Approximate Computing Extensions for the Clash HDL Compiler

Keszöcze O, Kießling M (2021)

Publication Language: English

Publication Type: Conference contribution, Original article

Publication year: 2021

Event location: virtuelle Konferenz DE

Abstract

Approximate Computing is a novel design paradigm sacrificing computational accuracy for gains in other non-functional
properties, such as area, latency, and power consumption. Many approximation techniques operate on the netlist of a
circuit. We believe that working on a higher level of abstraction is beneficial. For this, we chose the high-level hardware
description language Clash. The language is based on the functional programming language Haskell, i.e. describes
hardware in a very abstract manner allowing to focus on the functionality instead of technical details. We extend the
Clash compiler to support multiple approximation techniques. Initial experiments show promising results with respect to
reductions in area, latency, and power consumption of the synthesized approximated circuits.

Authors with CRIS profile

Oliver Keszöcze Juniorprofessur für Informatik

Related research project(s)

ACoF -- Approximate Computing on FPGAs (ACoF) June 1, 2021 - June 1, 2024 Preparation of DFG funding proposal: ACoF (ACoF) Jan. 15, 2019 - April 14, 2020

How to cite

APA:

Keszöcze, O., & Kießling, M. (2021). Approximate Computing Extensions for the Clash HDL Compiler. In Proceedings of the Workshop Methoden und Beschreibungssprachen zur Modellierung und Verifikation von Schaltungen und Systemen. virtuelle Konferenz, DE.

MLA:

Keszöcze, Oliver, and Michael Kießling. "Approximate Computing Extensions for the Clash HDL Compiler." Proceedings of the Workshop Methoden und Beschreibungssprachen zur Modellierung und Verifikation von Schaltungen und Systemen, virtuelle Konferenz 2021.

BibTeX: Download