Code efficiency of Keil C51

The efficiency of C51 program compilation to generate assembly code is determined by many factors. For Keil C51, it is mainly affected by the following two factors: 

Section 1 The impact of storage mode

    The storage mode determines the storage space of the default variables, and the complexity of the assembly code for accessing variables in each space determines the code rate.

For example, if an integer variable i is placed in memory space 18H and 19H, the ++i operation is compiled into four statements:

INC 0x19

MOV A,0x19

JNZ 0x272D

INC 0x18

0x272D:

If it is placed in the external memory space 0000H, 0001H, the ++i operation is compiled into nine statements:

MOV DPTR, 0001

MOVX A, @DPTR

INC A

MOVX @ DPTR,A

JNz #5

MOV OPTR,#0000

MOVX A,@DPTR

INC A

MOVX @ DPTR,A

As for the statements after assembly, the code rate of operations on external memory is much lower than that of internal memory operations. The generated statements are more than twice that of memory, and there are a large number of such operations in the program, which shows the impact of storage mode on code rate.

Therefore, the principle of programming is

1. Select the storage mode from small-compact-large in sequence. If there are too many variables, choose large mode.

2. Even if the large mode is selected, it is best to store some commonly used local variables or variables that can be placed in memory in order to maximize the code rate of the program.

Section 2 The impact of program structure

The structural units of a program include modules, functions, etc. For the same function, if the structure is more complex, it involves more operations, variables, functional modules and functions, etc., and the code rate is naturally much lower than that of a program with good structure and simple code.

In addition, the program's running control statements are also a key factor affecting the code rate. For example, many compilers translate switch-case statements into very complex forms, and Keil C51 is no exception. Relatively simple Switch-case statements are compiled into jump instruction forms with a higher code rate, but for more complex Switch-Case, a system library function ?C?ICASE must be called for processing, which is very complicated.

For example, if(), while(), etc. statements also have relatively low code length, but after compilation, they are much higher than switch-case.

Therefore, it is recommended that designers use as few statements as possible such as switch-case to control the program structure in order to improve the code rate.

In addition to the above two points, other factors will also affect the code rate, such as:

Whether to pass parameters using registers, that is, whether the NOAREGS option is enabled

Whether to include debugging information: DEBUG option

Whether to include extended debugging information: BJECTEXTEND