SPL to HP C/XL Migration Guide (30231-90001)
3- 3
CAUTION The numeric ranges AND the data storage formats for SPL and HP
C/XL 32-bit floating-point data are significantly different. If
your application uses REAL floating-point data that depend on
extreme values, bit manipulation, or file storage, you may have
a problem in migrating to HP C/XL.
However, floating-point values may be translated from MPE V
format to MPE XL format and back with the MPE XL HPFPCONVERT
intrinsic. See the
MPE XL Intrinsics Reference Manual
for
details.
LONG Format
Table 3-5. LONG Format
---------------------------------------------------------------------------------------------
| | |
| SPL | HP C/XL Equivalent |
| | |
---------------------------------------------------------------------------------------------
| | |
| Type: LONG | Type: double |
| | |
---------------------------------------------------------------------------------------------
| | |
| 64 bits (four words) in MPE V | 64 bits (two words) in IEEE floating-point |
| floating-point format: | format: |
| 1-bit sign, 9-bit exponent, 54-bit | 1-bit sign, 11-bit exponent, 52-bit |
| fraction. | fraction. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Approximate nonzero range: | Approximate nonzero range: |
| 8.63618555094445x10-78 | 2.2250738585072014x10-308 |
| to 1.157920892373162x1077 | to 1.79769313486231x10308 |
| | |
---------------------------------------------------------------------------------------------
CAUTION The numeric ranges AND the data storage formats for SPL and HP
C/XL 64-bit floating-point data are significantly different. If
your application uses LONG floating-point data that depend on
bit manipulation or file storage, you may have a problem in
migrating to HP C/XL.
However, floating-point values may be translated from MPE V
format to MPE XL format and back with the MPE XL HPFPCONVERT
intrinsic. See the
MPE XL Intrinsics Reference Manual
for
details.