All Topics
All Topics
Technology
Technology
AI
AI
Business
Business
Entertainment
Entertainment
News
News
Programming
Programming
Science
Science
Design
Design
Environment
Environment
Finance
Finance
Crypto
Crypto
Politics
Politics
Sports
Sports
Education
Education
Gaming
Gaming
Art
Art
Music
Music
Health
Health
Security
Security
Books
Books
Food
Food
Travel
Travel
Personal
Personal
Bluesky
Twitter

Shift-and-Add (SAdd): Hardware-Aware Channel Characterization With Polar NRZ Stimuli

1mo ago

Source

IEEEShift-and-Add (SAdd): Hardware-Aware Channel Characterization With Polar NRZ Stimuliieee.org
Snippet from the RSS feed
High-bandwidth wireline transceivers rely on stochastic adaptation algorithms, such as the Least Mean Squares (LMS) method for equalizer calibration. While effective, these methods cannot provide upper-bounded latency guarantees, require tuning step-size parameters, and can be burdened by error propagation when calibrating Decision Feedback Equalizers (DFEs). However, if a high-fidelity estimate of the channel response is known, an alternative approach is to configure an equalizer to directly compensate for channel-induced distortion. To address the challenges of low-latency adaptation in wireline transceivers, we co-design a novel hardware-aware algorithm and a digital architecture for training-based channel response identification. We define and leverage a subset of polar Non Return to Zero (NRZ) stimulus that we refer to as Shift-and-Add (SAdd) sequences, which are designed to simultaneously favor an efficient hardware design and provide low noise amplification with precisely defined estimator variance properties. For deployment in practical resource-constrained wireline transceivers, we present an algorithm and architecture with low latency, upper-bounded computation time, low memory usage, and few arithmetic operations.

You might also wanna read

FPGA-Based Hardware Architecture Enables Real-Time Decoding of Quantum LDPC Codes

Researchers from IMDEA Software Institute, Nokia Bell Labs, Complutense University of Madrid, Aalto University, and Quobly have developed an

quantumcomputingreport.com·13d ago

How AI Is Transforming the Complex Art of RFIC Chip Design

The article explores how artificial intelligence is being applied to the complex field of Radio Frequency Integrated Circuit (RFIC) design —

spectrum.ieee.org·14d ago

Binary normalized neural networks achieve 32x memory reduction with single-bit parameters while maintaining performance

This paper introduces binary normalized neural network layers where all parameters (kernel weights and biases) are constrained to single-bit

arxiv.org·9mo ago

Quandela Validates Low-Latency Photonic QPU Integration with NVIDIA Infrastructure via NVQLink

French photonic quantum computing company Quandela has experimentally validated a low-latency hardware integration path connecting its photo

iq.fp2.dev·14d ago

NeuEdge: A Neuromorphic Computing Framework for Energy-Efficient Edge AI with Adaptive Spiking Neural Networks

This paper presents NeuEdge, a framework for energy-efficient neuromorphic computing on edge devices using adaptive spiking neural networks

arxiv.org·11d ago

Chinese Researchers Develop Analog Chip Claimed to Be 1,000 Times Faster Than Nvidia GPUs for Specific Applications

Chinese researchers from Peking University have developed a new analog chip using resistive random-access memory (RRAM) technology that repo

livescience.com·8mo ago

Comments

Sign in to join the conversation.

No comments yet. Be the first.