software

Approximatecomputing and approximate circuits attracted a significant attention ofresearchers as well as industry in recent years. In order to accelerate theapproximate system design process and to support a fair benchmarking of approximationmethods, we propose EvoApproxLib - a comprehensivecollection of low-level approximate implementations of elementary arithmeticoperations.  The collection includes hundredsof implementations of 8-, 12- and 16-bit approximate adders and multipliersshowing unique trade-offs between several error metrics (such as the meanabsolute error and the worst case error) and electrical parameters(such as delay, area on a chip and power consumption for a particular fabricationtechnology). Each implementation can be downloaded in various formats (C, Matlab,Verilog). These implementations can be used to create high quality approximatecircuits or low-level software routines for approximate computing systems. EvoApproxLibwas automatically generated using an automated circuit design system (based on CartesianGenetic Programming) which is the original invention of the authors.EvoApproxLib extends the authors' results reported mainly in their paper whichwas awarded with the Best IP Award at DATE 2017. 

Approximate circuits, arithmetic circuits, low-power, accuracy

31. 10. 2019

Implementace a všechny dodatečné informace se nacházejí na https://ehw.fit.vutbr.cz/evoapproxlib/ a na https://github.com/ehw-fit/evoapproxlib/

K využití výsledku jiným subjektem je vždy nutné nabytí licence

Poskytovatel licence na výsledek nepožaduje licenční poplatek

https://www.fit.vut.cz/research/product/632/