with Large Language Modelscited by 3,637
  10. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,186
  11. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,719
  12. 2023Gorilla: Large Language Model Connected with Massive APIscited by 1,332
38 more
  1. 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,141
  2. 2022Diffusion-LM Improves Controllable Text Generationcited by 1,027
  3. 2025Large Language Diffusion Modelscited by 912
  4. 2024SnapKV: LLM Knows What You are Looking for Before Generationcited by 880
  5. 2024Simple and Effective Masked Diffusion Language Modelscited by 848
  6. 2023H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Modelscited by 832
  7. 2023Discrete Diffusion Modeling by Estimating the Ratios of the Data Distributioncited by 745
  8. 2024Simplified and Generalized Masked Diffusion for Discrete Datacited by 626
  9. 2021Argmax Flows and Multinomial Diffusion: Learning Categorical Distributionscited by 574
  10. 2025Dream 7B: Diffusion Large Language Modelscited by 544
  11. 2022Efficiently Scaling Transformer Inferencecited by 539
  12. 2023Scissorhands: Exploiting the Persistence of Importance Hypothesis for LLM KV Cache Compression at Test Timecited by 439
  13. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decodingcited by 420
  14. 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Modelscited by 409
  15. 2024Your Absorbing Discrete Diffusion Secretly Models the Conditional Distributions of Clean Datacited by 381
  16. 2025LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Modelscited by 285
  17. 2024Scaling Diffusion Language Models via Adaptation from Autoregressive Modelscited by 274
  18. 2023Prompt Cache: Modular Attention Reuse for Low-Latency Inferencecited by 268
  19. 2025FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Servingcited by 258
  20. 2025Mercury: Ultra-Fast Language Models Based on Diffusioncited by 235
  21. 2024CacheBlend: Fast Large Language Model Serving for RAG with Cached Knowledge Fusioncited by 223
  22. 2024BABILong: Testing the Limits of LLMs with Long Context Reasoning-in-a-Haystackcited by 215
  23. 2024Scaling up Masked Diffusion Models on Textcited by 199
  24. 2022SSD-LM: Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Controlcited by 194
  25. 2025Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inferencecited by 178
  26. 2025dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Cachingcited by 169
  27. 2025LLaDA2.0: Scaling Up Diffusion Language Models to 100Bcited by 166
  28. 2025dKV-Cache: The Cache for Diffusion Language Modelscited by 143
  29. 2025Fast-dLLM v2: Efficient Block-Diffusion LLMcited by 139
  30. 2025Accelerated Sampling from Masked Diffusion Models via Entropy Bounded Unmaskingcited by 131
  31. 2025Accelerating Diffusion LLMs via Adaptive Parallel Decodingcited by 112
  32. 2024ChunkAttention: Efficient Self-Attention with Prefix-Aware KV Cache and Two-Phase Partitioncited by 82
  33. 2025FlashDLM: Accelerating Diffusion Language Model Inference via Efficient KV Caching and Guided Diffusioncited by 77
  34. 2024Hydragen: High-Throughput LLM Inference with Shared Prefixescited by 73
  35. 2024Preble: Efficient Distributed Prompt Scheduling for LLM Servingcited by 70
  36. 2025Sparse-dLLM: Accelerating Diffusion LLMs with Dynamic Cache Evictioncited by 65
  37. 2025LLaDA-MoE: A Sparse MoE Diffusion Language Modelcited by 57
  38. 2025Diffusion Language Models are Super Data Learnerscited by 48

50 most cited of 87 All 87 (JSON)

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025WeDLM: Reconciling Diffusion Language Models with Standard Causal Attention for Fast Inference
  2. 2025ReFusion: A Diffusion Large Language Model with Parallel Autoregressive Decoding
  3. 2026Accelerating Masked Diffusion Large Language Models: A Survey of Efficient Inference Techniques
  4. 2026SPA-Cache: Singular Proxies for Adaptive Caching in Diffusion Language Models
  5. 2026Fixed State, Long Reach: What a Constant-Size Cache Buys Block Diffusion at Scale
  6. 2026Set Diffusion: Interpolating Token Orderings Between Autoregression and Diffusion for Fast and Flexible Decoding
  7. 2026GroupKV: Hierarchical KV Cache Management for Long-Context Diffusion LLM Inference
  8. 2026LaCache: Exact Caching and Precision-Adaptive Inference for Diffusion Large Language Models