AIDA: Associative DNN Inference Accelerator

2018/12/20 by Leonid Yavits, Yavits, Leonid, Roman Kaplan +3
Computer Science · Engineering · #Advanced Memory and Neural Computing #Advanced Neural Network Applications #Distributed #FOS: Computer and information sciences #Ferroelectric and Negative Capacitance Devices #Parallel #and Cluster Computing (cs.DC) #cs.DC

paper · pdf · doi:10.48550/arxiv.1901.04976

arxiv created 2018/12/20 · openalex publication_date 2018/12/20 · arxiv updated 2019/01/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We propose AIDA, an inference engine for accelerating fully-connected (FC) layers of Deep Neural Network (DNN). AIDA is an associative in-memory processor, where the bulk of data never leaves the confines of the memory arrays, and processing is performed in-situ. AIDA area and energy efficiency strongly benefit from sparsity and lower arithmetic precision. We show that AIDA outperforms the state of art inference accelerator, EIE, by 14.5x (peak performance) and 2.5x (throughput).

Citations

• ImageNet classification with deep convolutional neural networks

Related

• EIE: Efficient Inference Engine on Compressed Deep Neural Network
• Field-Programmable Deep Neural Network (DNN) Learning and Inference accelerator: a concept
• AnalogNAS: A Neural Network Design Framework for Accurate Inference with Analog In-Memory Computing
• DLAS: An Exploration and Assessment of the Deep Learning Acceleration Stack
• Proximu: Efficiently Scaling DNN Inference in Multi-core CPUs through Near-Cache Compute