# Communication Lower Bound in Convolution Accelerators

arXiv: https://arxiv.org/abs/1911.05662 · HTML: https://mlsanity.com/paper/1911.05662/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Nov 2019.

## Built on by (most cited first, 2 of 2)

- [NicePIM: Design Space Exploration for Processing-In-Memory DNN Accelerators with 3D-Stacked-DRAM](https://mlsanity.com/paper/2305.19041/) (2023)
- [DMC4ML: Data Movement Complexity for Machine Learning](https://mlsanity.com/paper/2312.14441/) (2023)

## Abstract (arXiv)

In current convolutional neural network (CNN) accelerators, communication (i.e., memory access) dominates the energy consumption. This work provides comprehensive analysis and methodologies to minimize the communication for CNN accelerators. For the off-chip communication, we derive the theoretical lower bound for any convolutional layer and propose a dataflow to reach the lower bound. This fundamental problem has never been solved by prior studies. The on-chip communication is minimized based on an elaborate workload and storage mapping scheme. We in addition design a communication-optimal CNN accelerator architecture. Evaluations based on the 65nm technology demonstrate that the proposed architecture nearly reaches the theoretical minimum communication in a three-level memory hierarchy and it is computation dominant. The gap between the energy efficiency of our accelerator and the theoretical best value is only 37-87%.