Proceedings ArticleDOI
Accelerating Discrete Fourier Transforms with dot-product engine
Miao Hu,John Paul Strachan +1 more
- pp 1-5
Reads0
Chats0
TLDR
This paper presents a solution of computing DFT using the dot-product engine (DPE) - a one transistor one memristor (1T1M) crossbar array with hybrid peripheral circuit support and the computing complexity is reduced to a constant O(λ) independent of the input data size.Abstract:
Discrete Fourier Transforms (DFT) are extremely useful in signal processing. Usually they are computed with the Fast Fourier Transform (FFT) method as it reduces the computing complexity from O(N2) to O(Nlog(N)). However, FFT is still not powerful enough for many real-time tasks which have stringent requirements on throughput, energy efficiency and cost, such as Internet of Things (IoT). In this paper, we present a solution of computing DFT using the dot-product engine (DPE) - a one transistor one memristor (1T1M) crossbar array with hybrid peripheral circuit support. With this solution, the computing complexity is further reduced to a constant O(λ) independent of the input data size, where λ is the timing ratio of one DPE operation comparing to one real multiplication operation in digital systems.read more
Citations
More filters
Proceedings ArticleDOI
Large Memristor Crossbars for Analog Computing
Can Li,Yunning Li,Hao Jiang,Wenhao Song,Peng Lin,Zhongrui Wang,Jianhua Yang,Qiangfei Xia,Miao Hu,Eric Montgomery,Jiaming Zhang,Noraica Davila,Catherine Graves,Zhiyong Li,John Paul Strachan,R. Stanley Williams,Ning Ge,Mark Barnell,Qing Wu +18 more
TL;DR: The recent process in analog computing using analog-voltage-amplitude-vector input and analog-memristor-conductance matrix, with applications in signal and image processing is reported.
Proceedings ArticleDOI
Learning Accuracy Analysis of Memristor-based Nonlinear Computing Module on Long Short-term Memory
TL;DR: To accelerate the training efficiency of neural network-based machine learning, a memristor-based nonlinear computing module is designed and analyzed and can potentially realize a monotonic nonlinear function by successively placing memristors in a series combing with a simple amplifier.
Proceedings ArticleDOI
Computational restructuring: rethinking image processing using memristor crossbar arrays
TL;DR: The reconstruction of the 2D DCT is proposed to restructure into an equivalent single linear transformation (or MVM operation), which eliminates the series computation and reduces the processed block sizes, resulting in both the robustness to errors and the image quality is improved.
Hybrid Analog-Digital Co-Processing for Scientific Computation
TL;DR: This research presents a novel approach called Hybrid Analog-Digital Co-Processing for Scientific Computation, which automates the very labor-intensive and therefore time-heavy and expensive process of analog-digital co-processing.
Proceedings ArticleDOI
Temporal and SFQ pulse-streams encoding for area-efficient superconducting accelerators
TL;DR: This work mitigates the stringent area constraints of superconducting technology by proposing a wave-pipelined Unary SFQ (U-SFQ) architecture that leverages the advantages of two data representations: pulse-streams and Race Logic.
References
More filters
Book
Discrete Cosine Transform: Algorithms, Advantages, Applications
TL;DR: This paper presents two Dimensional DCT Algorithms and their relations to the Karhunen-Loeve Transform, and some applications of the DCT, which demonstrate the ability of these algorithms to solve the discrete cosine transform problem.
Journal ArticleDOI
ISAAC: a convolutional neural network accelerator with in-situ analog arithmetic in crossbars
Ali Shafiee,Anirban Nag,Naveen Muralimanohar,Rajeev Balasubramonian,John Paul Strachan,Miao Hu,R. Stanley Williams,Vivek Srikumar +7 more
TL;DR: This work explores an in-situ processing approach, where memristor crossbar arrays not only store input weights, but are also used to perform dot-product operations in an analog manner.
Book
Computational Frameworks for the Fast Fourier Transform
TL;DR: The Radix-2 Frameworks, a collection of general and high performance FFTs designed to solve the multi-Dimensional FFT problem of Prime Factor and Convolution, are presented.
Proceedings ArticleDOI
Dot-product engine for neuromorphic computing: programming 1T1M crossbar to accelerate matrix-vector multiplication
Miao Hu,John Paul Strachan,Zhiyong Li,Emmanuelle J. Merced Grafals,Noraica Davila,Catherine Graves,Si-Ty Lam,Ning Ge,Jianhua Yang,R. Stanley Williams +9 more
TL;DR: The Dot-Product Engine (DPE) is developed as a high density, high power efficiency accelerator for approximate matrix-vector multiplication, invented a conversion algorithm to map arbitrary matrix values appropriately to memristor conductances in a realistic crossbar array.
Journal ArticleDOI
Gauss and the history of the fast fourier transform
TL;DR: The algorithm developed by Cooley and Tukey clearly had its roots in, though perhaps not a direct influence from, the early twentieth century, and remains the most Widely used method of computing Fourier transforms.