Some implementations of long double have an explicit leading bit in the significand. On such platforms it's possible to have unnormalized numbers (that is, numbers whose leading significant bit is 0) throughout the range of finite values. The CommonPoint system never creates such values, so you can construct them only by using integer operations to form the unnormal floating-point bit patterns. It's discouraged, however. On X86 platforms, unnormalized representations raise the invalid flag when they appear in operations.
[Contents]
[Previous]
[Next]
Click the icon to mail questions or corrections about this material to Taligent personnel.
Copyright©1995 Taligent,Inc. All rights reserved.
Generated with WebMaker