Comparative Analysis of Floating-Point Accuracy of IEEE 754 and Posit Standards

This article discusses the possibility of increasing the productivity and efficiency of calculations in embedded systems by using floating-point numbers of lower resolution, using a data type of the same resolution to increase the accuracy of calculations. It has been demonstrated that using numbers of type posit allows you to increase the accuracy when using the same bit depth, as well as reduce the amount of memory used when using data of the type posit lower bit.