The display of

floating-point numbers comprises two parts the significant (or mantissa) M and the exponent E.

One of the inputs is defined as IEEE 754 double precision

floating-point number and the other is 16-bit fixed-point number.

The IEEE specification for a 32-bit

floating-point number can be depicted as

The term ULP (for Unit in the Last Place) denotes the distance between two

floating-point numbers.

x] = L - 1, then the

floating-point number is either [+ or -] 0 or the denormalized number [+ or -]0.

A

floating-point number can be converted to fixed point by shifting the binary point of the mantissa by the appropriate number of sign bits.

cI, cU: conversion from a 64-bit signed, respectively unsigned, hardware integer to a

floating-point numberIrritatingly, the IEEE standard specifies a number of roundoff and truncation algorithms that may be employed when storing a

floating-point number.

The TC1-FPU supports the IEEE-754 single-precision

floating-point number format and all defined rounding modes.

4D-20 (for a

floating-point number system with sufficient precision).

IEEE denormalized number (denormal)--A

floating-point number with magnitude between zero and the smallest representable normalized number.

We also show how to produce the

floating-point number closest to the exact result with relatively few additional operations.

This permits correct behavior in the pathological case where half of a

floating-point number is loaded in one place, and the other half is loaded in some other basic block.

2) If a

floating-point number is communicated between processors, the communication layer guarantees the exact transmittal of the floating-point value.

Types specify the format of data, such as alphanumeric characters,

floating-point numbers or more complex data structures, such as Web pages and database queries.