Related
Is there a way to scale down with highest quality a font which is png image in opengl at startup? I tried gluScaleImage but there are many artefacts. Is there anything that uses lanczos or something like that? I don't want to write a shader or anything that does the scaling runtime.
This is based on an algorithm, I copied decades ago from the German c't Magazin, and still use it from time to time for similar issues like described by OP.
bool scaleDown(
const Image &imgSrc,
Image &imgDst,
int w, int h,
int align)
{
const int wSrc = imgSrc.w(), hSrc = imgSrc.h();
assert(w > 0 && w <= wSrc && h > 0 && h <= hSrc);
// compute scaling factors
const double sx = (double)wSrc / (double)w;
const double sy = (double)hSrc / (double)h;
const double sxy = sx * sy;
// prepare destination image
imgDst.resize(w, h, (w * 3 + align - 1) / align * align);
// cache some data
const uint8 *const dataSrc = imgSrc.data();
const int bPRSrc = imgSrc.bPR();
// perform scaling
for (int y = 0; y < h; ++y) {
const double yStart = sy * y;
const double yEnd = std::min(sy * (y + 1), (double)hSrc);
const int yStartInt = (int)yStart;
const int yEndInt = (int)yEnd - (yEndInt == yEnd);
const double tFrm = 1 + yStartInt - yStart, bFrm = yEnd - yEndInt;
for (int x = 0; x < w; ++x) {
const double xStart = sx * x;
const double xEnd = std::min(sx * (x + 1), (double)wSrc);
const int xStartInt = (int)xStart;
const int xEndInt = (int)xEnd - (xEndInt == xEnd);
double lFrm = 1 + xStartInt - xStart, rFrm = xEnd - xEndInt;
double pixel[3] = { 0.0, 0.0, 0.0 }; // values of target pixel
for (int i = yStartInt; i <= yEndInt; ++i) {
int jData = i * bPRSrc + xStartInt * 3;
for (int j = xStartInt; j <= xEndInt; ++j) {
double pixelAdd[3];
for (int k = 0; k < 3; ++k) {
pixelAdd[k] = (double)dataSrc[jData++] / sxy;
}
if (j == xStartInt) {
for (int k = 0; k < 3; ++k) pixelAdd[k] *= lFrm;
} else if (j == xEndInt) {
for (int k = 0; k < 3; ++k) pixelAdd[k] *= rFrm;
}
if (i == yStartInt) {
for (int k = 0; k < 3; ++k) pixelAdd[k] *= tFrm;
} else if (i == yEndInt) {
for (int k = 0; k < 3; ++k) pixelAdd[k] *= bFrm;
}
for (int k = 0; k < 3; ++k) pixel[k] += pixelAdd[k];
}
}
imgDst.setPixel(x, y,
(uint8)pixel[0], (uint8)pixel[1], (uint8)pixel[2]);
}
}
// done
return true;
}
If I got it right, this implements a bilinear interpolation.
I don't dare to call it a Minimal Complete Verifiable Example although this is what I intended to do.
The complete sample application:
A simplified class Image
image.h:
#ifndef IMAGE_H
#define IMAGE_H
#include <vector>
// convenience type for bytes
typedef unsigned char uint8;
// image helper class
class Image {
private: // variables:
int _w, _h; // image size
size_t _bPR; // bytes per row
std::vector<uint8> _data; // image data
public: // methods:
// constructor.
Image(): _w(0), _h(0), _bPR(0) { }
// destructor.
~Image() = default;
// copy constructor.
Image(const Image&) = delete; // = default; would work as well.
// copy assignment.
Image& operator=(const Image&) = delete; // = default; would work as well.
// returns width of image.
int w() const { return _w; }
// returns height of image.
int h() const { return _h; }
// returns bytes per row.
size_t bPR() const { return _bPR; }
// returns pointer to image data.
const uint8* data(
int y = 0) // row number
const {
return &_data[y * _bPR];
}
// returns data size (in bytes).
size_t size() const { return _data.size(); }
// clears image.
void clear();
// resizes image.
uint8* resize( // returns allocated buffer
int w, // image width
int h, // image height
int bPR); // bytes per row
// returns pixel.
int getPixel(
int x, // column
int y) // row
const;
// sets pixel.
void setPixel(
int x, // column
int y, // row
uint8 r, uint8 g, uint8 b);
// sets pixel.
