Granted Patent B1

US 7,237,228 · App. 10/008,952 · Granted Jun 26, 2007

Abstract

Claims (49)

1. A computer-implemented method for processing numerical values in a computer program executable on a computer system, comprising:

encapsulating in a large-integer datatype, large-integer data and associated large-integer operators, wherein the large-integer data has runtime expandable precision and maximum precision is limited only by system memory availability;

overloading language-provided arithmetic, logical, and type conversion operators with the large-integer operators that operate on large-integer variables in combination with other datatypes, and programmed usage of a variable of the large-integer datatype is equivalent to and interoperable with a variable of a system-defined integral datatype;

establishing a plurality of available storage nodes available for allocation to large-integer data;

allocating a subset of the plurality of available storage nodes for a large-integer variable, determining a number of storage nodes to be allocated as a function of a size of the large-integer variable, and storing in each node of the subset a subset of bit values that represent a numerical value in the allocated plurality of storage nodes and forming a linked list of the allocated plurality of storage nodes;

determining a total number of available storage nodes available for allocation to large-integer data; allocating memory for a first number of available storage nodes, responsive to the total number being less than first threshold value, and establishing the first number of available storage nodes; and

removing from the plurality of available storage nodes, responsive to the total number being greater than a second threshold value, a second number of storage nodes, and deallocating memory for the second number of storage nodes.

2. The method of claim 1 , further comprising converting a character string into large-integer data in response to a constant definition statement.

3. The method of claim 2 , further comprising converting large-integer data to and from a character string for input, output, and serialization.

4. The method of claim 1 , further comprising:

converting input data from language-provided input functions to large-integer data; and

converting large-integer data to a format compatible with language-provided output functions.

5. The method of claim 1 , further comprising allocating a selected number of bits for each storage node in response to a program-specified parameter.

6. The method of claim 1 , further comprising:

maintaining a set of available storage nodes that are not allocated to any large-integer variable;

allocating a storage node from the set of available storage nodes to a large-integer variable while performing a large-integer operation that generates a numerical value and stores the numerical value in the variable, if a number of bit values required to represent the numerical value exceeds storage available in storage nodes allocated to the large-integer variable; and

returning to the set of available storage nodes a storage node allocated to a large-integer variable while performing a large-integer operation that generates a numerical value for storage in the variable, if a number of bit values required to represent the numerical value is less than storage available in storage nodes allocated to the variable.

7. The method of claim 6 , further comprising overloading language-provided memory allocation and deallocation operators with large-integer operators that allocate and deallocate storage nodes.

8. The method of claim 1 , further comprising, responsive to a large-integer divide operation specifying an input dividend and divisor:

identifying a set of most-significant bits of the dividend and a set of least-significant bits of the dividend;

recursively performing a large-integer divide operation using the set of most-significant bits as the input dividend, and returning a quotient and a remainder;

finding a lower-part dividend as a function of the remainder and the set of least-significant bits;

recursively performing a large-integer divide operation using the lower-part dividend; and

concurrently solving for the quotient and the remainder.

9. The method of claim 8 , further comprising identifying an optimal set of most-significant bits of the dividend and a set of least-significant bits of the dividend as a function of a number of bits that represent the dividend and a number of bits that represent the divisor.

10. The method of claim 9 , further comprising identifying an optimal set of most-significant bits of the dividend and a set of least-significant bits of the dividend as a function of one-half a difference between the number of bits that represent the dividend and the number of bits that represent the divisor.

11. The method of claim 1 , further comprising emulating fixed-bit arithmetic on variables of the large-integer data type.

12. The method of claim 1 , further comprising transferring data associated with temporary variables of the large-integer datatype by moving pointers to the data.

13. The method of claim 1 , further comprising

encapsulating in a large-floating-point datatype, large-floating-point data and associated operators, wherein the large-floating-point data has runtime expandable precision and maximum precision is limited only by system memory availability; and

overloading language-provided arithmetic, logical, and type conversion operators for floating-point data with the large-floating-point datatype operators that operate on large-floating-point variables in combination with other datatypes, and programmed usage of a variable of the large-floating-point datatype is equivalent to and interoperable with a variable of a system-defined floating-point datatype.

14. The method of claim 1 , further comprising

encapsulating in a large-rational datatype, large-rational data and associated operators, wherein the large-rational data has runtime expandable precision and maximum precision is limited only by system memory availability; and

overloading language-provided arithmetic, logical, and type conversion operators for rational data with the large-rational datatype operators that operate on large-rational variables in combination with other datatypes, and programmed usage of a variable of the large-rational datatype is equivalent to and interoperable with a variable of a system-defined rational datatype.

15. An apparatus for processing numerical values in a computer program executable on a computer system, comprising:

means for encapsulating in a large-integer datatype, large-integer data and associated large-integer operators, wherein the large-integer data has runtime expandable precision and maximum precision is limited only by system memory availability;

means for overloading language-provided arithmetic, logical, and type conversion operators for integers with the large-integer datatype operators that operate on large-integer variables in combination with other datatypes, and programmed usage of a variable of the large-integer datatype is equivalent to and interoperable with a variable of a system-defined integral datatype;

means for establishing a plurality of allocable storage nodes available for allocation to large-integer data;

means for allocating, for a large-integer variable, a subset of the plurality of allocable storage nodes, determining a number of storage nodes to be allocated as a function of a size of the large-integer variable; and

means for storing in each node of the subset a subset of bit values that represent a numerical value in the allocated plurality of storage nodes and forming a linked list of the allocated plurality of storage nodes

means for determining a total number of available storage nodes available for allocation to large-integer data;

means for allocating memory for a first number of available storage nodes, responsive to the total number being less than first threshold value, and establishing the first number of available storage nodes; and

means for removing from the plurality of available storage nodes, responsive to the total number being greater than a second threshold value, a second number of storage nodes, and deallocating memory fro the second number of storage nodes.

16. The apparatus of claim 15 , further comprising

means for encapsulating in a large-floating-point datatype, large-floating-point data and associated operators, wherein the large-floating-point data has runtime expandable precision and maximum precision is limited only by system memory availability; and

means for overloading language-provided arithmetic, logical, and type conversion operators for floating-point data with the large-floating-point datatype operators that operate on large-floating-point variables in combination with other datatypes, and programmed usage of a variable of the large-floating-point datatype is equivalent to and interoperable with a variable of a system-defined floating-point datatype.

17. The apparatus of claim 15 , further comprising

means for encapsulating in a large-rational datatype, large-rational data and associated operators, wherein the large-rational data has runtime expandable precision and maximum precision is limited only by system memory availability; and

means for overloading language-provided arithmetic, logical, and type conversion operators for rational data with the large-rational datatype operators that operate on large-rational variables in combination with other datatypes, and programmed usage of a variable of the large-rational datatype is equivalent to and interoperable with a variable of a system-defined rational datatype.

Assignments (13)