/FLOAT=[option] Specifies the floating-point data format to be used in memory for single and double precision data items. You can use the /FLOAT qualifier at compile time to specify either VAX F_floating or IEEE S_floating formats for single precision data items or VAX D_floating or VAX G_floating or IEEE T_floating formats for double precision data items. You cannot mix VAX and IEEE formats in the same compilation unit. You can select one of the following options: D_FLOAT Specifies that the memory format for COMP-1 data is VAX F_floating and for COMP-2 data is VAX D_floating. The default is D_FLOAT on OpenVMS Alpha. Note that D_floating on Alpha and I64 is not 100% compatible with D_floating on VAX. G_FLOAT Specifies that the memory format for COMP-1 data is VAX F_floating and for COMP-2 data is VAX G_floating. Note that G_floating on I64 is not 100% compatible with G_floating on Alpha and VAX. IEEE_FLOAT Specifies that the memory format for COMP-1 data is IEEE S_floating and for COMP-2 data is IEEE T_floating. The default is IEEE_FLOAT on OpenVMS I64. The IEEE standard for binary floating-point arithmetic, ANSI/IEEE 754-1985, defines four floating-point formats in two groups, basic and extended, each group having two widths, single and double. The Alpha architecture supports the basic single and double formats. Refer to the Alpha Architecture Reference Manual for more information about using floating-point data types with the Alpha architecture. Because the Alpha and I64 architectures are IEEE-compliant, you can run existing COBOL programs containing IEEE floating-point data formats with VSI COBOL. Specifying /FLOAT without an option is equivalent to specifying /FLOAT=D_FLOAT on OpenVMS Alpha and /FLOAT=IEEE_FLOAT on OpenVMS I64.