# Double: Breaking the Acceleration Limit via Double Retrieval Speculative Parallelism

arXiv: https://arxiv.org/abs/2601.05524 · HTML: https://mlsanity.com/paper/2601.05524/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- Placed in Speculative decoding, one of 12 directions in Efficient LLM inference and serving.
- First posted on arXiv 9 Jan 2026.

## Placed in

- Efficient LLM inference and serving / [Speculative decoding](https://mlsanity.com/inference/speculative-decoding/)

## Built on by (most cited first, 16 of 16)

- [OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning](https://mlsanity.com/paper/2606.26790/) (2026)
- [SEED: Self-Evolving On-Policy Distillation for Agentic Reinforcement Learning](https://mlsanity.com/paper/2607.14777/) (2026)
- [ECHO: Elastic Speculative Decoding with Sparse Gating for High-Concurrency Scenarios](https://mlsanity.com/paper/2604.09603/) (2026)
- [Efficient Inference for Large Vision-Language Models: Bottlenecks, Techniques, and Prospects](https://mlsanity.com/paper/2604.05546/) (2026)
- [StreamMeCo: Long-Term Agent Memory Compression for Efficient Streaming Video Understanding](https://mlsanity.com/paper/2604.09000/) (2026)
- [FBS: Modeling Native Parallel Reading inside a Transformer](https://mlsanity.com/paper/2601.21708/) (2026)
- [Draft Less, Retrieve More: Hybrid Tree Construction for Speculative Decoding](https://mlsanity.com/paper/2605.20104/) (2026)
- [PERM: Psychology-grounded Empathetic Reward Modeling for Large Language Models](https://mlsanity.com/paper/2601.10532/) (2026)
- [See the Forest for the Trees: Loosely Speculative Decoding via Visual-Semantic Guidance for Efficient Inference of Video LLMs](https://mlsanity.com/paper/2604.05650/) (2026)
- [When Hidden States Drift: Can KV Caches Rescue Long-Range Speculative Decoding?](https://mlsanity.com/paper/2604.26412/) (2026)
- [D-cut: Adaptive Verification Depth Pruning for Batched Speculative Decoding](https://mlsanity.com/paper/2607.14647/) (2026)
- [FACT-E: Causality-Inspired Evaluation for Trustworthy Chain-of-Thought Reasoning](https://mlsanity.com/paper/2604.10693/) (2026)
- [TailSieve: Partial-Rollout-Guided Tail Routing for LLM Rollouts](https://mlsanity.com/paper/2608.22788/) (2026)
- [FlowHijack: A Dynamics-Aware Backdoor Attack on Flow-Matching Vision-Language-Action Models](https://mlsanity.com/paper/2604.09651/) (2026)
- [Semiotic logical hexagon theory for LLM logical reasoning](https://mlsanity.com/paper/2607.21933/) (2026)
- [SPECTRE: Hybrid Ordinary-Parallel Speculative Serving for Resource-Efficient LLM Inference](https://mlsanity.com/paper/2605.08151/) (2026)

## Abstract (arXiv)

Parallel Speculative Decoding (PSD) accelerates traditional Speculative Decoding (SD) by overlapping draft generation with verification. However, it remains hampered by two fundamental challenges: (1) a theoretical speedup ceiling dictated by the speed ratio between the draft and target models, and (2) high computational waste and pipeline stall due to mid-sequence token rejections of early errors. To address these limitations, we introduce textscDouble (Double Retrieval Speculative Parallelism). By bridging the gap between SD and PSD, our framework resolves the Retrieval Precision-Efficiency Dilemma through a novel synchronous mechanism. Specifically, we enable the draft model to execute iterative retrieval speculations to break the theoretical speedup limits; to alleviate rejections without rollback, the target model performs authoritative retrieval to generate multi-token guidance. textscDouble is entirely training-free and lossless. Extensive experiments demonstrate state-of-the-art speedup of 5.3× on LLaMA3.3-70B and 2.8× on Qwen3-32B, significantly outperforming the advanced method EAGLE-3 that requires extensive model training.