EEM 486 EEM 486: Computer Architecture Lecture 2 MIPS Instruction Set Architecture

Lec 2.2 Assembly Language  Basic job of a CPU: execute lots of instructions  Instructions are the primitive operations that the CPU may execute  Different CPUs implement different sets of instructions  The set of instructions a particular CPU implements is an Instruction Set Architecture (ISA) Examples: Intel 80x86 (Pentium 4), IBM/Motorola PowerPC (Macintosh), MIPS, Intel IA64,...

Lec 2.3 Instruction Set Architecture (ISA) “... the attributes of a [computing] system as seen by the programmer, i.e., the conceptual structure and functional behavior, as distinct from the organization of the data flows and controls the logic design, and the physical implementation.” Amdahl, Blaaw, and Brooks, 1964

Lec 2.4 ISA instruction set software hardware

Lec 2.5 ISA

Lec 2.6 Instruction Set Architectures  Early trend was to add more and more instructions to new CPUs to do elaborate operations VAX architecture had an instruction to multiply polynomials!  RISC philosophy – Reduced Instruction Set Computing Keep the instruction set small and simple, makes it easier to build fast hardware. Let software do complicated operations by composing simpler ones.

Lec 2.7 MIPS Architecture  MIPS – semiconductor company that built one of the first commercial RISC architectures  We will study the MIPS architecture in some detail in this class  Why MIPS instead of Intel 80x86? MIPS is simple, elegant. Don’t want to get bogged down in gritty details. MIPS widely used in embedded apps, x86 little used in embedded, and more embedded computers than PCs

Lec 2.8

Lec 2.9 Assembly Variables: Registers  Unlike HLL like C or Java, assembly cannot use variables Why not? Keep hardware simple  Assembly operands are registers limited number of special locations built directly into the hardware operations can only be performed on these!  Benefit: Since registers are directly in hardware, they are very fast (faster than 1 billionth of a second)

Lec 2.10 Assembly Variables: Registers  Drawback: Since registers are in hardware, there are a predetermined number of them Solution: MIPS code must be very carefully put together to efficiently use registers  How many registers?

Lec 2.11 Determining the number of registers? Design principle: Smaller is faster - A large number of registers would increase the clock cycle time - Balance the craving of programs for more registers with the desire to keep the clock cycle fast  32 registers in MIPS  Each MIPS register is 32 bits wide Groups of 32 bits called a word in MIPS

Lec 2.12 Assembly Variables: Registers  MIPS registers are numbered from 0 to 31  Each register can be referred to by number or name  Number references: $0, $1, $2, … $30, $31

Lec 2.13 Assembly Variables: Registers  By convention, each register also has a name to make it easier to code  For now: $16 - $23  $s0 - $s7 (correspond to C variables) $8 - $15  $t0 - $t7 (correspond to temporary variables)  In general, use names to make your code more readable

Lec 2.14 C variables vs. registers  In C (and most High Level Languages) variables declared first and given a type int fahr, celsius; char a, b, c, d, e;  Each variable can ONLY represent a value of the type it was declared as (cannot mix and match int and char variables).  In Assembly Language, the registers have no type; operation determines how register contents are treated

Lec 2.15 MIPS Addition and Subtraction  Syntax of Instructions: opopd1, opd2, opd3 where: op) operation by name opd1) operand getting result (destination) opd2) 1st operand for operation (source1) opd3) 2nd operand for operation (source2)  Syntax is rigid: 1 operator, 3 operands Why?

