NAME

fpa − direct interface to floating point accelerator

DESCRIPTION

This library is no longer required for IBM/4.3 floating point support.  However, it is provided in this release to provide compatibility with previous releases and to enable an orderly conversion to the new support.  Users are encouraged to recompile all floating point programs.  IBM/4.3 floating point support has been enhanced so that it will use the fastest floating point hardware available (unless use of the FPA environment variable specifies otherwise).  For more information, see the two articles mentioned below. 

For those systems where users have many Makefiles, scripts, and so forth, that depend on the -lfpa flag, the system administrator can install a dummy library to satisfy the loader.  A dummy library is provided for this purpose in /usr/src/old/fpa. 

SEE ALSO

“IBM/4.3 Linkage Convention” in Volume II, Supplementary Documents
“Floating Point Arithmetic” in Volume II, Supplementary Documents

PRPQs 5799-WZQ/5799-PFF: IBM/4.3  —  July 1987