INF: SQL Server Does Use Floating Point Math Coprocessor

  - Microsoft SQL Server version 4.2 for OS/2

SQL Server uses a special version of the emulator floating point package. This means that SQL Server will use the math coprocessor if it is present. SQL Server will emulate the math coprocessor if it is not present.

SQL Server only uses the math coprocessor for floating point operations ( contain express references to columns of the datatype FLOAT). All LOG, SIN, TAN, EXP, and co-related functions will utilize floating point math and take advantage of the math coprocessor.

By contrast, datatype MONEY is stored as a "double-precision" INT (8 bytes), not as a FLOAT (8 bytes). If you omit the dollar sign preceding a MONEY value, SQL Server will assume a default of FLOAT.

Descriptions of the math functions that perform calculations in floating-point math can be found in the "Mathematical Functions" section of the "Microsoft SQL Server Learning Transact-SQL" manual.