Control Techniques UD70 - Optimizing Programs

UD70
Issue code: 70nu2
4-16 DPL programming
4.8 Optimizing programs
In order for programs to run effectively, the following are recommended.
Integer variables
Use integer variables where possible, rather than floating-point variables.
The processing of a floating-point variable is 20 times slower than for an
integer variable. (See INT instruction in Chapter 7 Reference.)
Fixed-point arithmetic
To represent decimal places, use fixed-point arithmetic. For example, if a
resolution of .001.001 is required, let 11 be represented by 10001000. This allows
accuracy to be maintained throughout mathematical operations.
The output from an expression must then be corrected by a relevant
dividing factor.
Example a% = 1500 // “a% = 1.5”
b% = 2500 // “b% = 2.5”
c% = a% * b% // c% = 3750000
// Divide by 1000 to adjust c%
c% = c% / 1000 // “c% = 3750”
// To convert to the real value, we must divide by 1000 again
#1.21 = c% / 1000 // “c% = 3.75”
Temporary integer variables
Minimize the number of times parameters are accessed. Instead of
accessing a parameter repeatedly, use temporary integer variables if a
parameter value is needed more than once. The access time for a
parameter is 50 times greater than that for a variable.
Example IF #1.21 > 100 THEN
range% = 1
ELSEIF #1.21 > 200 THEN
range% = 2
ENDIF
This becomes:
temp% = #1.21
IF temp% > 100 THEN
range% = 1
ELSEIF temp% > 200 THEN
range% = 2
ENDIF