Converting MIPS Instructions to Binary for Computer Assembly Students

[basketball853]

Hey Guys! Thank you for checking out my Thread... I am taking a computer assembly class and i have homework due on Wed but i am stuck on a single problem. I have to convert MIPS instructions into binary, I want to learn how to do this because it will be on a test... and i could just go into Mars and assemble the code and read it from there but i want to know how to do it... okay soo the instructions are:
addi $sp, $sp, -8
sw $ra, 4($sp)

Now from looking into my book i see that addi has an Opcode of 001000 and it has a I instruction format...
Also sw has an Opcode of 101011 and is an I instruction format

How do i figure out what registers $sp?, $ra, the array 4($sp) are in binary? Please help... i would really appreciate it!

 

[Mark44]

Hey Guys! Thank you for checking out my Thread... I am taking a computer assembly class and i have homework due on Wed but i am stuck on a single problem. I have to convert MIPS instructions into binary, I want to learn how to do this because it will be on a test... and i could just go into Mars and assemble the code and read it from there but i want to know how to do it... okay soo the instructions are:
addi $sp, $sp, -8
sw $ra, 4($sp)

Now from looking into my book i see that addi has an Opcode of 001000 and it has a I instruction format...
Also sw has an Opcode of 101011 and is an I instruction format

How do i figure out what registers $sp?, $ra, the array 4($sp) are in binary?

See http://en.wikipedia.org/wiki/MIPS_architecture, in the section titled Compiler register usage. $sp is the stack pointer, and its value in MIPS instructions is $29. $ra is return address; its value is $31.

4($sp) would be an address on the stack - $sp + 4*(size of a machine word).

Here's another link to a site that lists the instructions and what they look like in binary - http://www.mrc.uidaho.edu/mrc/people/jff/digital/MIPSir.html.

For your ADDI instruction, the constant to be added is -8, which will need to be written in two's complement form.

 

[basketball853]

Thank you for your reply... i have already checked those sites out... but they have not helped me at all...

So how do i go about converting it into binary code? i think i am missing the simple concept...

So i take Addi for instance and use its opcode... okay cool got that done but what about the registers?
all it says on that website is: Used to add sign-extended constants (and also to copy one register to another: addi $1, $2, 0), executes a trap on overflow
001000ss sssttttt CCCCCCCC CCCCCCCC

what the hell does that mean lol

 

[Mark44]

This -001000ss sssttttt CCCCCCCC CCCCCCCC - shows how the instruction op code and registers and operands are encoded to make up an ADDI instruction.

First 6 bits - the instruction opcode (001000 for ADDI)
Next 5 bits - the source ($2 in your example)
Next 5 bits - the destination ($1 in your example)
Next 16 bits - the immediate value

ADDI effectively does this $t = $s + imm. value

So if you wanted to add 5 plus the value in $t1 to $t2, the MIPS assembly instruction would be:
ADDI $t2, $t1, 5

Note that the numbers associated with $t0 and $t1 are 8 and 9.

The binary form of this instuction would be

001000 01001 01000 0000000000000101
opcode
...| $t1..|
...| $t0.|
.....|...5....|

 

[basketball853]

Thank you again for clarifying that...but what if there is no value for the registers i see that i have a constant of -8 which i would use two's complement but for the other $sp register is it defaulted at 0?

 

[Mark44]

Are you asking about ADDI $sp, $sp, -8?

What do you mean "but for the other $sp register is it defaulted at 0"

 

[basketball853]

Yes i am

addi $sp, $sp, -8
would that in the end turn to be (opcode) 001000 (default 0 for $sp) 00000 (default 0 for $sp) 00000 1111111111110111 (-8)

 

[Mark44]

Yes i am

addi $sp, $sp, -8
would that in the end turn to be (opcode) 001000 (default 0 for $sp) 00000 (default 0 for $sp) 00000 1111111111110111 (-8)

0 is not the number for $sp, so the parts with 00000 and 00000 are not right. This link - http://en.wikipedia.org/wiki/MIPS_architecture- tells you the numbers of all the registers. I also told you what it was in post #2.

 

[basketball853]

oohhh i think i may understand... so the value of $sp ( the stack pointer) is $29 sooo
it would be

001000 11101 11101 111111111111011
opcode $sp $sp -8 constant

did i correctly get the value?

 

[Mark44]

Looks good!

 

[basketball853]

okay SWEET! soo and for the other one i got
101011 11111 11111 how do i figure out the 4($sp) ?

you stated before "4($sp) would be an address on the stack - $sp + 4*(size of a machine word)."

so do i take 29 + 4 * ( idk what the size of word is) ?

 

[Mark44]

sw $ra, 4($sp)

What this is doing is to store the word (== 32 bits) at $sp + 4 to the $ra register.

The size of a machine word seems to be 32 bits, since that's the size of the encoded instruction. Let's go with that assumption.

According to the reference I cited, the encoded instruction is
101011 sssss ttttt iiii iiii iiii iiii.

Here, the s bits hold the source register, the t bits the destination register, and the i bits the offset from the source register. I am thinking now that the offset is in bytes, so we use the offset as-is.

So the number for $sp is $29, the number for $ra is $31, and the offset is 4. I think this is the encoded instruction for SW $RA, 4($SP)
101011 11101 11111 0000 0000 0000 0100
The first 6 bits are the SW instruction. The next 5 bits are the source register, $SP. The next 5 bits are the $RA register, and the last 16 bits are the offset, which is 4.

 

[basketball853]

Thank you so much for your help! You have really helped me understand this! I am very grateful!

 

[Mark44]

You're very welcome. Thanks are always appreciated!