Microarchitecture

Microarchitecture connects circuits and ISA together. ISA = Instruction Set Architecture
ISA specifies what hardware does, but not how
- ISA defines the instruction code
Microarchitecture specifies how it does it
- Microarchitecture determine how it work it

The ISA is the interface of a processor as seen by an assembly language programmer or compiler writer
Micro-architecture transfers the ISA into an implementation
For a given ISA, there might be many different microarchitectures
An architecture is a collection of circuits connected

Von Neumann architecture

VNA

Memory
CPU
- Register
- ALU
- CU
I/O

Memory stores not only data, but coded instructions that make up a computer program
CPU fetches and executes – interprets - successive instructions of the program
Program is simply data for the interpreter – as in a Universal Turing Machine!
Single expandable resource pool – main memory
constrains both data and program size

Memory

Address space $2^{16}$
- means: there are $2^{16}$ block
Addressability $16$
Each block store 16-bit instruction
- Each block represent by a Hexadecimal code
  - x0000
  - x0001
  - ....
- Address 逐行递增
Memory is store element
Basic Operations
- LOAD
  - read a value from a memory location
- STORE
  - write a value to a memory location

Interface to Memory

MAR: Memory Address Register (D flip-flops)
MDR: Memory Data Register (D flip-flops)

To read a location (A):
1. Write the address (A) into the MAR.
2. Send a “read” signal to the memory.
3. Read the data from MDR.
To write a value (X) to a location (A):
1. Write the data (X) to the MDR.
2. Write the address (A) into the MAR.
3. Send a “write” signal to the memory.

CPU

The brain of the computer
It is the part that actually executes the machine instructions
Inside the CPU
- Data path
  - Registers
- Control Path
  - IR (instruction register), PC (program counter), FSM (finite state machine)

Register Register File

Register is store element
- Achieved by D filp-flop
8 Register
- Each register represent by 3 bit to record its address
  - 000 $R_{0}$
  - 001 $R_{1}$
  - 010 $R_{2}$
  - ...
  - 110 $R_{6}$
  - 111 $R_{7}$
Source Register 1 $S R_{1}$
Source Register 2 $S R_{2}$
Destination Register $D R$

Why register? Closer to processing unit, allow quicker access to intermediate results instead of going to memory

ALU Arithmetic Logic Unit

Perform arithmetic and logic operations (AND, NOT, ADD) on values stored in registers

ADD
AND
NOT A
PASS A

CU

Control Unit: a finite state machine coordinates execution of the program IR Get the Instruction from Memory through PC's address

PC Program Counter
- It stores the next instruction address
- In each clock, PC+=1
- In x86 architecture also called IP Instruction Pointer
IR Instruction Register
- Current Instruction
Control Unit as a Finite State Machine

I/O

Devices for getting data into and out of memory
Each device has its own interface, usually a set of registers like the memory’s MAR and MDR
- keyboard: data register (KBDR) and status register (KBSR)
- console: data register (DDR) and status register (DSR)

Instruction

The instruction is the most basic unit of computer processing.
One instruction specifies two things:
- opcode: operation to be performed
- operands: data/locations to be used for operation
An instruction is encoded as a sequence of bits (just like data!)
Control unit interprets instruction
A computer’s instructions and their formats is known as its Instruction Set Architecture (ISA).

Lc-3's ADD Instruction

opcode: 0001
steering bit IR[5]

If IR[5]=0

asm

ADD DST SR1 000 SR2
0001 xxx xxx 000 xxx

operands: Src1,Src2,Dst
- Src1+Src2->Dst

If IR[5]=1

asm

ADD DST SR1 1 xxxxx
0001 xxx xxx 1 xxxxx

operands: Src1,Dst,xxxxx
- Src1+xxxxx->Dst

LC-3's LDR Instruction

opcode: 0110
operands: Src,Dst
Load the value in memory location Src into register Dst
Move [Base + Offset] to Dst
Load memory content at address (R3 + 6) to R2

txt

LDR DST SRC offset
0110 xxx xxx xxxxxx

Instruction processsing

Fetch
Decode
Evaluate Address
Fetch operands
Execute
Store result

Fetch

PC-MAR-MDR-IR

Load next instruction (at address stored in PC) from memory into Instruction Register (IR).

Load contents of PC into MAR.
Send “read” signal to memory.
Read contents of MDR, store in IR.

Then increment PC, so that it points to the next instruction in sequence.

PC becomes PC+1.

Decode

First identify the opcode.

In LC-3, this is always the first four bits of instruction.
A 4-to-16 decoder asserts a control line corresponding to the desired opcode.

Evaluate Address 计算地址

For instructions that require memory access, compute address used for access.
Examples:
- add offset to base register (as in LDR)
- add offset to PC (or to part of PC)

Fetch Operands

Obtain source operands needed to perform operation.
Examples:
- load data from memory (LDR)
- read data from register file (ADD)

Execution

Perform the operation, using the source operands.
Examples:
- send operands to ALU and assert ADD signal
- do nothing (e.g., for loads and stores)

Store

Write results to destination (register or memory)
Examples:
- result of ADD is placed in destination register
- result of memory load is placed in destination register
for store instruction, data is stored to memory

Changing the Sequence of Execution

In the FETCH phase, PC is incremented by 1 automatically (counter)
Other Sequence
- if-then, loop, function call
- Achieved by special instruction that changes the content of PC
  - Jumps (unconditionals)
  - Branches (conditional)

LC-3's Jump Instruction

Set the PC to the value obtained by adding an offset to a register.

txt

JMPR 000 BASE offset
1100 000 011 000110

Add the value of 6 (offset) to the contents of R3 (Base),and load the result into the PC

This becomes the address of the next instruction to fetch.

Driving Force: the Clock

The clock is a signal that keeps the control unit moving.

At each clock “tick,” control unit moves to the next machine cycle -- may be next instruction or next phase of current instruction.
Stopping the Computer
- Stopping the instruction cycle requires stopping the clock

Instruction summary

Three basic kinds of instructions:
- computational instructions (ADD, AND, …)
- data movement instructions (LD, ST, …)
- control instructions (JMP, BRnz, …)