Lec 2.16 MIPS Addition and Subtraction “The natural number of operands for an operation like addition is three…requiring every instruction to have exactly three operands, no more and no less, conforms to the philosophy of keeping the hardware simple” Design Principle: Keep hardware simple by regularity

Lec 2.17 Addition and Subtraction of Integers  Addition in Assembly Example: add $s0, $s1, $s2 (in MIPS) Equivalent to: a = b + c (in C) where MIPS registers $s0, $s1, $s2 are associated with C variables a, b, c  Subtraction in Assembly Example: sub $s3, $s4, $s5 (in MIPS) Equivalent to: d = e - f (in C) where MIPS registers $s3, $s4, $s5 are associated with C variables d, e, f

Lec 2.18 Addition and Subtraction of Integers  How do the following C statement? a = b + c + d - e;  Break into multiple instructions add $t0, $s1, $s2 # temp = b + c add $t0, $t0, $s3 # temp = temp + d sub $s0, $t0, $s4 # a = temp - e  Notice: A single line of C may break up into several lines of MIPS.  Notice: Everything after the hash mark on each line is ignored (comments)

Lec 2.19 Addition and Subtraction of Integers  How do we do this? f = (g + h) - (i + j);  Use intermediate temporary register add $t0,$s1,$s2# temp = g + h add $t1,$s3,$s4# temp = i + j sub $s0,$t0,$t1# f=(g+h)-(i+j)

Lec 2.20 Register Zero  One particular immediate, the number zero (0), appears very often in code.  So, we define register zero ($0 or $zero ) to always have the value 0; eg add $s0, $s1, $zero (in MIPS) f = g (in C) where MIPS registers $s0, $s1 are associated with C variables f, g  Defined in hardware, so an instruction add $zero, $zero, $s0 will not do anything!

Lec 2.21 Immediates  Immediates are numerical constants  They appear often in code, so there are special instructions for them  Add Immediate: addi $s0,$s1,10 (in MIPS) f = g + 10 (in C) where MIPS registers $s0, $s1 are associated with C variables f, g  Syntax similar to add instruction, except that last argument is a number instead of a register.

Lec 2.22 Immediates  There is no Subtract Immediate in MIPS. Why?  Limit types of operations that can be done to absolute minimum if an operation can be decomposed into a simpler operation, don’t include it addi …, -X = subi …, X => so no subi  addi $s0,$s1,-10 (in MIPS) f = g - 10 (in C) where MIPS registers $s0, $s1 are associated with C variables f, g

Lec 2.23 Overflow in Arithmetic  Reminder: Overflow occurs when there is a mistake in arithmetic due to the limited precision in computers.  Example (4-bit unsigned numbers): But we don’t have room for 5-bit solution, so the solution would be 0010, which is +2, and wrong.

Lec 2.24 Overflow in Arithmetic  Some languages detect overflow (Ada), some don’t (C)  MIPS solution is 2 kinds of arithmetic instructions to recognize 2 choices: add ( add ), add immediate ( addi ) and subtract ( sub ) cause overflow to be detected add unsigned ( addu ), add immediate unsigned ( addiu ) and subtract unsigned ( subu ) do not cause overflow detection  Compiler selects appropriate arithmetic MIPS C compilers produce addu, addiu, subu

Lec 2.25 Assembly Operands: Memory  C variables map onto registers  What about large data structures like arrays? only 32 registers provided  Memory contains such data structures  But MIPS arithmetic instructions only operate on registers, never directly on memory  Data transfer instructions transfer data between registers and memory: Memory to register Register to memory

Lec 2.26 Anatomy: 5 components of any Computer Processor Computer Control (“brain”) Datapath Registers MemoryDevices Input Output Load (from) Store (to) These are “data transfer” instructions…  Registers are in the datapath of the processor  If operands are in memory, we must transfer them to the processor to operate on them, and then transfer back to memory when done

Lec 2.27 Memory Organization  Viewed as a large, single-dimension array, with an address  A memory address is an index into the array  "Byte addressing" means that the index points to a byte of memory bits of data

Lec 2.28 Memory Organization  Bytes are nice, but most data items use larger "words"  For MIPS, a word is 32 bits or 4 bytes.  2 32 bytes with byte addresses from 0 to  2 30 words with byte addresses 0, 4, 8, bits of data Registers hold 32 bits of data

Lec 2.29  Big Endian: address of most significant byte = word address (xx00 = Big End of word) IBM 360/370, Motorola 68k, MIPS, Sparc, HP PA  Little Endian: address of least significant byte = word address (xx00 = Little End of word) Intel 80x86, DEC Vax, DEC Alpha (Windows NT) msb lsb little endian byte big endian byte 0 Endianess How do byte addresses map onto words?

