Post compile optimizer for linkable object code

• US 5,469,572 A
• Filed: 12/01/1992
• Issued: 11/21/1995
• Est. Priority Date: 12/01/1992
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: James M Taylor
• Inventors: Taylor, James M.
• Primary Examiner(s): Kriess, Kevin A.
• Assistant Examiner(s): BUTLER, DENNIS
• Application Number:
  US07/984,082
• Time in Patent Office: 1,085 Days
• Field of Search: 395/700, 364/280.5, 364/973
• US Class Current: 717/152
• CPC Class Codes: G06F 8/443 Optimisation
• CPC Class Codes: G06F 8/445 Exploiting fine grain parallelism, i.e. parallelism at instruction level run-time instruction scheduling G06F9/3836
• CPC Class Codes: G06F 8/54 Link editing before load time

Abstract

A system for processing a complete object code data set, to be linked into an executable program. The system features means for facilitating optimization analysis based upon the complete object code data set and also means for modifying the object code data based upon the optimization analysis results. The system allows for the specification of static instruction ranges that are to be isolated from the optimization process. Indirect and external process transfer of control destination ranges are identified in order to formulate the proper control flow information necessary to retain the optimized object code data integrity. The system enables the implementation of a class of optimizations which are supplemental to the standard class of optimizations performed during the initial object code generation process.

Claims

1. A system for generating data representing an optimized object code set for a data processing system from data representing a complete object code set, comprising:

(a) a memory area for storing data;

(b) means for forming a plurality of data storage areas within said memory area;

(c) first converter means for converting said data representing the complete object code set into data representing a first plurality of instruction sequences to be stored in a first data storage area, a first plurality of symbol and relocation data to be stored in a second data storage area and a plurality of static data to be stored in a third data storage area;

(d) second converter means coupled with the first and second data storage areas for converting said data representing said first plurality of instruction sequences into data representing a plurality of optimized instruction sequences to be stored in a fourth data storage area;

(e) adjuster means coupled with the second storage area for adjusting said first symbol and relocation data into a plurality of adjusted symbol and relocation data to be stored in a fifth data storage area such that said adjusted symbol and relocation data corresponds to said data representing the plurality of optimized instruction sequences; and

(f) data generating means coupled to said third, fourth and fifth storage areas for generating said data representing the optimized object code set from said static data, said data representing the plurality of optimized instruction sequences and said plurality of adjusted symbol and relocation data.

2. The system of claim 1 wherein said data representing the complete object code set includes a plurality of static instruction ranges.

3. The system of claim 2 wherein said data representing the complete object code set includes a plurality of statically referenced procedure ranges such that said plurality of statically referenced procedure ranges correspond to said static instruction ranges comprising:

first identifying means for designating said plurality of statically referenced procedure ranges thereby allowing for the formulation of accurate control flow information corresponding to said data representing the complete object code set.

4. The system of claim 2 wherein said static instruction ranges are indicated in a separate storage area.

5. The system of claim 4 wherein said static instruction ranges are indicated as encompassing an object code module.

6. The system of claim 4 wherein said static instruction ranges are indicated as encompassing an object code library.

7. The system of claim 2 wherein said static instruction ranges are indicated within said data representing the complete object code set.

8. The system of claim 7 wherein said static instruction ranges are indicated as encompassing a instruction segment offset begin and end address boundary.

9. The system of claim i wherein said data representing the first plurality of instruction sequences contains data representing a plurality of indirect transfer of control instructions.

10. The system of claim 9 wherein said data representing the complete object code set includes a plurality of indirect transfer of control destination procedure ranges such that said plurality of indirect transfer of control destination procedure ranges correspond to said data representing the plurality of indirect transfer of control instructions comprising:

second identifying means for designating said plurality of indirect transfer of control destination procedure ranges corresponding to said data representing the plurality of indirect transfer of control instructions thereby allowing for the formulation of accurate control flow information corresponding to said data representing the complete object code set.

