cs.AINov 4, 2026

BlockLock: Reproducible LLM Inference with Fixed Batch Geometry

Authors: Xiuyuan Qi, Pedro Reviriego, Javier Conde, Shanshan Liu

Organizations: University of Electronic Science and Technology of China, Chengdu, Sichuan, China · Universidad Polit´ecnica de Madrid, Madrid, Community of Madrid, Spain.

Abstract

Reproducible LLM inference requires that a request produces the same output regardless of the other requests in the batch. Existing solutions pay for this in different ways: batch-invariant kernels constrain every operator, and isolating prefill with fixed-shape decode ({our CoRun-style baseline}) gives up shared prefill compute and pads decode. We present BlockLock, which keeps shared batched prefill and pins only the numerical path of the protected request: it fixes the geometry that the request sees, with {two modes: BlockLock-Rows uses fixed-width rows and retains batched decode of all requests; BlockLock-Pack uses terminal-anchored packing with an isolated dummy segment, then extracts the protected request's KV state for target-only decode. Under a stated per-deployment contract, both modes preserve the target's output bitwise when its companions change.} Across five {dense} models (Qwen3-8B/14B, Llama-3.1-8B-Instruct, and Gemma-3-4B/12B-IT) and three single-GPU deployments (NVIDIA A6000, A100, and H100), we evaluate our approach on both tinyBenchmarks and BFCL{; we also include initial tensor-parallel experiments on two NVIDIA A6000 GPUs. Both modes preserve target sequences and Top20-logit bits in the reported tests.} At equal delivered work, {BlockLock-Rows} reduces latency by 53.5--55.0% and improves throughput by 115.2--122.2% relative to vLLM Batch-Invariant; compared with batch-invariant kernels, it needs no custom kernel implementations and constrains only the protected requests.

Figures & tables

Appendix figures & tables15 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Detent: Batch-Invariant LLM Serving and RL Rollouts at Near-Stock Speed

    Nov 7, 2026Peixu Xin, Runchuan Feng, Jiazhen Guo +2LLM ServingLLM Inference

  2. Bit-Flip Vulnerability of Shared KV-Cache Blocks in LLM Serving Systems

    Apr 19, 2026Yuji Yamamoto, Satoshi MatsuuraLLM ServingAdversarial Attacks on LLMs