Rounding error study of calculator floating-point numbers arithmetic operation

To analyze the rounding error of calculator floating-point numbers arithmetic operation is the foundation of numerical calculation method’s error analysis. This paper studies roundly the basic theories of rounding error of calculator floating-point numbers arithmetic operation and demonstrates the related conclusions. It also illustrates with examples that the error margin calculated by arithmetic operation can establish more complicated calculation error margin.This paper still studies the mains of the forward error analysis and backward error analysis. By generalizing the above results, the paper estimates the rounding error of matrix basic calculation.