University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell

CS352H: Computer Systems Architecture

Lecture 4: Instruction Set Architectures III + MIPSALU

September 10, 2008

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 2

ISA Example: IA-32

A 20-year old ISA! (30 years if you count earlier x86)Blend of RR & RM ISA

Two-operand instructionsSeven addressing modesMode to indicate whether 8086, 286 or IA-32 (>=386)

More than 630 instructions!Evolved to include multimedia, graphicsSupport operations on 8, 16 and 32-bit operands

Variable-length instructions1 – 15 bytes long (3 bytes on average)

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 3

IA-32 Registers

Originally 16-bit registers

As of 386 eight GP 32-bit regs

Segment registers encodeaddress prefix

Certain accesses particularsegment registers by default

Condition code register

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 4

IA Instruction Formats

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 5

Principles of ISA Design

KISS — Keep It Simple, Stupid (Cray)Complexity increases

Logic areaPipe stage durationDevelopment time

Evolution leads to kludges

OrthogonalitySimple rules, few exceptionsAll ops on all registers in all addressing modes

FrequencyMake the common case fast

Instructions are not born equal!

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 6

Principles of ISA Design (cont)

GeneralityNot all problems need the same capabilities

Toaster oven vs. supercomputer

Performance should be easy to predictPeter Denning: Don’t build it if you can’t model it

Permit efficient implementationToday10 years from now

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 7

CISC vs. RISC

Complex Instruction Set Computer(CISC)Typically includes:

Variable-length instructionsRM instructionsMany, complex addressingmodesComplex instructions

Examples:VAX, IBM 360/370, x86

AdvantagesCode densityLegacy SW

Reduced Instruction Set Computer(RISC)Typically includes:

GP registersFixed 3-address formatStrict load/store conformanceFew, simple addressing modesSimple instructions

Examples:DEC Alpha, MIPS

AdvantagesGood compiler targetEasy to implement/pipeline

In practice, other than a handful of ISAs, most contain elements of both.

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 8

MIPS ALU

Needs to supportadd, addi, suband, andi, or, ori, xor, xorisll, srl, sra, sllv, srlv, sravslt, sltiAll the unsigned versions

A

B

ALU

32

SE32

1632

OP

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 9

Bit Slice Approach

• For subtraction: set invert & CarryIn to 1 (2’s complement arithmetic)• How about SLT?

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 10

LSB and MSB Need to Do Extra

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 11

Overflow DetectionOverflow occurs when the result is too large to represent in the

number of bits allocatedadding two positives yields a negativeor, adding two negatives gives a positiveor, subtract a negative from a positive gives a negativeor, subtract a positive from a negative gives a positive

On your own: Prove you can detect overflow by:Carry into MSB xor Carry out of MSB

1

1

1 10

1

0

1

1

0

0 1 1 1

0 0 1 1+

7

3

0

1

– 6

1 1 0 0

1 0 1 1+

–4

– 5

71

0

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 12

Overflow Detection Logic

For an N-bit ALU: Overflow = CarryIn[N-1] XOR CarryOut[N-1]

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 13

LSB and MSB Need to Do Extra

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 14

What About Performance?

Critical path of N-bit ripple-carry adder is N * Tadd

Throw hardware at it! (Next lecture)

University of Texas at Austin CS352H - Computer Systems Architecture Fall 2009 Don Fussell 15

Next Lecture

Finish MIPS ALUSee Appendix C (on CD)

In particular, Sections C-5 & C-6 for fast addersSee Chapter 3.3 & 3.4 for Multiplication & DivisionSee Chapter 3.5 for Floating point ops