The New IEEE-754 Standard for Floating Point Arithmetic

The IEEE-754 standard for Floating Point Arithmetic[1] that was in effect at the time of this seminar was adopted in 1985. That standard was intended for hardware implementation, although provisions were made for software implementation for operations. In addition to required operations, an appendix of recommended functions was also specified. Default exception handling was specified, and provisions for alternate exception handling were also provided. Lacking in the standard were means to access many of the features from higher level languages. Because of the standardization which IEEE-754 (1985) provided, it became possible to write algorithms using floating point arithmetic which could be executed on a variety of platforms and which would produce identical results. It became possible to prove statements about the behavior of floating point programs. In 2000, the IEEE chartered a new committee to examine the IEEE-754 standard with the goals of including decimal floating point arithmetic, incorporating good existing practice, providing for reproducible results, and clarifying the standard, while not invalidating conforming implementations of the IEEE-754(1985) standard.