void setPixel(
int x, // column
int y, // row
int value) // RGB value
{
setPixel(x, y, value & 0xff, value >> 8 & 0xff, value >> 16 & 0xff);
}
};
// helper functions:
inline uint8 getR(int value) { return value & 0xff; }
inline uint8 getG(int value) { return value >> 8 & 0xff; }
inline uint8 getB(int value) { return value >> 16 & 0xff; }
#endif // IMAGE_H
image.cc:
#include <cassert>
#include "image.h"
// clears image.
void Image::clear()
{
_data.clear(); _w = _h = _bPR = 0;
}
// allocates image data.
uint8* Image::resize( // returns allocated buffer
int w, // image width
int h, // image height
int bPR) // bits per row
{
assert(w >= 0 && 3 * w <= bPR);
assert(h >= 0);
_w = w; _h = h; _bPR = bPR;
const size_t size = h * bPR;
_data.resize(size);
return _data.data();
}
// returns pixel.
int Image::getPixel(
int x, // column
int y) // row
const {
assert(x >= 0 && x < _w);
assert(y >= 0 && y < _h);
const size_t offs = y * _bPR + 3 * x;
return _data[offs + 0]
| _data[offs + 1] << 8
| _data[offs + 2] << 16;
}
// sets pixel.
void Image::setPixel(
int x, // column
int y, // row
uint8 r, uint8 g, uint8 b) // R, G, B values
{
assert(x >= 0 && x < _w);
assert(y >= 0 && y < _h);
const size_t offs = y * _bPR + 3 * x;
_data[offs + 0] = r;
_data[offs + 1] = g;
_data[offs + 2] = b;
}
Image Scaling
imageScale.h:
#ifndef IMAGE_SCALE_H
#define IMAGE_SCALE_H
#include "image.h"
/* scales an image to a certain width and height.
*
* Note:
* imgSrc and imgDst may not be identical.
*/
bool scaleTo( // returns true if successful
const Image &imgSrc, // source image
Image &imgDst, // destination image
int w, int h, // destination width and height
int align = 4); // row alignment
/* scales an image about a certain horizontal/vertical scaling factor.
*
* Note:
* imgSrc and imgDst may not be identical.
*/
inline bool scaleXY( // returns true if successful
const Image &imgSrc, // source image
Image &imgDst, // destination image
double sX, // horizontal scaling factor (must be > 0 but not too large)
double sY, // vertical scaling factor (must be > 0 but not too large)
int align = 4) // row alignment
{
return sX > 0.0 && sY > 0.0
? scaleTo(imgSrc, imgDst,
(int)(sX * imgSrc.w()), (int)(sY * imgSrc.h()), align)
: false;
}
/* scales an image about a certain scaling factor.
*
* Note:
* imgSrc and imgDst may not be identical.
*/
inline bool scale( // returns true if successful
const Image &imgSrc, // source image
Image &imgDst, // destination image
double s, // scaling factor (must be > 0 but not too large)
int align = 4) // row alignment
{
return scaleXY(imgSrc, imgDst, s, s, align);
}
#endif // IMAGE_SCALE_H
imageScale.cc:
#include <cassert>
#include <algorithm>
#include "imageScale.h"
namespace {
template <typename VALUE>
VALUE clip(VALUE value, VALUE min, VALUE max)
{
return value < min ? min : value > max ? max : value;
}
bool scaleDown(
const Image &imgSrc,
Image &imgDst,
int w, int h,
int align)
{
const int wSrc = imgSrc.w(), hSrc = imgSrc.h();
assert(w > 0 && w <= wSrc && h > 0 && h <= hSrc);
// compute scaling factors
const double sx = (double)wSrc / (double)w;
const double sy = (double)hSrc / (double)h;
