All Downloads are FREE. Search and download functionalities are using the official Maven repository.

cern.colt.matrix.tfloat.impl.DiagonalFloatMatrix2D Maven / Gradle / Ivy

Go to download

Parallel Colt is a multithreaded version of Colt - a library for high performance scientific computing in Java. It contains efficient algorithms for data analysis, linear algebra, multi-dimensional arrays, Fourier transforms, statistics and histogramming.

The newest version!
/*
Copyright (C) 1999 CERN - European Organization for Nuclear Research.
Permission to use, copy, modify, distribute and sell this software and its documentation for any purpose 
is hereby granted without fee, provided that the above copyright notice appear in all copies and 
that both that copyright notice and this permission notice appear in supporting documentation. 
CERN makes no representations about the suitability of this software for any purpose. 
It is provided "as is" without expressed or implied warranty.
 */
package cern.colt.matrix.tfloat.impl;

import java.util.concurrent.Callable;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.Future;

import cern.colt.matrix.tfloat.FloatMatrix1D;
import cern.colt.matrix.tfloat.FloatMatrix2D;
import edu.emory.mathcs.utils.ConcurrencyUtils;

/**
 * Diagonal 2-d matrix holding float elements. First see the package summary and javadoc tree view to get the broad picture.
 * 

* * @author Piotr Wendykier ([email protected]) */ public class DiagonalFloatMatrix2D extends WrapperFloatMatrix2D { private static final long serialVersionUID = 1L; /* * The non zero elements of the matrix. */ protected float[] elements; /* * Length of the diagonal */ protected int dlength; /* * An m-by-n matrix A has m+n-1 diagonals. Since the DiagonalFloatMatrix2D can have only one * diagonal, dindex is a value from interval [-m+1, n-1] that denotes which diagonal is stored. */ protected int dindex; /** * Constructs a matrix with a copy of the given values. values is * required to have the form values[row][column] and have exactly * the same number of columns in every row. Only the values on the main * diagonal, i.e. values[i][i] are used. *

