Floating point operation

Introduction

Floating Point Operator (Floating Point Unit, Abnormal FPU) is a structure that runs floating point operations. Generally, it is implemented using a circuit, and is applied in a computer chip. It is a major development after the integer operator, because before the floating point operation inventions, the floating point operation in the computer is simulated with an integer operation, and the efficiency is very poor. The floating-point operator will have an error, but scientific and engineering calculations still rely significantly on floating-point computing - just in programming, accuracy issues must be considered.

FPU and other functions Integration

In some modern computer architecture, in the central processor built-in FPU, floating point calculation function and SIMD (single instructions ) Calculate integration. A main example is that in Intel and AMD new X86 and X64 processors, the SSE instruction set replaces the X87 instruction set.

External FPU

Main entry: Auxiliary processor

• weitek

• 3170

• intel

• 8087

• 80287

  80387, 80387dx (i387dx)

• 80487sx (i487sx, with 80486sx use, actual The 487SX is a complete 486DX. After plugging in 486SX work)

• 80486dx grade CPU has built-in FPU

• Motorola

68881

• 68882

• 68040 CPUs above the grade built-in FPU (68LC040 and 68ec040 without fpu)

  see

  • scientific count method

  • execution unit