Embedded Systems Design: A Unified Hardware/Software Introduction

Chapter 2: Custom single-purpose processors

Outline
Introduction Combinational logic Sequential logic Custom single-purpose processor design RT-level custom single-purpose processor design

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

Introduction
Processor
Digital circuit that performs a computation tasks Controller and datapath CCD General-purpose: variety of computation tasks Single-purpose: one particular lens computation task Custom single-purpose: non-standard task
Digital camera chip

A2D

CCD preprocessor

Pixel coprocessor

D2A

JPEG codec

Microcontroller

Multiplier/Accum

A custom single-purpose processor may be

Fast, small, low power But, high NRE, longer time-to-market, less flexible
Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

DMA controller

Display ctrl

Memory controller

ISA bus interface

UART

LCD ctrl

CMOS transistor on silicon

Transistor
The basic electrical component in digital systems Acts as an on/off switch Voltage at gate controls whether current flows from source to drain Dont confuse this gate with a logic gate gate
1

source Conducts if gate=1 drain

IC package

IC

source

gate oxide channel

drain Silicon substrate

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

CMOS transistor implementations

Complementary Metal Oxide Semiconductor We refer to logic levels
Typically 0 is 0V, 1 is 5V
source gate Conducts if gate=1 drain gate source Conducts if gate=0 drain

nMOS

pMOS

Two basic CMOS types

nMOS conducts if gate=1 pMOS conducts if gate=0 Hence complementary
1 x x F = x' x 0 y 0 inverter NAND gate x 0 NOR gate 1 y x y 1

F = (xy)'

F = (x+y)'
y

Basic gates
Inverter, NAND, NOR

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

Basic logic gates

x F

x 0 1

F 0 1

x y F

F=x Driver

F=xy AND

x 0 0 1 1

y 0 1 0 1

F 0 0 0 1

x y

F=x+y OR

x 0 0 1 1

y 0 1 0 1

F 0 1 1 1

x y F

F=xy XOR

x 0 0 1 1

y 0 1 0 1

F 0 1 1 0

x 0 1

F 1 0

x y F

F = x Inverter

F = (x y) NAND

x 0 0 1 1

y 0 1 0 1

F 1 1 1 0

x y

F = (x+y) NOR

x 0 0 1 1

y 0 1 0 1

F 1 0 0 0

x y F

F=x y XNOR

x 0 0 1 1

y 0 1 0 1

F 1 0 0 1

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

Combinational logic design

A) Problem description y is 1 if a is 1, or b and c are 1. z is 1 if b or c is to 1, but not both, or if all are 1.
a 0 0 0 0 1 1 1 1 B) Truth table Inputs b c 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 Outputs y z 0 0 0 1 0 1 1 0 1 0 1 1 1 1 1 1

C) Output equations
y = a'bc + ab'c' + ab'c + abc' + abc z = a'b'c + a'bc' + ab'c + abc' + abc

D) Minimized output equations y bc a 00 01 11 10 0 0 0 1 0 1 1 1 1 1 y = a + bc z bc a 00 01 11 10 0 0 1 0 1 1 0 1 1 1 z = ab + bc + bc

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

E) Logic Gates a b c

Combinational components
I(m-1) I1 I0 n S0 n-bit, m x 1 Multiplexor S(log m) n I(log n -1) I0 log n x n Decoder O(n-1) O1 O0 A n n-bit Adder n carry sum less equal greater B n A n B n A n B

n-bit Comparator

n bit, m function S0 ALU S(log m) n O

O= I0 if S=0..00 I1 if S=0..01 I(m-1) if S=1..11

O0 =1 if I=0..00 O1 =1 if I=0..01 O(n-1) =1 if I=1..11

sum = A+B (first n bits) carry = (n+1)th bit of A+B

less = 1 if A<B equal =1 if A=B greater=1 if A>B

O = A op B op determined by S.

With enable input e With carry-in input Ci all Os are 0 if e=0 sum = A + B + Ci

May have status outputs carry, zero, etc.

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

Sequential components
I n load clear n-bit Register n Q Q= 0 if clear=1, I if load=1 and clock=1, Q(previous) otherwise. Q = lsb - Content shifted - I stored in msb shift I n-bit Shift register n-bit Counter n Q Q= 0 if clear=1, Q(prev)+1 if count=1 and clock=1.

