HP 3000 Manuals HP 3000 Manuals
Overview This chapter contains guidance on programming techniques which will help you to develop efficient applications. The information in this chapter only applies to programs compiled to native code not intermediate code. The next section provides some guidance on determining when to use native code instead of intermediate code. Choosing Intermediate or Native Code If you want to run your code within an unlinked environment, you can choose whether you want the cob command to output intermediate or native code. See Chapter 2 , COBOL System Interface for details of how you can do this. For maximum performance, process your programs to native code, as native code programs usually execute more quickly than intermediate code programs. However, there is a qualification to this. A program that is I-O bound (that is, spends most of its time moving data to and from files and devices rather than performing arithmetic on it) derives relatively little benefit in speed from code generation. Only programs that are processor bound (that is, spend most of their time operating on data rather than transferring it) are likely to increase their run-time speed significantly as a result of code generation. Although in most instances native code gives better performance than intermediate code, you should be aware that the native code version of a program takes up more space than its intermediate code equivalent, as intermediate code is very compact. You must thus ensure that you have enough memory and disk space available to cope with the space overheads of native code. Native code cannot be animated, thus you must do all your testing on the intermediate code versions of your programs. If your program is divided into a main COBOL program and a number of subprograms called from the main program, some of the programs can be processed to intermediate code, others to native code. You can mix the two quite freely.