I have following problem:
I wanted to add two numbers -1.25 and 3.25 (the result is obviously 2) in RISC-V assembly using 32-bit registers.

I assumed Q format (imaginary decimal point), so that there would be 29 integer positions (I know that is not necessary and I can have better resolution with allocating more positions to fractional part, but this is just for simplicity) and 3 fractional positions in U2 format.

So I converted -1.25 in dec to bin (U2) (wihout imaginary decimal point):

11111111111111111111111111111110.110

And 3.25:

00000000000000000000000000000011.010

And then converted it to HEX in order to store it in my program.

My code is:

.data
    newLine: .asciz "n"
    num: .word 0xFFFFFFFEC 
    num2: .word 0x000000034
.text
    li a7, 4
    la a0, newLine
    ecall
    
    la t5, num
    lw t5, 0(t5)

    la t6, num2
    lw t6, 0(t6)
    
    add t3, t5, t6
    
    li a7, 4
    la a0, newLine
    ecall
    
    li a7, 35
    mv a0, t3
    ecall

And in the end I get:

00000000000000000000000000100000

What I’ve got is 32 in dec and I’m very confused.

Because I thought that if I have 3 slots after the decimal point it means that if, at the end, I shift right the result by 3 bits I would get the integer part. But for the moment if I shift my result by 3 bits I would get:

00000000000000000000000000000100

Which is 4, not 2.

Why doesn’t it work that way?

Also I have another question:

How to print the result in RARS simulator so that I would have this result in my desired format? For example if I would add 3.33 and 1.50 I would get: 4.83. I know how to get the integer part (by shifting certain amount of bits) but don’t know how to do it for fractional part.

Thank you in advance.