You are here: Home / Past Courses / Spring 2013 - ECPE 170 / Tutorials / MIPS Instruction Set

MIPS Instruction Set

This is a **partial list** of the available MIPS32 instructions, system calls, and assembler directives. For more MIPS instructions, refer to the Assembly Programming section on the class Resources page.

In all examples, $1, $2, $3 represent registers. For class, you should use the register names, not the corresponding register numbers.

 

Arithmetic Instructions

InstructionExampleMeaningComments
add add $1,$2,$3 $1=$2+$3
subtract sub $1,$2,$3 $1=$2-$3
add immediate addi $1,$2,100 $1=$2+100 "Immediate" means a constant number
add unsigned addu $1,$2,$3 $1=$2+$3 Values are treated as unsigned integers,
not two's complement integers
subtract unsigned subu $1,$2,$3 $1=$2-$3 Values are treated as unsigned integers,
not two's complement integers
add immediate unsigned addiu $1,$2,100 $1=$2+100 Values are treated as unsigned integers,
not two's complement integers
Multiply (without overflow) mul $1,$2,$3 $1=$2*$3 Result is only 32 bits!
Multiply mult $2,$3 $hi,$low=$2*$3 Upper 32 bits stored in special register hi
Lower 32 bits stored in special register lo
Divide div $2,$3 $hi,$low=$2/$3 Remainder stored in special register hi
Quotient stored in special register lo

 

Logical

InstructionExampleMeaningComments
and and $1,$2,$3 $1=$2&$3 Bitwise AND
or or $1,$2,$3 $1=$2|$3 Bitwise OR
and immediate andi $1,$2,100 $1=$2&100 Bitwise AND with immediate value
or immediate or $1,$2,100 $1=$2|100 Bitwise OR with immediate value
shift left logical sll $1,$2,10 $1=$2<<10 Shift left by constant number of bits
shift right logical srl $1,$2,10 $1=$2>>10 Shift right by constant number of bits

 

Data Transfer

InstructionExampleMeaningComments
load word lw $1,100($2) $1=Memory[$2+100] Copy from memory to register
store word sw $1,100($2) Memory[$2+100]=$1 Copy from register to memory
load upper immediate lui $1,100 $1=100x2^16 Load constant into upper 16 bits. 
Lower 16 bits are set to zero.
load address la $1,label $1=Address of label Pseudo-instruction (provided by assembler, not processor!)
Loads computed address of label (not its contents) into register
load immediate li $1,100 $1=100 Pseudo-instruction (provided by assembler, not processor!)
Loads immediate value into register
move from hi mfhi $2 $2=hi Copy from special register hi to general register
move from lo mflo $2 $2=lo Copy from special register lo to general register
move move $1,$2 $1=$2 Pseudo-instruction (provided by assembler, not processor!)
Copy from register to register.

Variations on load and store also exist for smaller data sizes:

  • 16-bit halfword: lh and sh
  • 8-bit byte: lb and sb

 

Conditional Branch

All conditional branch instructions compare the values in two registers together. If the comparison test is true, the branch is taken (i.e. the processor jumps to the new location). Otherwise, the processor continues on to the next instruction.

InstructionExampleMeaningComments
branch on equal beq $1,$2,100 if($1==$2) go to PC+4+100 Test if registers are equal
branch on not equal bne $1,$2,100 if($1!=$2) go to PC+4+100 Test if registers are not equal
branch on greater than bgt $1,$2,100 if($1>$2) go to PC+4+100 Pseduo-instruction
branch on greater than or equal bge $1,$2,100 if($1>=$2) go to PC+4+100 Pseduo-instruction
branch on less than blt $1,$2,100 if($1<$2) go to PC+4+100 Pseduo-instruction
branch on less than or equal ble $1,$2,100 if($1<=$2) go to PC+4+100 Pseduo-instruction

Note 1: It is much easier to use a label for the branch instructions instead of an absolute number.  For example: beq $t0, $t1, equal.  The label "equal" should be defined somewhere else in the code.

Note 2: There are many variations of the above instructions that will simplify writing programs! Consult the Resources for further instructions, particularly H&P Appendix A.

 

Comparison

InstructionExampleMeaningComments
set on less then slt $1,$2,$3 if($2<$3)$1=1;
else $1=0
Test if less than.
If true, set $1 to 1. Otherwise, set $1 to 0.
set on less then immediate slti $1,$2,100 if($2<100)$1=1;
else $1=0
Test if less than.
If true, set $1 to 1. Otherwise, set $1 to 0.

