paper-with-me

Papers

Correctness Forensics for Batch Speculative Decoding: Diagnosing the Ragged Tensor Problem

2025-10-26 · Ranran Haoran Zhang, Soumik Dey, Ashirbad Mishra, Hansi Wu, Binbin Li, Rui Zhang arxiv

Inference optimizations are routinely evaluated by throughput alone, without verifying output correctness. We conduct a forensic analysis of batch speculative decoding and find that several widely-used implementations silently produce corrupted outputs (repetitive tokens, <unk> symbols) while reporting competitive speed; failures invisible to metrics like ROUGE. We trace the root cause to the ragged tensor problem: variable token acceptance desynchronizes position IDs, attention masks, and KV-cache across a batch. We formalize the synchronization invariants (rectangular alignment and position-ID contiguity) that valid batched inference must preserve and show that maintaining them incurs superlinear alignment overhead under contiguous layouts. EQSPEC enforces the invariants without custom kernels; EXSPEC schedules same-length sequences to bypass realignment. On SpecBench across three model families, EXSPEC reaches 3 x throughput at batch size 8 with 95% exact match to standard decoding; residual divergence traces to floating-point non-determinism, not synchronization error. Code:https://github.com/eBay/spec_dec

📄 PDF Abstract BibTeX arXiv:2510.22876

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Accelerating LLM Inference with Staged Speculative Decoding

2023-08-08 · Benjamin Spector, Chris Re

Recent advances with large language models (LLM) illustrate their diverse capabilities. We propose a novel algorithm, staged speculative decoding, to accelerate LLM inference in small-batch, on-device scenarios. We addre…

BASS: Batched Attention-optimized Speculative Sampling

2024-04-24 · Haifeng Qian, Sujan Kumar Gonugondla, Sungsoo Ha, Mingyue Shang 외

Speculative decoding has emerged as a powerful method to improve latency and throughput in hosting large language models. However, most existing implementations focus on generating a single sequence. Real-world generativ…

GPUHumanEval

TETRIS: Optimal Draft Token Selection for Batch Speculative Decoding

2025-02-21 · Zhaoxuan Wu, Zijian Zhou, Arun Verma, Alok Prakash 외

We propose TETRIS, a novel method that optimizes the total throughput of batch speculative decoding in multi-request settings. Unlike existing methods that optimize for a single request or a group of requests as a whole,…

SPIRe: Boosting LLM Inference Throughput with Speculative Decoding

2025-04-08 · Sanjit Neelam, Daniel Heinlein, Vaclav Cvicek, Akshay Mishra 외

Speculative decoding (SD) has been shown to reduce the latency of autoregressive decoding (AD) by 2-3x for small batch sizes. However, increasing throughput and therefore reducing the cost per token requires decoding wit…

SpecMemo: Speculative Decoding is in Your Pocket

2025-05-16 · Selin Yıldırım, Deming Chen

Recent advancements in speculative decoding have demonstrated considerable speedup across a wide array of large language model (LLM) tasks. Speculative decoding inherently relies on sacrificing extra memory allocations t…

Large Language Model