一句话：提出 BlockDepend，用块级依赖分析简化算子间依赖并挖掘并行机会，显著加速 DNN 推理。
深度神经网络中的算子间优化依赖于准确的数据依赖分析，而传统机器学习编译器（MLC）在元素级与算子级做静态依赖分析，存在两个问题：复杂的依赖关系阻碍了高效的算子间优化，同时被忽略的可并行计算导致 GPU 资源利用不足。本文提出 BlockDepend，一个基于块级依赖分析的 MLC 框架。它通过考察编译过程的更低层次阶段，提取关键的块级依赖信息，从而简化算子之间的复杂关系，并揭示隐藏的并行化机会，据此实施针对性优化策略以提升访存效率与 GPU 利用率。实验表明，在多种工作负载上，BlockDepend 相比 NVIDIA TensorRT 与 AMD MIGraphX 分别取得 1.71× 和 2.88× 的加速。
Inter-operator optimization in deep neural networks (DNNs) relies on accurate data dependency analysis. Traditional machine learning compilers (MLCs) perform static data dependency analysis at the element and operator levels, leading to two key limitations: complex dependencies that hinder efficient inter-operator optimizations, and overlooked parallelizable computations that underutilize GPU resources. We introduce BlockDepend, a novel MLC framework that addresses these issues through block-level dependency analysis. By examining the lower-level phases of compilation, BlockDepend extracts crucial block-level dependency information, simplifying complex relationships between operators and uncovering hidden parallelization opportunities. This allows for targeted optimization strategies that enhance memory access efficiency and improve GPU utilization. Our experiments demonstrate BlockDepend's effectiveness, achieving speedups of 1.71× and 2.88× compared to NVIDIA TensorRT and AMD MIGraphX, respectively, across various workloads.
@inproceedings{DiWSMRHMZTS25,
title = {Optimizing Deep Learning Inference Efficiency through Block Dependency Analysis},
author = {Zhanyuan Di and Leping Wang and En Shao and Zhaojia Ma and Ziyi Ren and Feng Hua and Lixian Ma and Jie Zhao and Guangming Tan and Ninghui Sun},
booktitle = {ASPLOS},
year = {2025},
doi = {10.1145/3676641.3716264}
}