Note: There are many variations of the above instructions that will simplify writing programs! Consult the Resources for further instructions, particularly H&P Appendix A.

 

Unconditional Jump

InstructionExampleMeaningComments
jump j 1000 go to address 1000 Jump to target address
jump register jr $1 go to address stored in $1 For switch, procedure return
jump and link jal 1000 $ra=PC+4; go to address 1000 Use when making procedure call.
This saves the return address in $ra

Note: It is much easier to use a label for the jump instructions instead of an absolute number.  For example: j loop.  That label should be defined somewhere else in the code.

 

System Calls

The SPIM simulator provides a number of useful system calls.  These are simulated, and do not represent MIPS processor instructions. In a real computer, they would be implemented by the operating system and/or standard library.

System calls are used for input and output, and to exit the program.  They are initiated by the syscall instruction. In order to use this instruction, you must first supply the appropriate arguments in registers $v0, $a0-$a1, or $f12, depending on the specific call desired. (In other words, not all registers are used by all system calls). The syscall will return the result value (if any) in register $v0 (integers) or $f0 (floating-point).

Available syscall services in SPIM:

ServiceOperationCode 
(in $v0)
ArgumentsResults
print_int Print integer number (32 bit)
1
$a0 = integer to be printed None
print_float Print floating-point number (32 bit)
2
$f12 = float to be printed None
print_double Print floating-point number (64 bit)
3
$f12 = double to be printed None
print_string Print null-terminated character string
4
$a0 = address of string in memory None
read_int Read integer number from user
5
None Integer returned in $v0
read_float Read floating-point number from user
6
None Float returned in $f0
read_double Read double floating-point number from user
7
None Double returned in $f0
read_string Works the same as Standard C Library fgets() function.
8
$a0 = memory address of string input buffer
$a1 = length of string buffer (n)
None
sbrk Returns the address to a block of memory containing n additional bytes.
(Useful for dynamic memory allocation)
9
$a0 = amount address in $v0
exit Stop program from running
10
None None
print_char Print character 11 $a0 = character to be printed None
read_char Read character from user 12 None Char returned in $v0
exit2 Stops program from running and returns an integer 17 $a0 = result (integer number) None

Notes:

  • The print_string service expects the address to start a null-terminated character string. The directive .asciiz creates a null-terminated character string.
  • The read_int, read_float and read_double services read an entire line of input up to and including the newline character.
  • The read_string service has the same semantics as the C Standard Library routine fgets().
    • The programmer must first allocate a buffer to receive the string
    • The read_string service reads up to n-1 characters into a buffer and terminates the string with a null character.
    • If fewer than n-1 characters are in the current line, the service reads up to and including the newline and terminates the string with a null character.
  • There are a few additional system calls not shown above for file I/O:  open, read, write, close (with codes 13-16)

 

Assembler Directives

An assembler directive allows you to request the assembler to do something when converting your source code to binary code.

DirectiveResult
.word w1, ..., wn Store n 32-bit values in successive memory words
.half h1, ..., hn Store n 16-bit values in successive memory words
.byte b1, ..., bn Store n 8-bit values in successive memory words
.ascii str Store the ASCII string str in memory.
Strings are in double-quotes, i.e. "Computer Science"
.asciiz str Store the ASCII string str in memory and null-terminate it
Strings are in double-quotes, i.e. "Computer Science"
.space n Leave an empty n-byte region of memory for later use
.align n Align the next datum on a 2^n byte boundary.
For example, .align 2 aligns the next value on a word boundary

 

Registers

MIPS has 32 general-purpose registers that could, technically, be used in any manner the programmer desires. However, by convention, registers have been divided into groups and used for different purposes. Registers have both a number (used by the hardware) and a name (used by the assembly programmer).

This table omits special-purpose registers that will not be used in ECPE 170.

Register
Number
Register 
Name
Description
0
$zero
The value 0
2-3
$v0 - $v1
(values) from expression evaluation and function results
4-7
$a0 - $a3
(arguments) First four parameters for subroutine
8-15, 24-25
$t0 - $t9
Temporary variables
16-23
$s0 - $s7
Saved values representing final computed results
31
$ra
Return address

 

Based on: