US 6,263,429 B1 · Dynamic microcode for embedded processors

Abstract

A method of compressing programs, especially those used in embedded systems, is provided which allows greater program compression without significantly degrading system performance. The method provides: first, examining an entire program for sequences of lines of code, which may or may not constitute basic blocks; determining which sequences are identical or are identical except for a variation in a predetermined number of Elements within the sequence; designating and saving one uncompressed version of the identified sequences in memory as a specific microroutine, saving the Elements which differentiate the saved sequence from the various nearly identical sequences; and, assembling a version of the program consisting of original lines of code and microcalls. The microcall is a line of code which instructs a processor to implement a previously saved microroutine and provides an indication as to which Elements in the microroutine are to be replaced and where to find the substitute Elements. In a preferred embodiment, the Elements replaced are parameters such as register values. However, the present invention may be suitably modified to accommodate substitutions of any portion of a code including sequences and/or lines of instruction, operands, opcodes, or the like. Additionally, in an alternative embodiment, the determination of when the end of a microroutine code has been reached is facilitated by providing a single bit indicator table which corresponds to the lines of code stored as microroutines in memory, such that when the single bit reads a particular value, the last instruction of the microroutine has been reached.

Claims21

Dependent claims

A method of compressing a program utilized by a processor in controlling the operation of an associated device, wherein said program has at least one original sequence of lines of code and each line of code comprises at least one Element;

said method comprising the steps of:
a. identifying a first sequence of lines of code;
b. identifying a second sequence of lines of code wherein said first sequence and said second sequence are Substantially Similar;
c. identifying any varying Element in said second sequence which varies from a corresponding Element in said first sequence and wherein all non-identified Elements in said first sequence and said second sequence are identical;
d. assigning a microroutine designator to said first sequence;
e. assigning an Element designator to said varying Element, wherein said Element designator identifies that said corresponding Element is to be replaced by said varying Element;
f. assigning a microcall designator to each of said first and second sequences, wherein said microcall designator references said microroutine designator;
g. selecting a first sequence from a plurality of first sequences of lines of code which optimizes the compression of said program;
and
h. assembling a sequence of lines of code comprised of said lines of code of said program interleaved with said microcall designators, wherein said microcall designators replace those sequences of code Substantially Similar to those sequences of lines of code selected for optimization.

7 more elements

Dependencies are inferred from the claim text.

Dynamic microcode for embedded processors

Life of this patent

Dynamic microcode for embedded processors

Life of this patent

Abstract

Claims21