Loading...
MOTOROLA M68040 USER’S MANUAL 10-29
10.7 FLOATING-POINT UNIT INSTRUCTION TIMINGS
For floating-point instructions in the MC68040, the integer pipeline passes the decoded
instruction to the floating-point unit for execution, then supports the floating-point unit by
calculating effective addresses and transferring operands to and from this unit. For these
instructions, the execution times listed in the integer unit timing section show the overhead
required by the integer unit to support the floating-point unit, assuming the floating-point
unit is not busy with the previous floating-point instructions.
Times in parentheses are the total time that that stage uses to execute an instruction even
though the stage can pass data to the next stage early. The order of operands is generally
not significant for timing purposes. Different rounding modes (i.e., round to zero, etc.)
never incur a time penalty. Instructions with an S or D (e.g., FSADD) have the same effect
as setting the rounding precision to S or D. All FMOVEM instructions wait for the pipe to
idle before starting. Refer to Section 9 Floating-Point Unit (MC68040 Only) for details
on the operation of the floating-point unit pipeline.
10.7.1 Miscellaneous Integer Unit Support Timings
Instruction Condition <ea> Calculate Execute
FBcc Taken
Not Taken
7
6
7
6
FDBcc cc True
cc False
9
11
1L + 7
1L + 9
FNOP FPU Idle 6 6
FTRAPcc Not Taken 6 1L + 5
Loading...
Terms of Use | Privacy Policy | DMCA Policy
2006-2020 Rsmanuals.com