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

Sequential logic design

A) Problem Description B) Construct a clock divider. Slow down your preexisting clock so that you output a 1 for every four clock cycles
a C) Implementation Model Combinational logic x I1 I0 Q1 Q0 State register B) State Diagram a=0 x=0 x=1 a=1 a=0 I1 I0 Q1 0 0 0 0 1 1 1 1 D) State Table (Moore-type) Inputs Q0 a 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 Outputs I0 0 1 1 0 0 1 1 0

I1 0 0 0 1 1 1 1 0

x 0 0 0 1

0
a=1 1 a=0 x=0

3
a=1 2 x=0 a=0

Given this implementation model

Sequential logic design quickly reduces to combinational logic design
10

a=1

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

Sequential logic design (cont.)

E) Minimized Output Equations F) Combinational Logic

I1 Q1Q0 00 a 0
1

01

11

10

0
0

0
1

1
0

1
1

a I1 = Q1Q0a + Q1a + Q1Q0 x

I0 Q1Q0 00
a

01 1 0

11 1 0

10 0 1 I0 = Q0a + Q0a

I1

0 1

x Q1Q0 00 a 0 1 0 0

I0 01 0 0 11 1 1 10 0 0 x = Q1Q0 Q1 Q0

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

11

Custom single-purpose processor basic model

external control inputs controller datapath control inputs external data inputs datapath

controller next-state and control logic state register datapath registers

external control outputs

datapath control outputs

external data outputs

functiona l units

controller and datapath

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

a view inside the controller and datapath

12

Finite State Machine With Data

Program FSMD Each statement Assignment Statement Loop Statement Branch Statement

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

13

State diagram templates

Assignment statement Loop statement
while (cond) { loop-bodystatements } next statement
!cond
cond

Branch statement
if (c1) c1 stmts else if c2 c2 stmts else other stmts next statement
C: c1 c1 stmts !c1*c2 c2 stmts !c1*!c2 others

a=b next statement

C:

a=b next statement

J:

loop-bodystatements

J: next statement

next statement

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

14

Example: greatest common divisor

!1

First create algorithm Convert algorithm to complex state machine

Known as FSMD: finitestate machine with datapath Can use templates to perform such conversion

(a) black-box view

1: 1 2: !(!go_i)

(c) state diagram

go_i

x_i GCD

y_i
2-J: 3:

!go_i

d_o

x = x_i

4:

y = y_i !(x!=y) x!=y

0: int x, y; 1: while (1) { 2: while (!go_i); 3: x = x_i; 4: y = y_i; 5: while (x != y) { 6: if (x < y) 7: y = y - x; else 8: x = x - y; } 9: d_o = x; (b) desired functionality }

5:

6: x<y 7: y = y -x 6-J: !(x<y)

8: x = x - y

5-J: 9: 1-J: d_o = x

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

15

Creating the Datapath- The FOUR steps

1. Create a register for any declared variable 2. Create a functional unit for each arithmetic operation 3. Connect the ports, registers and functional units
Based on reads and writes Use multiplexors for multiple sources
7: !1 1: 1 2: !go_i 2-J: x_sel 3: x = x_i y_sel n-bit 2x1 n-bit 2x1 !(!go_i) x_i y_i

Datapath

x_ld
4: y = y_i !(x!=y) x!=y 6: x<y y = y -x 6-J: !(x<y) != 5: x!=y x_neq_y x_lt_y y_ld

0: x

0: y

5:

< 6: x<y

subtractor 8: x-y

subtractor 7: y-x

8: x = x - y

9: d d_o

d_ld

4. Create unique identifier

for each datapath component control input and output
Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

5-J: 9: 1-J: d_o = x

16

Creating the controllers FSM

!1 go_i

1:
1 2: !go_i 2-J: 3: x = x_i !(!go_i)

Controller
0000 0001 1: 1 2: !go_i 0010 2-J: 0011 x_sel = 0 3: x_ld = 1 y_sel = 0 4: y_ld = 1 5: 6:

!1 !(!go_i)

