Module -5

Chapter 6
Intermediate Code Generation

Prof. Maya B S
Assistant Professor
Department of CS&E
BIT, Bangalore

Prof.Maya B S, Assistant Professor, CSE,BIT 1

Prof.Maya B S, Assistant Professor, CSE,BIT 2
 Chapter 6
Intermediate Code Generation

6.1 Variants of Syntax tree

6.2 Three Address Code

Prof.Maya B S, Assistant Professor, CSE,BIT 3

Intermediate Code Generation

Prof.Maya B S, Assistant Professor, CSE,BIT 4

Prof.Maya B S, Assistant Professor, CSE,BIT 5
Prof.Maya B S, Assistant Professor, CSE,BIT 6
6.1 Variants of Syntax tree

Prof.Maya B S, Assistant Professor, CSE,BIT 7

6.1.1 Directed Acyclic Graph for Expressions

• The syntax tree for an expression, a DAG has leaves corresponding to

atomic operands and interior codes corresponding to operators.

• The difference is that a node ‘N’ in a DAG has more than one parent.

• if ‘N’ represents a common subexpression.

• In a syntax tree , the tree for the common subexpression would be

replicated as many times as the subexpression appears in the original
expression.

• DAG not only represent expression , it gives the compiler important clues
regarding the generation of efficient code to evaluate the expression.

Prof.Maya B S, Assistant Professor, CSE,BIT 8

1.write DAG for below expression a+a*(b-c)+(b-c)*d

• The leaf for ‘a’ has 2 parents, because ‘a’ appears in the expressions.
• Two occurrence of the common subexpression b-c are represented by one
node , he node labeled by –
• That node has 2 parents , representing its two uses in the subexpression
a*(b-c) and (b-c)*d.
• Even though ‘b’ and ‘c’ appears twice in the complete expression, their
nodes each have one parent, since both uses are in the common
subexpression b-c.

Prof.Maya B S, Assistant Professor, CSE,BIT 9

SDD for Directed Acyclic Graph for Expressions

Prof.Maya B S, Assistant Professor, CSE,BIT 10

• We assume that entry-1 points to the symbol table entry for ‘a’ and
similarly for other identifier.

• When the call to leaf(id,entry-1) is repeated at step2, the node created by

the previous call is returned so p2=p1.

• Similarly, the nodes returned at step8 and 9 are the same as those returned
at step 3 and 4(i.e p8=p3, p9=p4).

• Hence the nodes returned at step 10 must be the same at that returned at
step5. ie p10=p5.

Prof.Maya B S, Assistant Professor, CSE,BIT 11

 6.1.2 The value number method for constructing DAG

• The nodes of a syntax tree or DAG are stored in an array of records as

shown in below figure.
• Ex: Nodes of a DAG i=i+10 allocated in an array.
• Each row of the array represents one record and therefore one node.
• In each record, the first field is an operation code, indicating the label of the
node.
• Fig b leaves have one additional field ,which holds the lexical value(either
a symbol table pointer or constant) and interior nodes have two additional
fields indicating the left and right children.

Prof.Maya B S, Assistant Professor, CSE,BIT 12

• In this array , we refer to nodes by giving the integer index of the record for that
node within the array.

• This integer historically has been called the value number for the node or for the
expression represented by the node.

• Above figure ,the node labeled + have value number 3 and its left and right children
have number 1 and 2 respectively.

• Use pointer to record or reference to objects instead of integer indexes ,but we

shall still refer to the reference to a node as its “valve No” .If stored in an
appropriate data structure, value number helps us construct expressions DAG
efficiently.

Prof.Maya B S, Assistant Professor, CSE,BIT 13

Algorithm for value number method for constructing DAG
Input: Label op,node l,and node r
Output: The value number of a node in the array with
signature <op,l,r >

Prof.Maya B S, Assistant Professor, CSE,BIT 14

 Problems on DAG and value number method

1. ((x+y)-((x+y)*(x-y)))+((x+y)*(x-y))

2. Construct the DAG and identify the value number for the sub expression
of the following expressions, assuming + associates from left.

a) a+b+(a+b)

b) a+b+a+b

c) a+a+(a+a+a+(a+a+a+a))

Prof.Maya B S, Assistant Professor, CSE,BIT 15

6.2 Three Address Code

Prof.Maya B S, Assistant Professor, CSE,BIT 16

6.2.1 Address and Instruction

Prof.Maya B S, Assistant Professor, CSE,BIT 17

Instruction of 3-address code

Prof.Maya B S, Assistant Professor, CSE,BIT 18

 Continue

Prof.Maya B S, Assistant Professor, CSE,BIT 19

Example of three address instruction

Prof.Maya B S, Assistant Professor, CSE,BIT 20

Consider the statement do i=i+1;while(a[i]<v);write 2possible translation of above
statement using three address code instruction.
Note: the multiplication i*8 is appropriate for an array of elements that each take
8 units of space.

