Compressing microcode

• US 7,095,342 B1
• Filed: 03/31/2005
• Issued: 08/22/2006
• Est. Priority Date: 03/31/2005
• Status: Active Grant

• Current Assignee: Sony Corporation Of America (Sony Group Corp.)
• Original Assignee: Intel Corporation
• Inventors: Wu, Youfeng, Hum, Herbert, Kim, Sangwook, Breternitz, Mauricio Jr.
• Primary Examiner(s): Mai, Lam T.
• Application Number:
  US11/096,152
• Time in Patent Office: 509 Days
• Field of Search: 341 50- 70, 341/87, 341/102, 341/106, 714/42, 714/718
• US Class Current: 341/67
• CPC Class Codes: H03M 7/3084 using adaptive string matching, e.g. the Lempel-Ziv method

Abstract

In one embodiment, the present invention includes a method to compress data stored in a memory to reduce size and power consumption. The method includes segmenting each word of a code portion into multiple fields, forming tables having unique entries for each of the fields, and assigning a pointer to each of the unique entries in each of the tables. Other embodiments are described and claimed.

Claims

1. A method comprising:

segmenting each word of a code portion into a plurality of fields, wherein the code portion comprises microcode to be stored in a read only memory;

forming tables for each of the plurality of fields, each table having a plurality of unique entries; and

assigning one of a plurality of pointers to each of the plurality of unique entries in each of the tables.

2. The method of claim 1, further comprising forming the tables to minimize loading in a memory in which the tables are stored.

3. The method of claim 2, further comprising reducing power consumption of the memory via minimizing the loading.

4. The method of claim 2, further comprising assigning locations in the tables addressable by a lesser number of first bit states to the unique entries that occur most frequently in the code portion.

5. The method of claim 1, further comprising assigning locations in the tables in a generally decreasing order of occurrence in codewords of the code portion.

6. The method of claim 1, further comprising inverting a unique entry if the uninverted unique entry has more first bit states than second bit states.

7. The method of claim 6, further comprising storing a transformation indicator with the inverted unique entry.

8. An apparatus comprising:

a memory including a plurality of tables, each of the plurality of tables including unique entries for a field within microcode words; and

the memory including an array to store at least one sequence of pointers to access the plurality of tables.

9. The apparatus of claim 8, wherein the unique entries are to be stored in generally descending order of occurrence in the microcode words.

10. The apparatus of claim 9, wherein the generally descending order is to lower a number of first bit states in the array.

11. The apparatus of claim 10, wherein the lowered number of first bit states is to reduce power consumption of the memory.

12. The apparatus of claim 8, wherein the memory comprises a read only memory (ROM) of a processor.

13. The apparatus of claim 8, further comprising a logic operator coupled to at least one of the plurality of tables to perform a transformation on the unique entries according to a value of a transformation indicator.

14. The apparatus of claim 8, wherein the at least one sequence of pointers comprises a fixed length.

15. An article comprising a machine-readable storage medium that stores data representing an integrated circuit comprising:

a memory having a plurality of tables each including unique entries for a field of a microcode; and

the memory having an array to store a compressed representation of each field for words of the microcode.

16. The article of claim 15, wherein the integrated circuit comprises a processor.

17. The article of claim 16, wherein the memory comprises a read only memory (ROM) of the processor.

18. An article comprising a machine-readable storage medium that stores data representing an integrated circuit comprising:

a memory having a plurality of tables each including unique entries for a field of a microcode and a logic operator coupled to one or more of the plurality of tables to perform an operation on an output of the one or more of the plurality of tables according to an indicator; and

the memory having an array to store a compressed representation of each field for words of the microcode.

19. The article of claim 18, wherein the memory comprises a read only memory (ROM) of a processor.

20. A system comprising:

a first memory array to store pointer sequences;

a second memory array having a plurality of tables, each of the plurality of tables including unique entries for microcode words; and

a dynamic random access memory (DRAM) coupled to the first memory array and the second memory array.

21. The system of claim 20, further comprising a processor, wherein the first memory array and the second memory array comprise microcode storage of the processor.

22. The system of claim 20, wherein the unique entries are to be stored in generally descending order of occurrence in the microcode words.

23. The system of claim 22, wherein the first memory array is to store a lowered number of first bit states via the generally descending order.

24. The system of claim 22, wherein the generally descending order is to reduce power consumption of the first memory array.

25. The system of claim 20, further comprising a logic operator coupled to at least one of the plurality of tables to perform a transformation on the unique entries according to a value of a transformation indicator.

26. The system of claim 20, further comprising a plurality of logic operators coupled to at least one of the plurality of tables to perform transformations on a plurality of segments of the unique entries based upon a transformation indicator.

27. A method comprising:

segmenting each word of a code portion into a plurality of fields;

forming tables for each of the plurality of fields, each table having a plurality of unique entries, wherein locations in the tables are assigned in a generally decreasing order of occurrence in codewords of the code portion; and

assigning one of a plurality of pointers to each of the plurality of unique entries in each of the tables.

28. The method of claim 27, further comprising inverting a unique entry if the uninverted unique entry has more first bit states than second bit states.

29. The method of claim 28, further comprising storing a transformation indicator with the inverted unique entry.