const double sxy = sx * sy;
// prepare destination image
imgDst.resize(w, h, (w * 3 + align - 1) / align * align);
// cache some data
const uint8 *const dataSrc = imgSrc.data();
const int bPRSrc = imgSrc.bPR();
// perform scaling
for (int y = 0; y < h; ++y) {
const double yStart = sy * y;
const double yEnd = std::min(sy * (y + 1), (double)hSrc);
const int yStartInt = (int)yStart;
const int yEndInt = (int)yEnd - (yEndInt == yEnd);
const double tFrm = 1 + yStartInt - yStart, bFrm = yEnd - yEndInt;
for (int x = 0; x < w; ++x) {
const double xStart = sx * x;
const double xEnd = std::min(sx * (x + 1), (double)wSrc);
const int xStartInt = (int)xStart;
const int xEndInt = (int)xEnd - (xEndInt == xEnd);
double lFrm = 1 + xStartInt - xStart, rFrm = xEnd - xEndInt;
double pixel[3] = { 0.0, 0.0, 0.0 }; // values of target pixel
for (int i = yStartInt; i <= yEndInt; ++i) {
int jData = i * bPRSrc + xStartInt * 3;
for (int j = xStartInt; j <= xEndInt; ++j) {
double pixelAdd[3];
for (int k = 0; k < 3; ++k) {
pixelAdd[k] = (double)dataSrc[jData++] / sxy;
}
if (j == xStartInt) {
for (int k = 0; k < 3; ++k) pixelAdd[k] *= lFrm;
} else if (j == xEndInt) {
for (int k = 0; k < 3; ++k) pixelAdd[k] *= rFrm;
}
if (i == yStartInt) {
for (int k = 0; k < 3; ++k) pixelAdd[k] *= tFrm;
} else if (i == yEndInt) {
for (int k = 0; k < 3; ++k) pixelAdd[k] *= bFrm;
}
for (int k = 0; k < 3; ++k) pixel[k] += pixelAdd[k];
}
}
imgDst.setPixel(x, y,
(uint8)pixel[0], (uint8)pixel[1], (uint8)pixel[2]);
}
}
// done
return true;
}
bool scaleUp(
const Image &imgSrc,
Image &imgDst,
int w, int h,
int align)
{
const int wSrc = imgSrc.w(), hSrc = imgSrc.h();
assert(w && w >= wSrc && h && h >= hSrc);
// compute scaling factors
const double sx = (double)wSrc / (double)w;
const double sy = (double)hSrc / (double)h;
// prepare destination image
imgDst.resize(w, h, (w * 3 + align - 1) / align * align);
// cache some data
const uint8 *const dataSrc = imgSrc.data();
const int bPRSrc = imgSrc.bPR();
// perform scaling
for (int y = 0; y < h; ++y) {
const double yStart = sy * y;
const double yEnd = std::min(sy * (y + 1), (double)hSrc - 1);
const int yStartInt = (int)yStart;
const int yEndInt = (int)yEnd;
if (yStartInt < yEndInt) {
const double bFract = clip((double)((yEnd - yEndInt) / sy), 0.0, 1.0);
const double tFract = 1.0 - bFract;
for (int x = 0; x < w; ++x) {
const double xStart = sx * x;
const double xEnd = std::min(sx * (x + 1), (double)wSrc - 1);
const int xStartInt = (int)xStart, xEndInt = (int)xEnd;
double pixel[4];
if (xStartInt < xEndInt) {
const double rFract
= clip((double)((xEnd - xEndInt) / sx), 0.0, 1.0);
const double lFract = 1.0 - rFract;
int jData = yStartInt * bPRSrc + xStartInt * 3;
for (int k = 0; k < 3; ++k) {
pixel[k] = tFract * lFract * dataSrc[jData++];
}
for (int k = 0; k < 3; ++k) {
pixel[k] += tFract * rFract * dataSrc[jData++];
}
jData = yEndInt * bPRSrc + xStartInt * 3;
for (int k = 0; k < 3; ++k) {
pixel[k] += bFract * lFract *dataSrc[jData++];
}
for (int k = 0; k < 3; ++k) {
pixel[k] += bFract * rFract *dataSrc[jData++];
}
} else {
int jData = yStartInt * bPRSrc + xStartInt * 3;
for (int k = 0; k < 3; ++k) {
