org.apfloat.internal
Class IntElementaryModMath

java.lang.Object
  extended byorg.apfloat.internal.IntElementaryModMath
Direct Known Subclasses:
IntModMath

public class IntElementaryModMath
extends java.lang.Object

Elementary modulo arithmetic functions for int data.

Modular addition and subtraction are trivial, when the modulus is less than 231 and overflow can be detected easily.

Modular multiplication is more complicated, and since it is usually the single most time consuming operation in the whole program execution, the very core of the Number Theoretic Transform (NTT), it should be carefully optimized.

The obvious (but not very efficient) algorithm for multiplying two ints and taking the remainder is

(int) ((long) a * b % modulus)

The first observation is that since the modulus is practically constant, it should be more efficient to calculate (once) the inverse of the modulus, and then subsequently multiply by the inverse modulus instead of dividing by the modulus.

The second observation is that to get the remainder of the division, we don't necessarily need the actual result of the division (we just want the remainder). So, we should discard the topmost 32 bits of the full 64-bit result whenever possible, to save a few operations.

The basic approach is to get some approximation of a * b / modulus. The approximation should be within +1 or -1 of the correct result. Then calculate a * b - approximateDivision * modulus to get the remainder. This calculation needs to use only the lowest 32 bits. As the modulus is less than 231 it is easy to detect the case when the approximate division was off by one (and the remainder is ±modulus off).

There are different algorithms to calculate the approximate division a * b / modulus. This implementation simply converts all the operands to double and performs the mulciplications. This requires that converting between integer types and floating point types is efficient. On some platforms this may not be true; in that case it can be more efficient to perform the multiplications using 64-bit integer arithmetic.

To simplify the operations, we calculate the inverse modulus as 1.0 / (modulus + 0.5). Since the modulus is assumed to be prime, and a double has more bits for precision than an int, the approximate result of a * b / modulus will always be either correct or one too small (but never one too big).

Version:
1.0.2
Author:
Mikko Tommila

Constructor Summary
IntElementaryModMath()
          Default constructor.
 
Method Summary
 int getModulus()
          Get the modulus.
 int modAdd(int a, int b)
          Modular addition.
 int modMultiply(int a, int b)
          Modular multiplication.
 int modSubtract(int a, int b)
          Modular subtraction.
 void setModulus(int modulus)
          Set the modulus.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

IntElementaryModMath

public IntElementaryModMath()
Default constructor.

Method Detail

modMultiply

public final int modMultiply(int a,
                             int b)
Modular multiplication.

Parameters:
a - First operand.
b - Second operand.
Returns:
a * b % modulus

modAdd

public final int modAdd(int a,
                        int b)
Modular addition.

Parameters:
a - First operand.
b - Second operand.
Returns:
(a + b) % modulus

modSubtract

public final int modSubtract(int a,
                             int b)
Modular subtraction. The result is always >= 0.

Parameters:
a - First operand.
b - Second operand.
Returns:
(a - b + modulus) % modulus

getModulus

public final int getModulus()
Get the modulus.

Returns:
The modulus.

setModulus

public final void setModulus(int modulus)
Set the modulus.

Parameters:
modulus - The modulus.