org.apfloat.internal.FloatMediumConvolutionStrategy Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of apfloat Show documentation
High performance arbitrary precision arithmetic library
The newest version!
/*
 * MIT License
 *
 * Copyright (c) 2002-2023 Mikko Tommila
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in all
 * copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */
package org.apfloat.internal;

import org.apfloat.ApfloatContext;
import org.apfloat.ApfloatRuntimeException;
import org.apfloat.spi.ConvolutionStrategy;
import org.apfloat.spi.DataStorageBuilder;
import org.apfloat.spi.DataStorage;

/**
 * Medium-length convolution strategy.
 * Performs a simple O(n²) multiplication when the size of one operand is relatively short.
 *
 * @version 1.9.0
 * @author Mikko Tommila
 */

public class FloatMediumConvolutionStrategy
    extends FloatBaseMath
    implements ConvolutionStrategy
{
    // Implementation notes:
    // - Assumes that the operands have been already truncated to match resultSize (the resultSize argument is ignored)
    // - This class probably shouldn't be converted to a single class using generics because there is some performance impact

    /**
     * Creates a convolution strategy using the specified radix.
     *
     * @param radix The radix that will be used.
     */

    public FloatMediumConvolutionStrategy(int radix)
    {
        super(radix);
    }

    @Override
    public DataStorage convolute(DataStorage x, DataStorage y, long resultSize)
        throws ApfloatRuntimeException
    {
        DataStorage shortStorage, longStorage;

        if (x.getSize() > y.getSize())
        {
            shortStorage = y;
            longStorage = x;
        }
        else
        {
            shortStorage = x;
            longStorage = y;
        }

        long shortSize = shortStorage.getSize(),
             longSize = longStorage.getSize(),
             size = shortSize + longSize;

        if (shortSize > Integer.MAX_VALUE)
        {
            throw new ApfloatInternalException("Too long shorter number, size = " + shortSize);
        }

        int bufferSize = (int) shortSize;

        ApfloatContext ctx = ApfloatContext.getContext();
        DataStorageBuilder dataStorageBuilder = ctx.getBuilderFactory().getDataStorageBuilder();
        DataStorage resultStorage = dataStorageBuilder.createDataStorage(size * Float.BYTES);
        resultStorage.setSize(size);

        DataStorage.Iterator src = longStorage.iterator(DataStorage.READ, longSize, 0),
                             dst = resultStorage.iterator(DataStorage.WRITE, size, 0),
                             tmpDst = new DataStorage.Iterator()                        // Cyclic iterator
                             {
                                 @Override
                                 public void next()
                                 {
                                     this.position++;
                                     this.position = (this.position == bufferSize ? 0 : this.position);
                                 }

                                 @Override
                                 public float getFloat()
                                 {
                                     return this.buffer[this.position];
                                 }

                                 @Override
                                 public void setFloat(float value)
                                 {
                                     this.buffer[this.position] = value;
                                 }

                                 private static final long serialVersionUID = 1L;

                                 private float[] buffer = new float[bufferSize];
                                 private int position = 0;
                             };

        for (long i = 0; i < longSize; i++)
        {
            DataStorage.Iterator tmpSrc = shortStorage.iterator(DataStorage.READ, shortSize, 0);        // Sub-optimal: this could be cyclic also

            float factor = src.getFloat(),          // Get one word of source data
                    carry = baseMultiplyAdd(tmpSrc, tmpDst, factor, 0, tmpDst, shortSize),
                    result = tmpDst.getFloat();       // Least significant word of the result

            dst.setFloat(result);     // Store one word of result

            tmpDst.setFloat(carry);   // Set carry from calculation as new last word in cyclic buffer

            tmpDst.next();              // Cycle buffer; current first word becomes last
            src.next();
            dst.next();
        }

        // Exhaust last words from temporary cyclic buffer and store them to result data
        for (int i = 0; i < bufferSize; i++)
        {
            float result = tmpDst.getFloat();
            dst.setFloat(result);

            tmpDst.next();
            dst.next();
        }

        return resultStorage;
    }

    private static final long serialVersionUID = -6697305140738370764L;
}