pixel[k] = tFract * dataSrc[jData++];
}
jData = yEndInt * bPRSrc + xStartInt * 3;
for (int k = 0; k < 3; ++k) {
pixel[k] += bFract * dataSrc[jData++];
}
}
imgDst.setPixel(x, y,
(uint8)pixel[0], (uint8)pixel[1], (uint8)pixel[2]);
}
} else {
for (int x = 0; x < w; ++x) {
const double xStart = sx * x;
const double xEnd = std::min(sx * (x + 1), (double)wSrc - 1);
const int xStartInt = (int)xStart, xEndInt = (int)xEnd;
double pixel[3];
if (xStartInt < xEndInt) {
const double rFract
= clip((double)((xEnd - xEndInt) / sx), 0.0, 1.0);
const double lFract = 1.0 - rFract;
int jData = yStartInt * bPRSrc + xStartInt * 3;
for (int k = 0; k < 3; ++k) {
pixel[k] = lFract * dataSrc[jData++];
}
for (int k = 0; k < 3; ++k) {
pixel[k] += rFract * dataSrc[jData++];
}
} else {
int jData = yStartInt * bPRSrc + xStartInt * 3;
for (int k = 0; k < 3; ++k) pixel[k] = dataSrc[jData++];
}
imgDst.setPixel(x, y,
(uint8)pixel[0], (uint8)pixel[1], (uint8)pixel[2]);
}
}
}
// done
return true;
}
} // namespace
bool scaleTo(const Image &imgSrc, Image &imgDst, int w, int h, int align)
{
Image imgTmp;
return w <= 0 || h <= 0 ? false
: w >= imgSrc.w() && h >= imgSrc.h()
? scaleUp(imgSrc, imgDst, w, h, align)
: w <= imgSrc.w() && h <= imgSrc.h()
? scaleDown(imgSrc, imgDst, w, h, align)
: w >= imgSrc.w()
? scaleUp(imgSrc, imgTmp, w, imgSrc.h(), 1)
&& scaleDown(imgTmp, imgDst, w, h, align)
: scaleDown(imgSrc, imgTmp, w, imgSrc.h(), 1)
&& scaleUp(imgTmp, imgDst, w, h, align);
}
PPM file IO
imagePPM.h:
#ifndef IMAGE_PPM_H
#define IMAGE_PPM_H
#include <iostream>
#include "image.h"
// reads a binary PPM file.
bool readPPM( // returns true if successful
std::istream &in, // input stream (must be opened with std::ios::binary)
Image &img, // image to read into
int align = 4); // row alignment
// writes binary PPM file.
bool writePPM( // returns true if successful
std::ostream &out, // output stream (must be opened with std::ios::binary)
const Image &img); // image to write from
#endif // IMAGE_PPM_H
imagePPM.cc:
#include <sstream>
#include <string>
#include "imagePPM.h"
// reads a binary PPM file.
bool readPPM( // returns true if successful
std::istream &in, // input stream (must be opened with std::ios::binary)
Image &img, // image to read into
int align) // row alignment
{
// parse header
std::string buffer;
if (!getline(in, buffer)) return false;
if (buffer != "P6") {
std::cerr << "Wrong header! 'P6' expected.\n";
return false;
}
int w = 0, h = 0, t = 0;
for (int i = 0; i < 3;) {
if (!getline(in, buffer)) return false;
if (buffer.empty()) continue; // skip empty lines
if (buffer[0] == '#') continue; // skip comments
std::istringstream str(buffer);
switch (i) {
case 0:
if (!(str >> w)) continue;
++i;
case 1:
if (!(str >> h)) continue;
++i;
case 2:
if (!(str >> t)) continue;
++i;
}
}
if (t != 255) {
std::cerr << "Unsupported format! t = 255 expected.\n";
return false;
}
// allocate image buffer
uint8 *data = img.resize(w, h, (w * 3 + align - 1) / align * align);
// read data
for (int i = 0; i < h; ++i) {
if (!in.read((char*)data, 3 * img.w())) return false;
data += img.bPR();
}
// done
return true;
}
// writes binary PPM file.
bool writePPM( // returns true if successful
std::ostream &out, // output stream (must be opened with std::ios::binary)
const Image &img) // image to write from
{
// write header
if (!(out << "P6\n" << img.w() << ' ' << img.h() << " 255\n")) return false;
// write image data
for (size_t y = 0; y < img.h(); ++y) {
const uint8 *const data = img.data(y);
if (!out.write((const char*)data, 3 * img.w())) return false;
}
// done
return true;
}
The main application
scaleRGBImg.cc:
#include <iostream>
#include <fstream>
#include <string>
#include "image.h"
#include "imagePPM.h"
#include "imageScale.h"
int main(int argc, char **argv)
{
// read command line arguments
if (argc <= 3) {
std::cerr << "Missing arguments!\n";
std::cout
<< "Usage:\n"
<< " scaleRGBImg IN_FILE SCALE OUT_FILE\n";
return 1;
}
const std::string inFile = argv[1];
char *end;
const double s = std::strtod(argv[2], &end);
if (end == argv[2] || *end != '\0') {
std::cerr << "Invalid scale factor '" << argv[2] << "'!\n";
return 1;
}
if (s <= 0.0) {
std::cerr << "Invalid scale factor " << s << "!\n";
return 1;
}
const std::string outFile = argv[3];
// read image
Image imgSrc;
{ std::ifstream fIn(inFile.c_str(), std::ios::binary);
if (!readPPM(fIn, imgSrc)) {
std::cerr << "Reading '" << inFile << "' failed!\n";
return 1;
}
}
// scale image
Image imgDst;
if (!scale(imgSrc, imgDst, s)) {
std::cerr << "Scaling failed!\n";
return 1;
}
// write image
{ std::ofstream fOut(outFile.c_str(), std::ios::binary);
if (!writePPM(fOut, imgDst) || (fOut.close(), !fOut.good())) {
std::cerr << "Writing '" << outFile << "' failed!\n";
return 1;
}
}
// done
return 0;
}
Test
Compiled in cygwin64:
$ g++ -std=c++11 -o scaleRGBImg scaleRGBImg.cc image.cc imagePPM.cc imageScale.cc
$
A sample image test.ppm for a test – converted to PPM in GIMP:
Test with the sample image:
$ for I in 0.8 0.6 0.4 0.2 ; do echo ./scaleRGBImg test.ppm $I test.$I.ppm ; done
./scaleRGBImg test.ppm 0.8 test.0.8.ppm
./scaleRGBImg test.ppm 0.6 test.0.6.ppm
./scaleRGBImg test.ppm 0.4 test.0.4.ppm
./scaleRGBImg test.ppm 0.2 test.0.2.ppm
$ for I in 0.8 0.6 0.4 0.2 ; do ./scaleRGBImg test.ppm $I test.$I.ppm ; done
$
This is what came out:
test.0.8.ppm:
test.0.6.ppm:
test.0.4.ppm:
test.0.2.ppm:
I don't really know how to explain my problem properly, but hopefully you can understand it using the images i provide.
I made this Mandelbrot image generator using a template and a tutorial on the internet, and i'm trying to make the generation process multi-threaded so the image gets split up into 4 equal parts and each thread calculates that part. The problem is, the first half of the image turns out black, and the second half generates fine. I have no idea what the problem is.
This code works fine without the multi-threading, so even though i feel like the problem resides there i still can't find it.
Code:
// mandelbrot.cpp
// compile with: g++ -std=c++11 -pthread mandelbrot.cpp -o mandelbrot
// view output with: eog mandelbrot.ppm
#include <fstream>
#include <iostream>
#include <complex> // if you make use of complex number facilities in C++
#include <thread>
#include <vector>
template <class T>
struct RGB
{
T r, g, b;
};
template <class T>
class Matrix
{
public:
Matrix(const size_t rows, const size_t cols) : _rows(rows), _cols(cols)
{
_matrix = new T *[rows];
for (size_t i = 0; i < rows; ++i)
{
_matrix[i] = new T[cols];
}
}
Matrix(const Matrix &m) : _rows(m._rows), _cols(m._cols)
{
_matrix = new T *[m._rows];
for (size_t i = 0; i < m._rows; ++i)
{
_matrix[i] = new T[m._cols];
for (size_t j = 0; j < m._cols; ++j)
{
_matrix[i][j] = m._matrix[i][j];
}
}
}
~Matrix()
{
for (size_t i = 0; i < _rows; ++i)
{
delete[] _matrix[i];
}
delete[] _matrix;
}
T *operator[](const size_t nIndex)
{
return _matrix[nIndex];
}
size_t width() const { return _cols; }
size_t height() const { return _rows; }
