Casting from Double to Integer in DLX Processor

Question

I have some problem understanding how the casting from double to integer happens in DLX. hopefully an assembly geek can help me out!

Having the data below:

    .data   0   
    .global     a   
a:  .double     3.14159265358979    
    .global     x   
x:  .double     1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16  
    .double     17,18,19,20,21,22,23,24,25,26,27    
    .global     xtop    
xtop:   .double     28 

The assembly DLX code follows:

    ld  f2,a    
    add     r1,r0,xtop  
            ; how does r1 get the value of 224 instead of 28???!!!
loop:   ld  f0,0(r1)    
            ; load stall occurs here
    multd   f4,f0,f2    
            ; 4 FP stalls
    sd  0(r1),f4    
    sub     r1,r1,#8    
    bnez    r1,loop     
    nop         ; branch delay slot
    trap    #0  ; terminate simulation 

I know that double is 64 bit divided in 8 bytes. A r or general purpose register can only have integer/fixed points number of 32 bit in group of 4 bytes.

In binary, 28 is 11100 and 224 is 11100000.

I suspect since both of them have 11100 in common, so I guess when trying to fit a double into a R register, it takes the 11100000 and put it somehow in the 32bit register as 11100...

really confusing stuff....any help would be appreciated!

Answer 1

r1 is loaded with the address of xtop. It's just an accident that it starts with 11100. There is no "double to integer cast" in that code. It's 224 because you have 28 doubles defined before xtop, each taking 8 bytes (the pi followed by the numbers 1-27).