Lec Aligned Not Aligned Alignment Can a word be placed on any byte boundary? Alignment: objects fall on address that is multiple of their size MIPS require words to be always aligned, i.e,. start at addresses that are multiples of four

Lec 2.31 Data Transfer: Memory to Reg  To transfer a word of data, we need to specify two things: Register: specify this by # ($0 - $31) or symbolic name ($s0,…, $t0, …) Memory address: more difficult -We can address it simply by supplying a pointer to a memory address -Other times, we want to be able to offset from this pointer  Remember: Load FROM memory

Lec 2.32 Data Transfer: Memory to Reg  To specify a memory address to copy from, specify two things: A register containing a pointer to memory A numerical offset (in bytes)  The desired memory address is the sum of these two values.  Example: 8($t0) specifies the memory address pointed to by the value in $t0, plus 8 bytes

Lec 2.33 Data Transfer: Memory to Reg  Load Instruction Syntax: op opd1, opd2(opd3) where op) operation name op1) register that will receive value op2) numerical offset in bytes op3) register containing pointer to memory  MIPS Instruction Name: lw (meaning Load Word, so 32 bits or one word are loaded at a time)

Lec 2.34 Data Transfer: Memory to Reg Example: lw $t0,12($s0) This instruction will take the pointer in $s0, add 12 bytes to it, and then load the value from the memory pointed to by this calculated sum into register $t0  Notes: $s0 is called the base register 12 is called the offset offset is generally used in accessing elements of array or structure: base reg points to beginning of array or structure Data flow

Lec 2.35 Data Transfer: Reg to Memory  Also want to store from register into memory Store instruction syntax is identical to Load’s  MIPS Instruction Name: sw (meaning Store Word, so 32 bits or one word are loaded at a time)  Example: sw $t0,12($s0) This instruction will take the pointer in $s0, add 12 bytes to it, and then store the value from register $t0 into that memory address  Remember: Store INTO memory Data flow

Lec 2.36 Compilation with Memory  What offset in lw to select A[5] in C?  4x5=20 to select A[5] : byte v. word  Compile by hand using registers: g = h + A[5]; g : $s1, h : $s2, $s3 :base address of A  1st transfer from memory to register: lw $t0,20($s3) # $t0 gets A[5] add $s1,$s2,$t0 # $s1 = h + A[5]

Lec 2.37 Compilation with Memory  Example: C code: A[12] = h + A[8]; MIPS code: lw $t0, 32($s3) add $t0, $s2, $t0 sw $t0, 48($s3)  Store word has destination last  Remember arithmetic operands are registers, not memory! Can’t write: add 48($s3), $s2, 32($s3)

Lec 2.38 Notes about Memory  Pitfall: Forgetting that sequential word addresses in machines with byte addressing do not differ by 1 Remember that for both lw and sw, the sum of the base address and the offset must be a multiple of 4 (to be word aligned)

Lec 2.39 Pointers v. Values  Key Concept: A register can hold any 32-bit value. That value can be a (signed) int, an unsigned int, a pointer (memory address), and so on  If you write add$t2,$t1,$t0 then $t0 and $t1 better contain values  If you write lw $t2,0($t0) then $t0 better contain a pointer  Don’t mix these up!

Lec 2.40 Role of Registers vs. Memory  What if more variables than registers? Compiler tries to keep most frequently used variable in registers Less common in memory: spilling  Why not keep all variables in memory? Smaller is faster: registers are faster than memory Registers more versatile: -MIPS arithmetic instructions can read 2, operate on them, and write 1 per instruction -MIPS data transfer only read or write 1 operand per instruction, and no operation

Lec 2.41 So far we’ve learned:  MIPS - arithmetic on registers and immediates only - loading words but addressing bytes  InstructionMeaning add $s1, $s2, $s3$s1 = $s2 + $s3 sub $s1, $s2, $s3$s1 = $s2 – $s3 addi $s1, $s2, 10 $s1 = $s lw $s1, 100($s2)$s1 = Memory[$s2+100] sw $s1, 100($s2)Memory[$s2+100] = $s1

Lec 2.42 Loading, Storing Bytes  In addition to word data transfers ( lw, sw ), MIPS has byte data transfers:  load byte: lb  store byte: sb  same format as lw, sw

Lec 2.43 x Loading, Storing Bytes  What do with other 24 bits in the 32 bit register? lb : sign extends to fill upper 24 bits byte loaded …is copied to “sign-extend” This bit xxxx xxxx xxxx zzz zzzz Normally don't want to sign extend chars MIPS instruction that doesn’t sign extend when loading bytes: load byte unsigned: lbu

Lec 2.44 So Far...  All instructions so far only manipulate data…we’ve built a calculator.  In order to build a computer, we need ability to make decisions…  C (and MIPS) provide labels to support “ goto ” jumps to places in code. C: Horrible style; MIPS: Necessary!

Lec 2.45 C Decisions: if Statements  2 kinds of if statements in C if ( condition ) clause if ( condition ) clause1 else clause2  Rearrange 2nd if into following: if ( condition ) goto L1; clause2; goto L2; L1: clause1; L2:  Not as elegant as if-else, but same meaning

Lec 2.46 MIPS Decision Instructions  Decision instruction in MIPS: beq register1, register2, L1 beq is “Branch if (registers are) equal” Same meaning as (using C): if (register1==register2) goto L1  Complementary MIPS decision instruction bne register1, register2, L1 bne is “Branch if (registers are) not equal” Same meaning as (using C): if (register1!=register2) goto L1  Called conditional branches

Lec 2.47 MIPS Goto Instruction  In addition to conditional branches, MIPS has an unconditional branch: jlabel  Called a Jump Instruction: jump (or branch) directly to the given label without needing to satisfy any condition  Same meaning as (using C): goto label  Technically, it’s the same as: beq $0,$0,label since it always satisfies the condition.

Lec 2.48 Compiling C if into MIPS  Compile by hand if (i == j) f=g+h; else f=g-h;  Use this mapping: f : $s0 g : $s1 h : $s2 i : $s3 j : $s4 Exit i == j? f=g+hf=g-h (false) i != j (true) i == j

Lec 2.49 Compiling C if into MIPS  Final compiled MIPS code: beq $s3,$s4,True # branch i==j sub $s0,$s1,$s2 # f=g-h(false) j Fin # goto Fin True: add $s0,$s1,$s2 # f=g+h (true) Fin:  Compiler automatically creates labels to handle decisions (branches) – Generally not found in HLL code. Exit i == j? f=g+hf=g-h (false) i != j (true) i == j Compile by hand if (i == j) f=g+h; else f=g-h;

Lec 2.50 Example: The C Switch Statement  Choose among four alternatives depending on whether k has the value 0, 1, 2 or 3. Compile this C code: switch (k) { case 0: f=i+j; break; /* k=0 */ case 1: f=g+h; break; /* k=1 */ case 2: f=g–h; break; /* k=2 */ case 3: f=i–j; break; /* k=3 */ }

Lec 2.51 Example: The C Switch Statement  This is complicated, so simplify.  Rewrite it as a chain of if-else statements, which we already know how to compile: if(k==0) f=i+j; else if(k==1) f=g+h; else if(k==2) f=g–h; else if(k==3) f=i–j;  Use this mapping: f:$s0, g:$s1, h:$s2, i:$s3, j:$s4, k:$s5

Lec 2.52 Example: The C Switch Statement  Final compiled MIPS code: bne $s5,$0,L1 # branch k!=0 add $s0,$s3,$s4 #k==0 so f=i+j j Exit # end of case so Exit L1: addi $t0,$s5,-1 # $t0=k-1 bne $t0,$0,L2 # branch k!=1 add $s0,$s1,$s2 #k==1 so f=g+h j Exit # end of case so Exit L2: addi $t0,$s5,-2 # $t0=k-2 bne $t0,$0,L3 # branch k!=2 sub $s0,$s1,$s2 #k==2 so f=g-h j Exit # end of case so Exit L3: addi $t0,$s5,-3 # $t0=k-3 bne $t0,$0,Exit # branch k!=3 sub $s0,$s3,$s4 #k==3 so f=i-j Exit:

Lec 2.53 Loops in C/Assembly  Simple loop in C; A[] is an array of int s do { g = g + A[i]; i = i + j; } while (i != h);  Rewrite this as: Loop:g = g + A[i]; i = i + j; if (i != h) goto Loop;  Use this mapping: g, h, i, j, base of A $s1, $s2, $s3, $s4, $s5

Lec 2.54 Loops in C/Assembly  Final compiled MIPS code: Loop: add $t1,$s3,$s3 #$t1= 2*i add $t1,$t1,$t1 #$t1= 4*i add $t1,$t1,$s5 #$t1=addr A lw $t1,0($t1) #$t1=A[i] add $s1,$s1,$t1 #g=g+A[i] add $s3,$s3,$s4 #i=i+j bne $s3,$s2,Loop # goto Loop # if i!=h  Original code: Loop:g = g + A[i]; i = i + j; if (i != h) goto Loop;

Lec 2.55 Compiling a While Loop C code: while (save[i] == k) i= i+j; MIPS code: Loop: add $t1, $s3, $s3# $t1= 2*i add $t1, $t1, $t1# $t1= 4*i add $t1, $t1, $s6# $t1= address of save[i] lw $t0, 0($t1)# $t0= save[i] bne $t0, $s5, Exit# go to Exit if save[i] ≠ k add $s3, $s3, $s4# i= i+j j Loop # go to Loop Exit:

Lec 2.56 Inequalities in MIPS  Until now, we’ve only tested equalities ( == and != in C)  General programs need to test as well.  Create a MIPS Inequality Instruction: “Set on Less Than” Syntax: slt reg1,reg2,reg3 Meaning: if (reg2 < reg3) reg1 = 1; else reg1 = 0;

Lec 2.57 Inequalities in MIPS  How do we use this? Compile by hand: if (g < h) goto Less; #g:$s0, h:$s1  Answer: compiled MIPS code… slt $t0,$s0,$s1 # $t0 = 1 if g<h bne $t0,$zero,Less # goto Less # if $t0!=0 # (if (g<h)) Less:  Branch if $t0 != 0  ( g < h )  Register $0 always contains the value 0, so bne and beq often use it for comparison after an slt instruction.  A slt  bne pair means if(… < …)goto…

Lec 2.58 Inequalities in MIPS  Now, we can implement, ≤ and ≥ ?  We could add 3 more instructions, but: MIPS goal: Simpler is Better  Can we implement ≤ in one or more instructions using just slt and the branches?  What about > ?  What about ≥ ?

