OCCAM: Optimal Data Reuse for Convolutional Neural Networks

TL;DR

Occam introduces an optimal partitioning and pipelining approach for CNNs that significantly reduces off-chip data transfers, improving performance and energy efficiency in image recognition tasks.

Contribution

It presents a novel method for full data reuse in CNNs through optimal partitioning, dependence closure analysis, and asynchronous pipelining, enhancing efficiency.

Findings

Reduces off-chip data transfer by 21x.

Achieves over 2x performance improvement.

Improves energy efficiency by up to 33%.

Abstract

Convolutional neural networks (CNNs) are emerging as powerful tools for image processing in important commercial applications. We focus on the important problem of improving the latency of image recognition. CNNs' large data at each layer's input, filters, and output poses a memory bandwidth problem. While previous work captures only some of the enormous data reuse, full reuse implies that the initial input image and filters are read once from off chip and the final output is written once off chip without spilling the intermediate layers' data to off-chip. We propose Occam to capture full reuse via four contributions. (1) We identify the necessary condition for full reuse. (2) We identify the dependence closure as the sufficient condition to capture full reuse using the least on-chip memory. (3) Because the dependence closure is often too large to fit in on-chip memory, we propose a dynamic programming algorithm that optimally partitions a given CNN to guarantee the least off-chip traffic at the partition boundaries for a given on-chip capacity. Occam's partitions reside on different chips forming a pipeline so that a partition's filters and dependence closure remain on-chip as different images pass through (i.e., each partition incurs off-chip traffic only for its inputs and outputs). (4) because the optimal partitions may result in an unbalanced pipeline, we propose staggered asynchronous pipelines (STAP) which replicates the bottleneck stages to improve throughput by staggering the mini-batches across the replicas. Importantly, STAP achieves balanced pipelines without changing Occam's optimal partitioning. Our simulations show that Occam cuts off-chip transfers by 21x and achieves 2.06x and 1.36x better performance, and 33\% and 24\% better energy than the base case and Layer Fusion, respectively. On an FPGA implementation, Occam performs 5.1x better than the base case.