Universal Number Library

Software · Sat Sep 17 00:00:00 EDT 2016

OSTI ID:1328681

Lloyd, G. Scott ^[1]

---

1. Lawrence Livermore National Laboratory

This floating-point arithmetic library contains a software implementation of Universal Numbers (unums) as described by John Gustafson [1]. The unum format is a superset of IEEE 754 floating point with several advantages. Computing with unums provides more accurate answers without rounding errors, underflow or overflow. In contrast to fixed-sized IEEE numbers, a variable number of bits can be used to encode unums. This all allows number with only a few significant digits or with a small dynamic range to be represented more compactly.

• Short Name / Acronym: Unum; 004970MLTPL00
• Site Accession Number: LLNL-CODE-704762
• Version: 00
• Programming Language(s): Medium: X; OS: Linux, FreeBSD, Mac OS, Cygwin
• Research Organization: Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
• Sponsoring Organization: USDOE Office of Science (SC)
• Contributing Organization: Lawrence Livermore National Laboratory
• DOE Contract Number: AC52-07NA27344
• OSTI ID: 1328681
• Country of Origin: United States