Skip to content

Latest commit

 

History

History
42 lines (26 loc) · 4.26 KB

README.md

File metadata and controls

42 lines (26 loc) · 4.26 KB

MulApprox

MulApprox - A comprehensive library of state-of-the-art approximate multipliers

Overview

The area of approximate multiplier lacks the comprehensive library of state-of-the-art work. Here, we fill this gap by presenting a collection of approximate multipliers in the form of a library. The library enables researchers in this field to compare their solutions with state-of-the-art designs. All the Verilog/CMOS implementations and behaviour models of the state-of-the-art multipliers are made available publicly, together with the software used for the error assessment and application study. To the best of our knowledge, there is no such comprehensive library of state-of-the-art approximate multipliers.

Code organisation

  • verilog/ -> Verilog source files for approximate multipliers grouped according to their class.

  • behaviour_models/ -> C source files that describe the functionality of approximate multipliers.

  • utils/OpenROAD/ -> Used constraints and configuration in OpenROAD synthesis flow

  • utils/ErrorAssessment/ -> C source files that calculate NMED and MRE

  • utils/Caffe_src/ -> Cuda files for implementing convolutional and fully connected layers with approximate multiplication in the Caffe framework

List

Paper Multiplier acronym Verilog Behaviour model
Design of Approximate Radix-4 Booth Multipliers for Error-TolerantComputing R4ABM ✔️ ✔️
Approximate Hybrid High Radix Encoding for Energy-Efficient Inexact Multipliers RAD1024 ✔️ ✔️
Hybrid Low Radix Encoding-Based Approximate Booth Multipliers HLR_BM ✔️ ✔️
On the Design of Logarithmic Multiplier Using Radix-4 Booth Encoding LOBO ✔️ ✔️
A Hybrid Radix-4 and Approximate Logarithmic Multiplier for Energy Efficient Image Processing HRALM ✔️ ✔️
Computer Multiplication and Division Using Binary Logarithms MITCHEL ✔️ ✖️
DRUM: A Dynamic Range Unbiased Multiplier for approximate applications DRUM ✔️ ✖️
RoBA Multiplier: A Rounding-Based Approximate Multiplier for High-Speed yet Energy-Efficient Digital Signal Processing (AS)ROBA ✔️ ✔️
Design and Evaluation of Approximate Logarithmic Multipliers for Low Power Error-Tolerant Applications ALM_SOA ✔️ ✔️
Efficient Mitchell’s Approximate Log Multipliers for Convolutional Neural Networks MITCHEL_TRUNC ✔️ ✔️
An Improved Logarithmic Multiplier for Energy-Efficient Neural Computing ILM_AA ✔️ ✖️
Design and Analysis of Energy-Efficient Dynamic Range Approximate Logarithmic Multipliers for Machine Learning DRALM ✔️ ✔️
A Two-Stage Operand Trimming Approximate Logarithmic Multiplier TL ✔️ ✔️