* The values are copied. So subsequent changes in values are not * reflected in the matrix, and vice-versa. * * @param values * The values to be filled into the new matrix. * @param dindex * index of the diagonal. * @throws IllegalArgumentException * if * * for any 1 <= row < values.length: values[row].length != values[row-1].length || index < -rows+1 || index > columns - 1 * . */ public DiagonalFloatMatrix2D(float[][] values, int dindex) { this(values.length, values.length == 0 ? 0 : values[0].length, dindex); assign(values); } /** * Constructs a matrix with a given number of rows and columns. All entries * are initially 0. * * @param rows * the number of rows the matrix shall have. * @param columns * the number of columns the matrix shall have. * @param dindex * index of the diagonal. * @throws IllegalArgumentException * if size<0 (float)size > Integer.MAX_VALUE. */ public DiagonalFloatMatrix2D(int rows, int columns, int dindex) { super(null); try { setUp(rows, columns); } catch (IllegalArgumentException exc) { // we can hold rows*columns>Integer.MAX_VALUE cells ! if (!"matrix too large".equals(exc.getMessage())) throw exc; } if ((dindex < -rows + 1) || (dindex > columns - 1)) { throw new IllegalArgumentException("index is out of bounds"); } else { this.dindex = dindex; } if (dindex == 0) { dlength = Math.min(rows, columns); } else if (dindex > 0) { if (rows >= columns) { dlength = columns - dindex; } else { int diff = columns - rows; if (dindex <= diff) { dlength = rows; } else { dlength = rows - (dindex - diff); } } } else { if (rows >= columns) { int diff = rows - columns; if (-dindex <= diff) { dlength = columns; } else { dlength = columns + dindex + diff; } } else { dlength = rows + dindex; } } elements = new float[dlength]; } public FloatMatrix2D assign(final cern.colt.function.tfloat.FloatFunction function) { if (function instanceof cern.jet.math.tfloat.FloatMult) { // x[i] = mult*x[i] final float alpha = ((cern.jet.math.tfloat.FloatMult) function).multiplicator; if (alpha == 1) return this; if (alpha == 0) return assign(0); if (alpha != alpha) return assign(alpha); // the funny definition of isNaN(). This should better not happen. for (int j = dlength; --j >= 0;) { elements[j] *= alpha; } } else { for (int j = dlength; --j >= 0;) { elements[j] = function.apply(elements[j]); } } return this; } public FloatMatrix2D assign(float value) { for (int i = dlength; --i >= 0;) elements[i] = value; return this; } public FloatMatrix2D assign(final float[] values) { if (values.length != dlength) throw new IllegalArgumentException("Must have same length: length=" + values.length + " dlength=" + dlength); int nthreads = ConcurrencyUtils.getNumberOfThreads(); if ((nthreads > 1) && (dlength >= ConcurrencyUtils.getThreadsBeginN_2D())) { nthreads = Math.min(nthreads, dlength); Future[] futures = new Future[nthreads]; int k = dlength / nthreads; for (int j = 0; j < nthreads; j++) { final int firstRow = j * k; final int lastRow = (j == nthreads - 1) ? dlength : firstRow + k; futures[j] = ConcurrencyUtils.submit(new Runnable() { public void run() { for (int r = firstRow; r < lastRow; r++) { elements[r] = values[r]; } } }); } ConcurrencyUtils.waitForCompletion(futures); } else { for (int r = dlength; --r >= 0;) { elements[r] = values[r]; } } return this; } public FloatMatrix2D assign(final float[][] values) { if (values.length != rows) throw new IllegalArgumentException("Must have same number of rows: rows=" + values.length + "rows()=" + rows()); int r, c; if (dindex >= 0) { r = 0; c = dindex; } else { r = -dindex; c = 0; } for (int i = 0; i < dlength; i++) { if (values[i].length != columns) { throw new IllegalArgumentException("Must have same number of columns in every row: columns=" + values[r].length + "columns()=" + columns()); } elements[i] = values[r++][c++]; } return this; } public FloatMatrix2D assign(FloatMatrix2D source) { // overriden for performance only if (source == this) return this; // nothing to do checkShape(source); if (source instanceof DiagonalFloatMatrix2D) { DiagonalFloatMatrix2D other = (DiagonalFloatMatrix2D) source; if ((dindex != other.dindex) || (dlength != other.dlength)) { throw new IllegalArgumentException("source is DiagonalFloatMatrix2D with different diagonal stored."); } // quickest System.arraycopy(other.elements, 0, this.elements, 0, this.elements.length); return this; } else { return super.assign(source); } } public FloatMatrix2D assign(final FloatMatrix2D y, final cern.colt.function.tfloat.FloatFloatFunction function) { checkShape(y); if (y instanceof DiagonalFloatMatrix2D) { DiagonalFloatMatrix2D other = (DiagonalFloatMatrix2D) y; if ((dindex != other.dindex) || (dlength != other.dlength)) { throw new IllegalArgumentException("y is DiagonalFloatMatrix2D with different diagonal stored."); } if (function instanceof cern.jet.math.tfloat.FloatPlusMultSecond) { // x[i] = x[i] + alpha*y[i] final float alpha = ((cern.jet.math.tfloat.FloatPlusMultSecond) function).multiplicator; if (alpha == 0) { return this; // nothing to do } } final float[] otherElements = other.elements; int nthreads = ConcurrencyUtils.getNumberOfThreads(); if ((nthreads > 1) && (dlength >= ConcurrencyUtils.getThreadsBeginN_2D())) { nthreads = Math.min(nthreads, dlength); Future[] futures = new Future[nthreads]; int k = dlength / nthreads; for (int j = 0; j < nthreads; j++) { final int firstRow = j * k; final int lastRow = (j == nthreads - 1) ? dlength : firstRow + k; futures[j] = ConcurrencyUtils.submit(new Runnable() { public void run() { if (function instanceof cern.jet.math.tfloat.FloatPlusMultSecond) { // x[i] = x[i] + alpha*y[i] final float alpha = ((cern.jet.math.tfloat.FloatPlusMultSecond) function).multiplicator; if (alpha == 1) { for (int j = firstRow; j < lastRow; j++) { elements[j] += otherElements[j]; } } else { for (int j = firstRow; j < lastRow; j++) { elements[j] = elements[j] + alpha * otherElements[j]; } } } else if (function == cern.jet.math.tfloat.FloatFunctions.mult) { // x[i] = x[i] * y[i] for (int j = firstRow; j < lastRow; j++) { elements[j] = elements[j] * otherElements[j]; } } else if (function == cern.jet.math.tfloat.FloatFunctions.div) { // x[i] = x[i] / y[i] for (int j = firstRow; j < lastRow; j++) { elements[j] = elements[j] / otherElements[j]; } } else { for (int j = firstRow; j < lastRow; j++) { elements[j] = function.apply(elements[j], otherElements[j]); } } } }); } ConcurrencyUtils.waitForCompletion(futures); } else { if (function instanceof cern.jet.math.tfloat.FloatPlusMultSecond) { // x[i] = x[i] + alpha*y[i] final float alpha = ((cern.jet.math.tfloat.FloatPlusMultSecond) function).multiplicator; if (alpha == 1) { for (int j = dlength; --j >= 0;) { elements[j] += otherElements[j]; } } else { for (int j = dlength; --j >= 0;) { elements[j] = elements[j] + alpha * otherElements[j]; } } } else if (function == cern.jet.math.tfloat.FloatFunctions.mult) { // x[i] = x[i] * y[i] for (int j = dlength; --j >= 0;) { elements[j] = elements[j] * otherElements[j]; } } else if (function == cern.jet.math.tfloat.FloatFunctions.div) { // x[i] = x[i] / y[i] for (int j = dlength; --j >= 0;) { elements[j] = elements[j] / otherElements[j]; } } else { for (int j = dlength; --j >= 0;) { elements[j] = function.apply(elements[j], otherElements[j]); } } } return this; } else { return super.assign(y, function); } } public int cardinality() { int cardinality = 0; int nthreads = ConcurrencyUtils.getNumberOfThreads(); if ((nthreads > 1) && (dlength >= ConcurrencyUtils.getThreadsBeginN_2D())) { nthreads = Math.min(nthreads, dlength); Future[] futures = new Future[nthreads]; Integer[] results = new Integer[nthreads]; int k = dlength / nthreads; for (int j = 0; j < nthreads; j++) { final int firstRow = j * k; final int lastRow = (j == nthreads - 1) ? dlength : firstRow + k; futures[j] = ConcurrencyUtils.submit(new Callable() { public Integer call() throws Exception { int cardinality = 0; for (int r = firstRow; r < lastRow; r++) { if (elements[r] != 0) cardinality++; } return cardinality; } }); } try { for (int j = 0; j < nthreads; j++) { results[j] = (Integer) futures[j].get(); } cardinality = results[0]; for (int j = 1; j < nthreads; j++) { cardinality += results[j]; } } catch (ExecutionException ex) { ex.printStackTrace(); } catch (InterruptedException e) { e.printStackTrace(); } } else { for (int r = 0; r < dlength; r++) { if (elements[r] != 0) cardinality++; } } return cardinality; } public float[] elements() { return elements; } public boolean equals(float value) { float epsilon = cern.colt.matrix.tfloat.algo.FloatProperty.DEFAULT.tolerance(); for (int r = 0; r < dlength; r++) { float x = elements[r]; float diff = Math.abs(value - x); if ((diff != diff) && ((value != value && x != x) || value == x)) diff = 0; if (!