Operands must be Consistent with the Operation

The addu instruction performs the binary addition algorithm with whatever bit patterns it is given as operands. In the program, one operand is in IEEE format, the other in two's complement. The instruction will combine these bit patterns using the binary addition algorithm and produce garbage.

Of course, using floating point registers and floating point addition will also not work because the floating point addition require two floating point operands:


##  Another Failure
##
        .text
        .globl main
main:   l.s     $f0,af              # get a one
        l.s     $f2,ai              # get another one
        add.s   $f4,$f0, $f2        # 1 + 1 = ??

        
        .data
af:     .float  1.0
ai:     .word   1

Recall that the data movement instructions (mtc1 and mfc1) merely copy bit patterns, they don't convert formats. So they are of no help, here.

Answer:

Operands must be Consistent with the Operation

QUESTION 21: