PL/I program efficiency

ABSTRACT

All PL/I Programmers should be aware of and genuinely concerned about PL/I Program efficiency. This paper addresses the following question:"How do you write a PL/I program which the PL/I Complier will reduce to the smallest and fastest executing machine language module?" The real world payoffs of knowing how the PL/I Optizing Compiler handles different syntactical representations of similar semantic relationships with respect to code generation and storage allocation can increase program runtime efficiency by many orders of magnitude thus saving many hours of CPU time. Research gathered by the authors in this paper deals with several areas within the PL/I programming realm: logical switches, bit string manipulation, character string manipulation, and multiway branching. Time and money did not permit for an exhaustive study of other areas of PL/I coding, but the results of these four areas have been rewarding and in some instances very surprising.

REFERENCES

• 1: IBM Corporate Data Processing Guidelines 3--101, Mar. 15, 1971
• 2: PL/I lang. Ref. GC33-0009- p. 427, Initial Item 16.
• 3: PL/I lang. Ref. GC-0009 - p. 429, Label Item 4.
• 4: PL/I lang. Ref. GC33-0009 - p. 427, Initial Item 5, p. 429 Label Item 6.
• 5: PL/I lang. Ref. GC33-0009 - p. 264, 1.c.
• 6: Author's note: A good question to ask at this point is, "How familiar is the PL/I Programming Community with the PL/I Preprocessor?"