Prof.Maya B S, Assistant Professor, CSE,BIT 21

1.Difference between syntax tree and DAG for x=(a+b*c)/(a-b*c)

Prof.Maya B S, Assistant Professor, CSE,BIT 22

2.Difference b/w syntax tree &DAG for a+b*c-d/b*c

Prof.Maya B S, Assistant Professor, CSE,BIT 23

6.2.2 Quadruples

• The description of 3-address instruction specify the components of each

type of instruction, but it does not specify the representation of these
instruction in a data structure.
• Ina compiler, these instruction can be implemented as objects or as records
with fields for the operator and operands.
• Three such representation are called “Quadruple”, “Triples”, and “Indirect
Triples ”.

Prof.Maya B S, Assistant Professor, CSE,BIT 24

Prof.Maya B S, Assistant Professor, CSE,BIT 25
Example of different form of records

Prof.Maya B S, Assistant Professor, CSE,BIT 26

Example for 3-address code

Prof.Maya B S, Assistant Professor, CSE,BIT 27

 Quadruple

• A “Quadruple” has 4 fields, which we call op,arg1,arg2 and result.

• Op field contains an internal code for the operator.

• Ex: 3-address instruction x=y+z is represented by ‘+’ in op, y in arg1,z in

arg2 and ‘x’ in result.

• The following are some exception to this rule

• Instructions with unary operator like x=minus y or x=y do not use arg2.
note that for a copy statement like x=y, op is = while for most other
operations, the assignment operator is implied.

• Operator like param use neither arg2 nor result.

• Conditional and unconditional jumps put the target label in result.

Prof.Maya B S, Assistant Professor, CSE,BIT 28

1.Write the 3-addres code for the assignment a=b*-c+b*-c
and also quadruple.(VVVVIMP)

Prof.Maya B S, Assistant Professor, CSE,BIT 29

2 Write the 3-addres code for the assignment (a+b*c)-(d/b*c) and also
quadruple.

Prof.Maya B S, Assistant Professor, CSE,BIT 30

3.Write the 3-addres code for the assignment a=b+c*d and
also quadruple.

Prof.Maya B S, Assistant Professor, CSE,BIT 31

6.2.3 Triples
 A Triple has only 3 fields, which we call op,arg1, and arg2.

 The result field in a quadruples is used primarily for temporary names.

Using triples we refer to the result of an operation x op y by its position,
rather than by an explicit temporary name.

 Instead of the temporary ‘t1’ in quadruple , a triple representation would

refer to position(0).

 Parenthesized numbers represents pointer into the triple structure itself.

Prof.Maya B S, Assistant Professor, CSE,BIT 32

Write syntax tree and triples for expression a+a*(b-c)+(b-c)*d

• We construct syntax tree in fig a and fig b shows Triples.

Prof.Maya B S, Assistant Professor, CSE,BIT 33

2 Write the 3-addres code for the assignment (a+b*c)-(d/b*c) and also
Triples

• Triples shown in figure

Prof.Maya B S, Assistant Professor, CSE,BIT 34

3.Write the 3-addres code for the assignment a=b+c*d and also Triples.

Prof.Maya B S, Assistant Professor, CSE,BIT 35

Indirect Triples

• In this, an optimizing compiler can move an instruction by reordering the

instruction list, without affecting the triples themselves.

• When implemented in java, an array of instruction objects is analogous to

an indirect triple representation, since JAVA treats the array elements as
reference to objects.

Prof.Maya B S, Assistant Professor, CSE,BIT 36

Examples of Indirect Triples for (a+b*c)-(d/b*c)

Prof.Maya B S, Assistant Professor, CSE,BIT 37

Translate the arithmetic expression a+-(b+c) into
a) Syntax tree b) Quaduples c)Triples d)Indirect Triples

• Home work

Prof.Maya B S, Assistant Professor, CSE,BIT 38

6.2.4 Static single Assignment Form(SSA)

• SSA is an intermediate representation that facilities certain code

optimization.

• Two distinctive aspects distinguish SSA from 3 address code.

 All assignment in SSA are to variables with distinct names; Hence the term
static single assignment .

 Below figure shows the same intermediate program in 3 address code and
in static single assignment form .subscripts distinguish each definition of
variables p and q in the SSA representation.

Prof.Maya B S, Assistant Professor, CSE,BIT 39

The same variable may be defined in 2 different control flow path in a program.

Prof.Maya B S, Assistant Professor, CSE,BIT 40

Prof.Maya B S, Assistant Professor, CSE,BIT 41
• The source program,

• If(flag) x=-1;else x=1;

• Y=x*a;

• Has two control flow paths in which the variable ‘X’ gets defined.

• If we use different names for X in the true part and the false part of the
conditional statement , then which name should we use in assignment
y=x*a?

Prof.Maya B S, Assistant Professor, CSE,BIT 42

• Here is where the 2nd distinctive aspect of SSA comes into play?
 SSA uses a notational convention called the ᶲ function to combine
the two definition of x:
If(flag) x1=-1;else x2=1;
X3= ᶲ(x1,x2);
Here, ᶲ(x1,x2) has value ‘x1’ if the control flow passes through the true
part of the conditional and the value x2 if the control flow passes
through the false part.
i.e ᶲ-function returns the value of its argument that corresponds to the
control flow path that was taken to get to the assignment statement
containing the ᶲ-function.

Prof.Maya B S, Assistant Professor, CSE,BIT 43

 THANK YOU

Prof.Maya B S, Assistant Professor, CSE,BIT 44