11. The system of claim i wherein said complete object code data set contains a plurality of external process transfer of control destination ranges comprising:

third identifying means for designating said external process transfer of control destination ranges within said complete object code data set thereby allowing for the formulation of accurate control flow information corresponding to said data representing the complete object code set.

12. The system of claim i wherein said data representing the first plurality of instruction sequences contains data representing a plurality of limited displacement instructions.

13. The system of claim 12 wherein said data representing the plurality of limited displacement instructions contains a plurality of adjusted displacement range boundary violations comprising:

substitution means for converting selected data within said data representing the plurality of limited displacement instructions whereby said selected data contains said plurality of adjusted displacement range boundary violations into data representing a plurality of logically equivalent instruction sequences such that said logically equivalent instruction sequences include sufficiently sized adjusted displacement ranges.

14. A method for generating data representing an optimized object code set for a data processing system from data representing a complete object code set, comprising the steps of:

(a) forming a plurality of data storage areas within a memory area;

(b) converting said data representing the complete object code set into data representing a first plurality of instruction sequences to be stored in a first data storage area, a first plurality of symbol and relocation data to be stored in a second data storage area and a plurality of static data to be stored in a third data storage area;

(c) converting said data representing said first plurality of instruction sequences into data representing a plurality of optimized instruction sequences to be stored in a fourth data storage area;

(d) adjusting said first symbol and relocation data into a plurality of adjusted symbol and relocation data to be stored in a fifth data storage area such that said adjusted symbol and relocation data corresponds to said data representing the plurality of optimized instruction sequences; and

(e) generating said data representing the optimized object code set from said static data, said data representing the plurality of optimized instruction sequences and said plurality of adjusted symbol and relocation data.

15. The method of claim 14 wherein said data representing the complete object code set includes a plurality of static instruction ranges.

16. The method of claim 15 wherein said data representing the complete object code set includes a plurality of statically referenced procedure ranges such that said plurality of statically referenced procedure ranges correspond to said static instruction ranges further comprising the steps of:

designating said plurality of statically referenced procedure ranges thereby allowing for the formulation of accurate control flow information corresponding to said data representing the complete object code set.

17. The method of claim 15 wherein said static instruction ranges are indicated in a separate storage area.

18. The method of claim 17 wherein said static instruction ranges are indicated as encompassing an object code module.

19. The method of claim 17 wherein said static instruction ranges are indicated as encompassing an object code library.

20. The method of claim 15 wherein said static instruction ranges are indicated within said data representing the complete object code set.

21. The method of claim 20 wherein said static instruction ranges are indicated as encompassing a instruction segment offset begin and end address boundary.

22. The method of claim 14 wherein said data representing the first plurality of instruction sequences contains data representing a plurality of indirect transfer of control instructions.

23. The method of claim 22 wherein said data representing the complete object code set includes a plurality of indirect transfer of control destination procedure ranges such that said plurality of indirect transfer of control destination procedure ranges correspond to said data representing the plurality of indirect transfer of control instructions further comprising the steps of:

designating said plurality of indirect transfer of control destination procedure ranges corresponding to said data representing the plurality of indirect transfer of control instructions thereby allowing for the formulation of accurate control flow information corresponding to said data representing the complete object code set.

24. The method of claim 14 wherein said complete object code data set contains a plurality of external process transfer of control destination ranges further comprising the steps of:

designating said external process transfer of control destination ranges within said complete object code data set thereby allowing for the formulation of accurate control flow information corresponding to said data representing the complete object code set.

25. The method of claim 14 wherein said data representing the first plurality of instruction sequences contains data representing a plurality of limited displacement instructions.

26. The method of claim 25 wherein said data representing the plurality of limited displacement instructions contains a plurality of adjusted displacement range boundary violations further comprising the steps of:

converting selected data within said data representing the plurality of limited displacement instructions whereby said selected data contains said plurality of adjusted displacement range boundary violations into data representing a plurality of logically equivalent instruction sequences such that said logically equivalent instruction sequences include sufficiently sized adjusted displacement ranges.