Some threeaddress statements that will be used later:

1

• Some three-address statements that will be used
  later
• Assignment statements
• With a binary operation x y op z
• With a unary operation x op y
• With no operation(copy) x y
• Branch statements
• Unconditional jump goto L
• Conditional jumps if x relop y goto L
• Statement for procedure calls
• Param x, set a parameter for a procedure call
• Call p, n call procedure p with n parameters
• Return y return from a procedure with return
  value y

2

• Example instructions for procedure call p(x1,
  x2, x3, , xn)
• param x1
• param x2
• param xn
• call p, n
• Indexed assignments
• x yi and xi y
• Address and pointer assignments
• x y, x y

3

• Translation scheme to produce three address code
  for assignment statement.
• Emit (instruction) emit a three address
  statement to the output
• newtemp return a new temporary
• lookup(identifier) check if the identifier is in
  the symbol table.
• Grammar
• S-gtidE
• E-gtE1E2
• E-gtE1E2
• E-gt-E
• E-gt(E1)
• E-gtid
• E-gtnum

4

• S-gtidE plookup(id.name)
• if (p!nil) then emit(p
  E.place) else error
• E-gtE1E2 E.place newtemp
• emit(E.place E1.place
  E2.place)
• E-gtE1E2 E.place newtemp
• emit(E.place E1.place
  E2.place)
• E-gt-E1 E.place newtemp
• emit(E.place -
  E1.place)
• E-gt(E1) E.place E1.place
• E-gtid p lookup(id.name)
• if (p! nil) then E.place p
  else error
• E-gtnum E.place num.val
• Example x a 10 20 -30 with bottom-up
  parsing.

5

• Temporary names can be reused
• The code generated by E-gtE1E2 has the general
  form
• Evaluate E1 into t1
• Evaluate E2 into t2
• t3 t1 t2
• All temporaries used in E1 are dead after t1 is
  evaluated
• All temporaries used in E2 are dead after t2 is
  evaluated
• T1 and t2 are dead after t3 is assigned
• Temporaries can be managed as a stack
• Keep a counter c, newtemp increases c, when a
  temporary is used, decreases c.
• See the previous example.

6

• Addressing array elements
• Arrays are typically stored in block of
  consecutive locations.
• One-dimensional arrays
• A arraylow..high of
• the ith elements is at
• base (i-low)width ? iwidth (base
  lowwidth)
• Multi-dimensional arrays
• Row major or column major forms
• Row major a1,1, a1,2, a1,3, a2,1,
  a2,2, a2,3
• Column major a1,1, a2,1, a1, 2, a2,
  2,a1, 3,a2,3
• In raw major form, the address of ai1, i2 is
• Base((i1-low1)(high2-low21)i2-low2)width

7

• In general, for any dimensional array, the
  address of ai1, i2, i3, , ik can be computed
  as follows
• Let highj and lowj be for bounds for ij
• Let nj highj lowj 1
• The address is
• (((i1n2i2)n3)nkik)width base
  ((low1n2low2)n3)nklowk) width
• When generating code for array references, we
  need to explicitly compute the address.

8

• Translation scheme for array elements
• Limit(array, j) returns njhighj-lowj1
• .place the temporarys or variables
• .offset offset from the base, null if not an
  array reference
• Grammar
• S-gtL E
• E-gtEE
• E-gt(E)
• E-gtL
• L-gtElist
• L-gtid
• Elist-gtElist, E
• Elist-gtidE

9

• S-gtL E if L.offset null then emit(L.place
  E.place)
• else emit(L.placeL.offset
  E.place)
• E-gtE1E2 E.place newtemp
• emit(E.place E1.place
  E2.place)
• E-gt(E1) E.place E1.place
• E-gtL if L.offset null then E.place L.place
• else E.place newtemp
• emit(E.place L.place
  L.offset )
• L-gtElist L.place newtemp L.offset
  newtemp
• emit(L.place
  c(Elist.array))
• emit(L.offset Elist.place
  width(Elist.array)

10

• L-gtid L.place lookup(id.name)
• L.offset null
• Elist-gtElist1, E
• t newtemp
• m Elist1.ndim 1
• emit(t Elist1.place limit(Elist1.array
  , m))
• emit(t, t E.place)
• Elist.array Elist1.array
• Elist.place t
• Elist.ndim m
• Elist-gtidE Elist.array lookup(id.name)
• Elist.place E.place
• Elist.ndim 1

11

• Example A array1..10, 1..20 of integer
• Translate x Ay, z
• Accessing fields in records?
• a.b.c.d x
• X a.b.c.d