protected:
size_t _rows, _cols;
T **_matrix;
};
// Portable PixMap image
class PPMImage : public Matrix<RGB<unsigned char>>
{
public:
PPMImage(const size_t height, const size_t width) : Matrix(height, width) {}
void save(const std::string &filename)
{
std::ofstream out(filename, std::ios_base::binary);
out << "P6" << std::endl
<< _cols << " " << _rows << std::endl
<< 255 << std::endl;
for (size_t y = 0; y < _rows; y++)
for (size_t x = 0; x < _cols; x++)
out << _matrix[y][x].r << _matrix[y][x].g << _matrix[y][x].b;
}
};
const void calculateImage(PPMImage &image, unsigned &width, unsigned &height, unsigned &minX, unsigned &maxX, unsigned &minY, unsigned &maxY)
{
double MinRe = -2.0;
double MaxRe = 1.0;
double MinIm = -1.2;
double MaxIm = MinIm + (MaxRe - MinRe) * width / height;
double Re_facor = (MaxRe - MinRe) / (width - 1);
double Im_factor = (MaxIm - MinIm) / (height - 1);
unsigned MaxIterations = 30;
for (unsigned y = minY; y < maxY; ++y)
{
double c_im = MaxIm - y * Im_factor;
for (unsigned x = minX; x < maxX; ++x)
{
double c_re = MinRe + x * Re_facor;
double Z_re = c_re, Z_im = c_im;
bool isInside = true;
for (unsigned n = 0; n < MaxIterations; ++n)
{
double Z_re2 = Z_re * Z_re, Z_im2 = Z_im * Z_im;
if (Z_re2 + Z_im2 > 4)
{
isInside = false;
break;
}
Z_im = 2 * Z_re * Z_im + c_im;
Z_re = Z_re2 - Z_im2 + c_re;
}
if (isInside)
{
image[y][x].r = 255;
image[y][x].g = 0;
image[y][x].b = 0;
}
}
}
}
int main()
{
unsigned width = 1600;
unsigned height = 1600;
unsigned minX = 0;
unsigned minY = 0;
unsigned maxX = 800;
unsigned maxY = 800;
std::cout << "Creating image...\n";
PPMImage image(height, width);
std::vector<std::thread> workers;
for (int i = 0; i < 4; i++)
{
workers.push_back(std::thread(calculateImage, std::ref(image), std::ref(width), std::ref(height), std::ref(minX), std::ref(maxX), std::ref(minY), std::ref(maxY)));
std::cout << "Thread #" << i << " created. Calculating (" << minX << "," << minY << ") to (" << maxX << "," << maxY << ").\n";
if (i == 0)
{
minY = 800;
maxY = 1600;
}
else if (i == 1)
{
minX = 800;
maxX = 1600;
minY = 0;
maxY = 800;
}
else if (i == 2)
{
minX = 800;
maxX = 1600;
minY = 800;
maxY = 1600;
}
}
for (auto &th : workers)
{
th.join();
}
std::cout << "Saving to image...\n";
image.save("mandelbrot.ppm");
std::cout << "Finished.\n";
return 0;
}
Resulting image:
You can see a part of the first half is calculated, but most of it stays black, in fact it's "overwritten" by the third and fourth thread.
It's probably a very obvious mistake but i can't figure it out.
If i try to do it without multi-threading, e.g.:
calculateImage(image, width, height, minX, maxX, minY, maxY);
It works fine.
You are passing references to minX, minY, maxX, maxY to the threads, but then change their values immediately after launching them.
All threads will be referring to the same variables in main. Changing the value may be observed by all the threads.
Pass the four variables by value instead.
I want to genetically recreate a given image
I try to do this in sfml.
I have created a self-parting square that tries to evolve to look
like source image
Sadly, this thing crashes and I have no idea why, I suppose everything is handled nice and the vector appending shouldn't be a problem.
Please check out the code:
The main function:
#include "divisablesquare.h"
#include <SFML/Graphics.hpp>
#include <iostream>
#include <cstring>
#include <string>
#include <error.h>
#include <algorithm>
namespace GLOBAL
{
bool DEBUG_MODE = false;
};
int IDX = 0;
int main(int argc, char * argv[])
{
srand(time(NULL));
std::string def;
for(int i = 1; i < argc; i++)
{
def = argv[i];
std::string def2 = def;
std::transform(def2.begin(), def2.end(), def2.begin(), ::tolower);
if(strcmp(def2.c_str(), "--debug") == 0)
{
GLOBAL::DEBUG_MODE = true;
std::cerr << "Running in debug mode" << std::endl;
}
else
{
break;
}
}
sf::Image sourceImage;
sf::Texture sample;
if(!sourceImage.loadFromFile(def) && GLOBAL::DEBUG_MODE)
{
std::cerr << "Failed to open specified image!" << std::endl;
}
sample.loadFromImage(sourceImage);
sf::RectangleShape sourceRect;
sourceRect.setSize((sf::Vector2f)sourceImage.getSize());
sourceRect.setTexture(&sample);
sf::RenderWindow mainWindow(sf::VideoMode(sourceImage.getSize().x*2+10, sourceImage.getSize().y), "Genetic Image Generator");
std::vector<DivisableSquare> dSquares;
{
DivisableSquare starter(&dSquares, &sourceImage);
starter.init(128, 128, 128, sourceImage.getSize().x, sourceImage.getSize().y, sourceImage.getSize().x + 10, 0);
starter.Shape.setPosition({(float)sourceImage.getSize().x + 10, 0});
starter.Shape.setFillColor({128,128,128});
starter.Shape.setSize({(float)sourceImage.getSize().x, (float)sourceImage.getSize().y});
dSquares.push_back(starter);
}
sf::Clock clock;
while(mainWindow.isOpen())
{
sf::Time elapsed = clock.getElapsedTime();
if(elapsed.asMilliseconds() > 1000)
{
clock.restart();
dSquares.at(rand() % dSquares.size()).Partup();
}
sf::Event mainEvent;
while(mainWindow.pollEvent(mainEvent))
{
if(mainEvent.type == sf::Event::Closed)
mainWindow.close();
}
mainWindow.clear();
mainWindow.draw(sourceRect);
for(auto &&ref: dSquares)
{
mainWindow.draw(ref.Shape);
}
mainWindow.display();
}
}
divisablesquare header:
#ifndef DIVISABLESQUARE_H
#define DIVISABLESQUARE_H
#include <vector>
#include <SFML/Graphics.hpp>
class DivisableSquare
{
private:
sf::Image * parentImage;
std::vector<DivisableSquare> * ParentContainter;
unsigned short red, green, blue;
double width, height;
double posX, posY;
int id;
public:
~DivisableSquare();
sf::RectangleShape Shape;
DivisableSquare(std::vector<DivisableSquare>*, sf::Image*);
void init(unsigned short, unsigned short, unsigned short, unsigned int, unsigned int, unsigned int, unsigned int);
void Partup();
};
#endif // DIVISABLESQUARE_H
and the c++ file:
#include "divisablesquare.h"
#include <random>
#include <algorithm>
#include <iostream>
extern int IDX;
DivisableSquare::DivisableSquare(std::vector<DivisableSquare> *pc, sf::Image*tp)
{
this->ParentContainter = pc;
this->parentImage = tp;
this->id = IDX;
IDX++;
}
DivisableSquare::~DivisableSquare()
{
}
void DivisableSquare::init(unsigned short r, unsigned short g, unsigned short b,
unsigned int width, unsigned int height, unsigned int posX, unsigned int posY)
{
this->red = r;
this->blue = b;
this->green = g;
this->width = width;
this->height = height;
this->posX = posX;
this->posY = posY;
}
void DivisableSquare::Partup()
{
if(this->width < 2 && this->height < 2)
return;
double percentCut = (rand()%60 + 20)/100;
bool horizontalCut = rand()%2;
double posX1, posX2;
double posY1, posY2;
double width1, width2;
double height1, height2;
if(horizontalCut)
{
posX1 = this->posX;
posX2 = (this->posX+this->width)*percentCut;
posY1 = this->posY;
posY2 = this->posY;
width1 = this->width*percentCut;
width2 = this->width*(1-percentCut);
height1 = this->height;
height2 = this->height;
}
else
{
posX1 = this->posX;
posX2 = this->posX;
posY1 = this->posY;
posY2 = (this->posY + this->height)*percentCut;
width1 = this->width;
width2 = this->width;
height1 = this->height*percentCut;
height2 = this->height*(1-percentCut);
}
struct RGB
{
float r, g, b;
float parentCmp;
float originalCmp;
float averageCmp;
/**
* Make sure to append originalCmp later
* also remove "= 0"
* DONE
*/
};
std::vector<RGB> originalPixels1;
std::vector<RGB> originalPixels2;
for(unsigned int i = posX1; i < posX1+width1; i++)
{
for(unsigned int j = posY1; j < posY1+height1; j++)
{
if(this->parentImage->getSize().x > i && this->parentImage->getSize().y > j)
{
RGB pixel;
pixel.r = this->parentImage->getPixel(i, j).r;
pixel.g = this->parentImage->getPixel(i, j).g;
pixel.b = this->parentImage->getPixel(i, j).b;
