System designers integrate the M68040FPSP into the system so that the user object code
runs unchanged and remains totally transparent to the end user. The M68040FPSP can
be installed into any operating system. It provides kernel routines to support
unimplemented instructions and unsupported data types. Unimplemented instructions for
end-user applications compiled for the MC68881/MC68882 are contained in a library for
improved performance. For all MC68040 floating-point instructions, the coprocessor ID
field must be 001. Table E-1 lists the floating-point functions implemented as instructions
by the MC68040.
Table E-1. MC68040 Floating-Point Instructions
Floating-Point Instructions
Name Description Name Description
FMOVE Move to FPx or CR FDMOVE Double-Precision Move
FSMOVE Single-Precision Move FABS Absolute Value
FCMP Compare FDABS Double-Precision Absolute Value
FSABS Single-Precision Absolute Value FNEG Negate
FTST Test FDNEG Double-Precision Negate
FSNEG Single-Precision Negate FSUB Subtract
FADD Add FMUL Multiply
FDIV Divide FScc Set According to Condition
FBcc Branch Conditionally FTRAPcc Trap Conditionally
FDBcc Test Condition, Decrement, and Branch FSSUB Single-Precision Subtract
FSADD Single-Precision Add FSDIV Single-Precision Divide
FSMUL Single-Precision Multiply FDSUB Double-Precision Subtract
FDADD Double-Precision Add FDDIV Double-Precision Divide
FDMUL Double-Precision Multiply FSSQRT Single-Precision Square Root
FSQRT Square Root FNOP No Operation
FSAVE Save Internal State FSGLMUL Single-Precision Multiply
FMOVEM Move Multiple Registers FRESTORE Restore Internal State
Terms of Use | Privacy Policy | DMCA Policy
2006-2020 Rsmanuals.com