03-MPIS 寄存器

1 .MIPS registers register assembly name Comment r0 $zero Always 0 r1 $at Reserved for assembler r2-r3 $v0-$v1 Stores results r4-r7 $a0-$a3 Stores arguments r8-r15 $t0-$t7 Temporaries, not saved r16-r23 $s0-$s7 Contents saved for later use r24-r25 $t8-$t9 More temporaries, not saved r26-r27 $k0-$k1 Reserved by operating system r28 $gp Global pointer r29 $sp Stack pointer r30 $fp Frame pointer r31 $ra Return address 

2 .MIPS insruction formats Instruction “add” belongs to the R-type format. opcode rs rt rd shift amt function 6 5 5 5 5 6 src src dst add $s1, $s2, $t0 will be coded as 0 18 8 17 0 32 6 5 5 5 5 6 The “function” field is an extension of the opcode, and they together determine the operation. Note that “sub” has a similar format. 

3 .Instruction “lw” (load word) belongs to I-type format. opcode rs rt address 6 5 5 16 base dst offset lw $t0, 32($s3) will be coded as 35 19 8 32 6 5 5 16 Both “lw” and “sw” (store word) belong to I-format. MIPS has (fortunately) only three different instruction formats. The operation codes determine the format. This is how the control unit interprets the instructions. 

4 .

5 . What is an Assembler? A simple piece of software Assembly Machine Assembler Language Language lw t0, 32($s3) Binary code: add $s1, $s2, $t0 Consists of 0’s and 1’s only If you know the instruction formats, then you can translate it. The machine language consists of 0’s and 1’s 

6 .Pseudo-instructions (Makes your life a bit simpler) These are simple assembly language instructions that do not have a direct machine language equivalent. During assembly, the assembler translates each pseudo- instruction into one or more machine language instructions. Example move $t0, $t1 # $t0 ← $t1 The assembler will translate it to add $t0, $zer0, $t1 We will see more of these soon. 

7 .Think about these Q1. How will you load a constant into a memory location (i.e. consider implementing x :=3)? (Need some immediate mode instructions, like li which is a pseudo-instruction) Q2. How will you implement x:= x+1 in assembly language? What do you think? Q3. Why is the load (and store too) instruction so “crooked?” Used for its flexibility, let us discuss it. Q4. How will you load a constant (say 5) into a register? (Need the immediate mode instructions, like addi) 

8 .Loading a 32-bit constant into a register The pseudo-instruction “load immediate” li $s0, 0x 003A0012 hexadecimal means “load the 32-bit constant into register $s0.” Internally it is translated into lui $s0, 42 # load upper-half immediate ori $s0, $s0, 18 # (one can also use andi) 

9 .Logical Operations Shift left (logical) sll Shift right (logical) srl Bit-by-bit AND and, andi (and immediate) opcode rs rt rd shift amt function 6 5 5 5 5 6 src src dst sll $t2, $s0, 4 means $t2 = $s0 << 4 bit position (s0 = $16, t2 = $10) 0 0 16 10 4 0 6 5 5 5 5 6 s0 = 0000 0000 0000 0000 0000 0000 0000 1001 t2 = 0000 0000 0000 0000 0000 0000 1001 0000 Why are these instructions useful?