Lec 2.59 Immediates in Inequalities  There is also an immediate version of slt to test against constants: slti Helpful in for loops if (g >= 1) goto Loop Loop:... slti $t0,$s0,1 # $t0 = 1 if # $s0 =1)) C MIPSMIPS

Lec 2.60 What about unsigned numbers?  Also unsigned inequality instructions: sltu, sltiu …which sets result to 1 or 0 depending on unsigned comparisons  What is value of $t0, $t1 ? ( $s0 = FFFF FFFA hex, $s1 = 0000 FFFA hex slt $t0, $s0, $s1 sltu $t1, $s0, $s1

Lec 2.61 MIPS Signed vs. Unsigned – diff meanings!  MIPS Signed v. Unsigned is an “overloaded” term Do/Don't sign extend ( lb, lbu ) Don't overflow ( addu, addiu, subu, multu, divu ) Do signed/unsigned compare ( slt, slti / sltu, sltiu )

Lec 2.62 Bitwise Operations  Up until now, we’ve done arithmetic ( add, sub,addi ), memory access ( lw and sw ), and branches and jumps  All of these instructions view contents of register as a single quantity (such as a signed or unsigned integer)  New Perspective: View register as 32 raw bits rather than as a single 32-bit number  Since registers are composed of 32 bits, we may want to access individual bits (or groups of bits) rather than the whole  Introduce two new classes of instructions: Logical & Shift Ops

Lec 2.63 Logical Operators  Two basic logical operators: AND: outputs 1 only if both inputs are 1 OR: outputs 1 if at least one input is 1  Truth Table: standard table listing all possible combinations of inputs and resultant output for each A B A AND B A OR B

Lec 2.64 Logical Operators  Logical Instruction Syntax: op opd1, opd2, opd3 where op) operation name opd1) register that will receive value opd2) first operand (register) opd3) second operand (register) or immediate

Lec 2.65 Logical Operators  Instruction Names: and, or : Both of these expect the third argument to be a register andi, ori : Both of these expect the third argument to be an immediate  MIPS Logical Operators are all bitwise, meaning that bit 0 of the output is produced by the respective bit 0’s of the inputs, bit 1 by the bit 1’s, etc. C: Bitwise AND is & (e.g., z = x & y; ) C: Bitwise OR is | (e.g., z = x | y; )

Lec 2.66 Uses for Logical Operators  Note that and ing a bit with 0 produces a 0 at the output while and ing a bit with 1 produces the original bit.  This can be used to create a mask. Example: The result of and ing these:

Lec 2.67 Uses for Logical Operators  The second bitstring in the example is called a mask. It is used to isolate the rightmost 12 bits of the first bitstring by masking out the rest of the string (e.g. setting it to all 0s).  Thus, the and operator can be used to set certain portions of a bitstring to 0s, while leaving the rest alone. In particular, if the first bitstring in the above example were in $t0, then the following instruction would mask it: andi $t0,$t0,0xFFF

Lec 2.68 Uses for Logical Operators  Similarly, note that or ing a bit with 1 produces a 1 at the output while or ing a bit with 0 produces the original bit.  This can be used to force certain bits of a string to 1s. For example, if $t0 contains 0x , then after this instruction: ori$t0, $t0, 0xFFFF … $t0 contains 0x1234FFFF (e.g. the high-order 16 bits are untouched, while the low-order 16 bits are forced to 1s).

Lec 2.69 Shift Instructions  Move (shift) all the bits in a word to the left or right by a number of bits. Example: shift right by 8 bits Example: shift left by 8 bits

Lec 2.70 Shift Instructions  Shift Instruction Syntax: op opd1, opd2, opd3 where op) operation name opd1) register that will receive value opd2) first operand (register) opd3) shift amount (constant < 32)  MIPS shift instructions: 1. sll (shift left logical): shifts left and fills emptied bits with 0s 2. srl (shift right logical): shifts right and fills emptied bits with 0s 3. sra (shift right arithmetic): shifts right and fills emptied bits by sign extending

Lec 2.71 Shift Instructions  Example: shift right arith by 8 bits  Example: shift right arith by 8 bits

Lec 2.72 Shift Instructions  Since shifting may be faster than multiplication, a good compiler usually notices when C code multiplies by a power of 2 and compiles it to a shift instruction: a *= 8; (in C) would compile to: sll $s0,$s0,3 (in MIPS)  Likewise, shift right to divide by powers of 2 remember to use sra