(diff <= epsilon)) { return false; } } return true; } public boolean equals(Object obj) { if (obj instanceof DiagonalFloatMatrix2D) { DiagonalFloatMatrix2D other = (DiagonalFloatMatrix2D) obj; float epsilon = cern.colt.matrix.tfloat.algo.FloatProperty.DEFAULT.tolerance(); if (this == obj) return true; if (!(this != null && obj != null)) return false; final int rows = this.rows(); final int columns = this.columns(); if (columns != other.columns() || rows != other.rows()) return false; if ((dindex != other.dindex) || (dlength != other.dlength)) { return false; } float[] otherElements = other.elements; for (int r = 0; r < dlength; r++) { float x = elements[r]; float value = otherElements[r]; float diff = Math.abs(value - x); if ((diff != diff) && ((value != value && x != x) || value == x)) diff = 0; if (!(diff <= epsilon)) { return false; } } return true; } else { return super.equals(obj); } } public FloatMatrix2D forEachNonZero(final cern.colt.function.tfloat.IntIntFloatFunction function) { for (int j = dlength; --j >= 0;) { float value = elements[j]; if (value != 0) { elements[j] = function.apply(j, j, value); } } return this; } /** * Returns the length of the diagonal * * @return the length of the diagonal */ public int diagonalLength() { return dlength; } /** * Returns the index of the diagonal * * @return the index of the diagonal */ public int diagonalIndex() { return dindex; } public float[] getMaxLocation() { int location = 0; float maxValue = 0; int nthreads = ConcurrencyUtils.getNumberOfThreads(); if ((nthreads > 1) && (dlength >= ConcurrencyUtils.getThreadsBeginN_2D())) { nthreads = Math.min(nthreads, dlength); Future[] futures = new Future[nthreads]; float[][] results = new float[nthreads][2]; int k = dlength / nthreads; for (int j = 0; j < nthreads; j++) { final int firstRow = j * k; final int lastRow = (j == nthreads - 1) ? dlength : firstRow + k; futures[j] = ConcurrencyUtils.submit(new Callable() { public float[] call() throws Exception { int location = firstRow; float maxValue = elements[location]; float elem; for (int r = firstRow + 1; r < lastRow; r++) { elem = elements[r]; if (maxValue < elem) { maxValue = elem; location = r; } } return new float[] { maxValue, location, location }; } }); } try { for (int j = 0; j < nthreads; j++) { results[j] = (float[]) futures[j].get(); } maxValue = results[0][0]; location = (int) results[0][1]; for (int j = 1; j < nthreads; j++) { if (maxValue < results[j][0]) { maxValue = results[j][0]; location = (int) results[j][1]; } } } catch (ExecutionException ex) { ex.printStackTrace(); } catch (InterruptedException e) { e.printStackTrace(); } } else { maxValue = elements[0]; float elem; for (int r = 1; r < dlength; r++) { elem = elements[r]; if (maxValue < elem) { maxValue = elem; location = r; } } } int rowLocation; int columnLocation; if (dindex > 0) { rowLocation = location; columnLocation = location + dindex; } else if (dindex < 0) { rowLocation = location - dindex; columnLocation = location; } else { rowLocation = location; columnLocation = location; } return new float[] { maxValue, rowLocation, columnLocation }; } public float[] getMinLocation() { int location = 0; float minValue = 0; int nthreads = ConcurrencyUtils.getNumberOfThreads(); if ((nthreads > 1) && (dlength >= ConcurrencyUtils.getThreadsBeginN_2D())) { nthreads = Math.min(nthreads, dlength); Future[] futures = new Future[nthreads]; float[][] results = new float[nthreads][2]; int k = dlength / nthreads; for (int j = 0; j < nthreads; j++) { final int firstRow = j * k; final int lastRow = (j == nthreads - 1) ? dlength : firstRow + k; futures[j] = ConcurrencyUtils.submit(new Callable() { public float[] call() throws Exception { int location = firstRow; float minValue = elements[location]; float elem; for (int r = firstRow + 1; r < lastRow; r++) { elem = elements[r]; if (minValue > elem) { minValue = elem; location = r; } } return new float[] { minValue, location, location }; } }); } try { for (int j = 0; j < nthreads; j++) { results[j] = (float[]) futures[j].get(); } minValue = results[0][0]; location = (int) results[0][1]; for (int j = 1; j < nthreads; j++) { if (minValue > results[j][0]) { minValue = results[j][0]; location = (int) results[j][1]; } } } catch (ExecutionException ex) { ex.printStackTrace(); } catch (InterruptedException e) { e.printStackTrace(); } } else { minValue = elements[0]; float elem; for (int r = 1; r < dlength; r++) { elem = elements[r]; if (minValue > elem) { minValue = elem; location = r; } } } int rowLocation; int columnLocation; if (dindex > 0) { rowLocation = location; columnLocation = location + dindex; } else if (dindex < 0) { rowLocation = location - dindex; columnLocation = location; } else { rowLocation = location; columnLocation = location; } return new float[] { minValue, rowLocation, columnLocation }; } public float getQuick(int row, int column) { if (dindex >= 0) { if (column < dindex) { return 0; } else { if ((row < dlength) && (row + dindex == column)) { return elements[row]; } else { return 0; } } } else { if (row < -dindex) { return 0; } else { if ((column < dlength) && (row + dindex == column)) { return elements[column]; } else { return 0; } } } } public FloatMatrix2D like(int rows, int columns) { return new SparseFloatMatrix2D(rows, columns); } public FloatMatrix1D like1D(int size) { return new SparseFloatMatrix1D(size); } public void setQuick(int row, int column, float value) { if (dindex >= 0) { if (column < dindex) { //do nothing } else { if ((row < dlength) && (row + dindex == column)) { elements[row] = value; } else { // do nothing } } } else { if (row < -dindex) { //do nothing } else { if ((column < dlength) && (row + dindex == column)) { elements[column] = value; } else { //do nothing; } } } } public FloatMatrix1D zMult(FloatMatrix1D y, FloatMatrix1D z, float alpha, float beta, final boolean transposeA) { int rowsA = rows; int columnsA = columns; if (transposeA) { rowsA = columns; columnsA = rows; } boolean ignore = (z == null); if (z == null) z = new DenseFloatMatrix1D(rowsA); if (!(this.isNoView && y instanceof DenseFloatMatrix1D && z instanceof DenseFloatMatrix1D)) { return super.zMult(y, z, alpha, beta, transposeA); } if (columnsA != y.size() || rowsA > z.size()) throw new IllegalArgumentException("Incompatible args: " + ((transposeA ? viewDice() : this).toStringShort()) + ", " + y.toStringShort() + ", " + z.toStringShort()); if ((!ignore) && ((beta) != 1)) z.assign(cern.jet.math.tfloat.FloatFunctions.mult(beta)); DenseFloatMatrix1D zz = (DenseFloatMatrix1D) z; final float[] elementsZ = zz.elements; final int strideZ = zz.stride(); final int zeroZ = (int) z.index(0); DenseFloatMatrix1D yy = (DenseFloatMatrix1D) y; final float[] elementsY = yy.elements; final int strideY = yy.stride(); final int zeroY = (int) y.index(0); if (elementsY == null || elementsZ == null) throw new InternalError(); if (!transposeA) { if (dindex >= 0) { for (int i = dlength; --i >= 0;) { elementsZ[zeroZ + strideZ * i] += alpha * elements[i] * elementsY[dindex + zeroY + strideY * i]; } } else { for (int i = dlength; --i >= 0;) { elementsZ[-dindex + zeroZ + strideZ * i] += alpha * elements[i] * elementsY[zeroY + strideY * i]; } } } else { if (dindex >= 0) { for (int i = dlength; --i >= 0;) { elementsZ[dindex + zeroZ + strideZ * i] += alpha * elements[i] * elementsY[zeroY + strideY * i]; } } else { for (int i = dlength; --i >= 0;) { elementsZ[zeroZ + strideZ * i] += alpha * elements[i] * elementsY[-dindex + zeroY + strideY * i]; } } } return z; } protected FloatMatrix2D getContent() { return this; } }





© 2015 - 2025 Weber Informatics LLC | Privacy Policy