NeuralMatrix: Compute the Entire Neural Networks with Linear Matrix Operations for Efficient Inference

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sun, Ruiqi, Ye, Siwei, Zhao, Jie, He, Xin, Lin, Jianzhe, Li, Yiran, Zou, An
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916362188750848
author Sun, Ruiqi
Ye, Siwei
Zhao, Jie
He, Xin
Lin, Jianzhe
Li, Yiran
Zou, An
author_facet Sun, Ruiqi
Ye, Siwei
Zhao, Jie
He, Xin
Lin, Jianzhe
Li, Yiran
Zou, An
contents The inherent diversity of computation types within the deep neural network (DNN) models often requires a variety of specialized units in hardware processors, which limits computational efficiency, increasing both inference latency and power consumption, especially when the hardware processor needs to support and execute different neural networks. In this study, we introduce NeuralMatrix, which elastically transforms the computations of entire DNNs into linear matrix operations. This transformation allows seamless execution of various DNN models all with matrix operations and paves the way for running versatile DNN models with a single General Matrix Multiplication (GEMM) accelerator.Extensive experiments with both CNN and transformer-based models demonstrate the potential of NeuralMatrix to accurately and efficiently execute a wide range of DNN models, achieving 2.17-38.72 times computation efficiency (i.e., throughput per power) compared to CPUs, GPUs, and SoC platforms. This level of efficiency is usually only attainable with the accelerator designed for a specific neural network.
format Preprint
id arxiv_https___arxiv_org_abs_2305_14405
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle NeuralMatrix: Compute the Entire Neural Networks with Linear Matrix Operations for Efficient Inference
Sun, Ruiqi
Ye, Siwei
Zhao, Jie
He, Xin
Lin, Jianzhe
Li, Yiran
Zou, An
Machine Learning
Artificial Intelligence
Hardware Architecture
The inherent diversity of computation types within the deep neural network (DNN) models often requires a variety of specialized units in hardware processors, which limits computational efficiency, increasing both inference latency and power consumption, especially when the hardware processor needs to support and execute different neural networks. In this study, we introduce NeuralMatrix, which elastically transforms the computations of entire DNNs into linear matrix operations. This transformation allows seamless execution of various DNN models all with matrix operations and paves the way for running versatile DNN models with a single General Matrix Multiplication (GEMM) accelerator.Extensive experiments with both CNN and transformer-based models demonstrate the potential of NeuralMatrix to accurately and efficiently execute a wide range of DNN models, achieving 2.17-38.72 times computation efficiency (i.e., throughput per power) compared to CPUs, GPUs, and SoC platforms. This level of efficiency is usually only attainable with the accelerator designed for a specific neural network.
title NeuralMatrix: Compute the Entire Neural Networks with Linear Matrix Operations for Efficient Inference
topic Machine Learning
Artificial Intelligence
Hardware Architecture
url https://arxiv.org/abs/2305.14405