Same structure as FSMD Replace complex actions/conditions with datapath configurations

x_i y_i

4:

y = y_i 0100 !(x!=y) 0101 x!=y

Datapath
!x_neq_y x_sel n-bit 2x1 n-bit 2x1

5:

6: x<y 7: y = y -x 6-J: !(x<y)

0110

x_neq_y !x_lt_y x_sel = 1 8: x_ld = 1 1000

y_sel x_ld 0: x 0: y

8: x = x - y

x_lt_y 7: y_sel = 1 y_ld = 1 0111 1001 6-J:

y_ld

!= 5: x!=y x_neq_y x_lt_y

< 6: x<y

subtractor 8: x-y

subtractor 7: y-x

5-J:

1010 5-J: d_o = x 1011 9: d_ld = 1

9:
1-J:

9: d d_o

d_ld

1100 1-J:

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

17

Splitting into a controller and datapath

go_i

Controller implementation model

go_i Combinational logic x_sel y_sel x_ld y_ld x_neq_y x_lt_y d_ld

Controller
0000 0001 1: 1 2: !go_i 0010 2-J: 0011 x_sel = 0 3: x_ld = 1 y_sel = 0 4: y_ld = 1 5: 6:

!1 x_i !(!go_i) x_sel y_sel x_ld y_ld 0: x 0: y n-bit 2x1 n-bit 2x1 y_i

(b) Datapath

0100 0101 Q3 Q2 Q1 Q0 0110 State register I3 I2 I1 I0

!= x_neq_y=0 5: x!=y x_neq_y x_lt_y d_ld

< 6: x<y

subtractor 8: x-y

subtractor 7: y-x

x_neq_y=1 x_lt_y=0 x_sel = 1 8: x_ld = 1 1000

9: d d_o

x_lt_y=1 7: y_sel = 1 y_ld = 1 0111

1001 6-J:
1010 5-J: 1011 9: d_ld = 1

1100 1-J:

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

18

Controller state table for the GCD example

Inputs Q3 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Q2 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1 1 Q1 0 0 0 1 1 0 0 0 1 1 1 0 0 1 1 0 0 1 1 Q0 0 1 1 0 1 0 1 1 0 0 1 0 1 0 1 0 1 0 1
x_neq _y *

Outputs x_lt_y * * * * * * * * 0 1 * * * * * * * * * go_i * 0 1 * * * * * * * * * * * * * * * * I3 0 0 0 0 0 0 1 0 1 0 1 1 1 0 1 0 0 0 0 I2 0 0 0 0 1 1 0 1 0 1 0 0 0 1 1 0 0 0 0 I1 0 1 1 0 0 0 1 1 0 1 0 0 1 0 0 0 0 0 0 I0 1 0 1 1 0 1 1 0 0 1 1 1 0 1 0 0 0 0 0 x_sel X X X X 0 X X X X X X 1 X X X X X X X y_sel X X X X X 0 X X X X 1 X X X X X X X X x_ld 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 y_ld 0 0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 d_ld 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0

* * * * * 0 1 * * * * * * * * * * *

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

19

Completing the GCD custom single-purpose processor design

We finished the datapath We have a state table for the next state and control logic
All thats left is combinational logic design
controller datapath registers

next-state and control logic

state register

functional units

This is not an optimized design, but we see the basic steps

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

a view inside the controller and datapath

20

RT-level custom single-purpose processor design

Problem Specification

We often start with a state machine

Rather than algorithm Cycle timing often too central to functionality

Sende r

rdy_in clock data_in(4)

Bridge A single-purpose processor that converts two 4-bit inputs, arriving one at a time over data_in along with a rdy_in pulse, into one 8-bit output on data_out along with a rdy_out pulse.

rdy_out

Rece iver

data_out(8)

Example
Bus bridge that converts 4-bit bus to 8-bit bus Start with FSMD Known as register-transfer (RT) level Exercise: complete the design
Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

rdy_in=0 rdy_in=1 WaitFirst4

Bridge RecFirst4Start data_lo=data_in rdy_in=0 rdy_in=1 RecSecond4Start data_hi=data_in rdy_in=0

rdy_in=1 RecFirst4End

rdy_in=0 WaitSecond4

rdy_in=1 RecSecond4End

FSMD

Send8Start data_out=data_hi & data_lo rdy_out=1

Send8End rdy_out=0

Inputs rdy_in: bit; data_in: bit[4]; Outputs rdy_out: bit; data_out:bit[8] Variables data_lo, data_hi: bit[4];

21

RT-level custom single-purpose processor design (cont)

Bridge

(a) Controller
rdy_in=0 WaitFirst4 rdy_in=0 WaitSecond4 rdy_in=1 RecFirst4Start data_lo_ld=1 rdy_in=0 rdy_in=1 RecSecond4Start data_hi_ld=1 rdy_in=1 RecFirst4End

rdy_in=1 RecSecond4End

Send8Start data_out_ld=1 rdy_out=1

rdy_in clk

Send8End rdy_out=0

rdy_out

data_in(4)
data_out_ld data_hi_ld to all registers data_hi data_out data_lo data_lo_ld

data_out

(b) Datapath

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

22

Optimizing single-purpose processors

Optimization is the task of making design metric values the best possible Optimization opportunities
original program FSMD datapath FSM

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

23

Optimizing the original program

Analyze program attributes and look for areas of possible improvement
number of computations size of variable time and space complexity operations used
multiplication and division very expensive

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

24

Optimizing the original program (cont)

original program 0: int x, y; 1: while (1) { 2: while (!go_i); 3: x = x_i; 4: y = y_i; 5: while (x != y) { 6: if (x < y) 7: y = y - x; else 8: x = x - y; } 9: d_o = x; } optimized program 0: int x, y, r; 1: while (1) { 2: while (!go_i); // x must be the larger number 3: if (x_i >= y_i) { 4: x=x_i; 5: y=y_i; } 6: else { 7: x=y_i; 8: y=x_i; } 9: while (y != 0) { 10: r = x % y; 11: x = y; 12: y = r; } 13: d_o = x; } GCD(42,8) - 3 iterations to complete the loop x and y values evaluated as follows: (42, 8), (8,2), (2,0)

replace the subtraction operation(s) with modulo operation in order to speed up program

GCD(42, 8) - 9 iterations to complete the loop x and y values evaluated as follows : (42, 8), (43, 8), (26,8), (18,8), (10, 8), (2,8), (2,6), (2,4), (2,2). Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

25

Optimizing the FSMD

Areas of possible improvements
merge states
states with constants on transitions can be eliminated, transition taken is already known states with independent operations can be merged

separate states
states which require complex operations (a*b*c*d) can be broken into smaller states to reduce hardware size

scheduling

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

26

Optimizing the FSMD (cont.)

int x, y;
1: 1 2: !go_i 2-J: x = x_i y = y_i !(x!=y) x!=y 6: x<y !(x<y) y = y -x 6-J: 5-J: d_o = x 8: x = x - y !(!go_i) !1

original FSMD eliminate state 1 transitions have constant values

optimized FSMD int x, y;

2: go_i !go_i x = x_i y = y_i

3: 4: 5:

merge state 2 and state 2J no loop operation in between them

3:

5:

merge state 3 and state 4 assignment operations are independent of one another
merge state 5 and state 6 transitions from state 6 can be done in state 5 eliminate state 5J and 6J transitions from each state can be done from state 7 and state 8, respectively eliminate state 1-J transition from state 1-J can be done directly from state 9

x<y 7: y = y -x

x>y 8: x = x - y

9:

d_o = x

7:

9: 1-J:

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

27

Optimizing the datapath

Sharing of functional units
one-to-one mapping, as done previously, is not necessary if same operation occurs in different states, they can share a single functional unit

Multi-functional units
ALUs support a variety of operations, it can be shared among operations occurring in different states

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

28

Optimizing the FSM

State encoding
task of assigning a unique bit pattern to each state in an FSM size of state register and combinational logic vary can be treated as an ordering problem

State minimization
task of merging equivalent states into a single state
state equivalent if for all possible input combinations the two states generate the same outputs and transitions to the next same state

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

29

Summary
Custom single-purpose processors
Straightforward design techniques Can be built to execute algorithms Typically start with FSMD CAD tools can be of great assistance

Embedded Systems Design: A Unified Hardware/Software Introduction, (c) 2000 Vahid/Givargis

30