originalPixels1.push_back(pixel);
}
}
}
for(unsigned int i = posX2; i < posX2+width2; i++)
{
for(unsigned int j = posY2; j < posY2+height2; j++)
{
if(this->parentImage->getSize().x > i && this->parentImage->getSize().y > j)
{
RGB pixel;
pixel.r = this->parentImage->getPixel(i, j).r;
pixel.g = this->parentImage->getPixel(i, j).g;
pixel.b = this->parentImage->getPixel(i, j).b;
originalPixels2.push_back(pixel);
}
}
}
RGB pix1 = {0,0,0,0,0,0}, pix2={0,0,0,0,0,0};
for(auto &&ref : originalPixels1)
{
pix1.r += ref.r;
pix1.g += ref.g;
pix1.b += ref.b;
}
pix1.r /= originalPixels1.size();
pix1.g /= originalPixels1.size();
pix1.b /= originalPixels1.size();
for(auto &&ref : originalPixels2)
{
pix2.r += ref.r;
pix2.g += ref.g;
pix2.b += ref.b;
}
pix2.r /= originalPixels1.size();
pix2.g /= originalPixels1.size();
pix2.b /= originalPixels1.size();
auto comparVal = [](RGB v1, RGB v2)
{
float val1 = 0.2126*v1.r + 0.7152*v1.g + 0.0722*v1.b;
float val2 = 0.2126*v2.r + 0.7152*v2.g + 0.0722*v2.b;
return (val1 > val2) ? val1-val2 : val2-val1;
};//smaller - better
RGB first[100];
RGB second[100];
for(int i = 0; i < 100; i++)
{
first[i].r = rand() % 255;
first[i].g = rand() % 255;
first[i].b = rand() % 255;
second[i].r = rand() % 255;
second[i].g = rand() % 255;
second[i].b = rand() % 255;
}
// insert orginalcmp here
for(int i = 0; i < 100; i++)
{
first[i].originalCmp = comparVal(first[i], pix1);
second[i].originalCmp = comparVal(second[i], pix2);
}
RGB pRgb;
pRgb.r = this->red;
pRgb.b = this->blue;
pRgb.b = this->blue;
for(int i = 0; i < 100; i++)
{
first[i].parentCmp = comparVal(first[i], pRgb);
second[i].parentCmp = comparVal(second[i], pRgb);
first[i].averageCmp = (first[i].originalCmp+first[i].parentCmp)/2;
second[i].averageCmp = (second[i].originalCmp+second[i].parentCmp)/2;
}
std::sort(first, first+100, [](const RGB& l, const RGB& r){return r.averageCmp > l.averageCmp;});
std::sort(second, second+100, [](const RGB& l, const RGB& r){return r.averageCmp > l.averageCmp;});
RGB bestfirst = first[rand()%10], bestsecond = second[rand()%10];
DivisableSquare firstSQ(this->ParentContainter, this->parentImage);
DivisableSquare secondSQ(this->ParentContainter, this->parentImage);
firstSQ.init(bestfirst.r, bestfirst.g, bestfirst.b, width1, height1, posX1, posY1);
secondSQ.init(bestsecond.r, bestsecond.g, bestsecond.b, width2, height2, posX2, posY2);
firstSQ.Shape.setFillColor({(sf::Uint8)bestfirst.r, (sf::Uint8)bestfirst.g, (sf::Uint8)bestfirst.b});
secondSQ.Shape.setFillColor({(sf::Uint8)bestsecond.r, (sf::Uint8)bestsecond.g, (sf::Uint8)bestsecond.b});
firstSQ.Shape.setSize({(float)width1, (float)height1});
secondSQ.Shape.setSize({(float)width2, (float)height2});
firstSQ.Shape.setPosition({(float)posX1 + this->parentImage->getSize().x + 10, (float)posY1});
secondSQ.Shape.setPosition({(float)posX2 + this->parentImage->getSize().x + 10, (float)posY2});
this->ParentContainter->push_back(firstSQ);
this->ParentContainter->push_back(secondSQ);
//crash here
for(unsigned int i = 0; i < this->ParentContainter->size(); i++)
{
if(this->ParentContainter->at(i).id == this->id)
this->ParentContainter->erase(this->ParentContainter->begin()+i);
}
}
I know this is a poor code but i just wanted to test things out, what could cause a vector::push_back to crash my app?
The problem is that you're adding to dSquares while executing code using a member of the vector. When the vector is resized (during the this->ParentContainter->push_back(firstSQ); call), the object that this points to is moved (since it is part of the vector). However, this keeps pointing at the previous location of the object, and when you try to push the second new square you access this deallocated memory, resulting in Undefined Behavior and (in this case) a crash.
A possible fix is to call dSquares.reserve(dSquares.size() + 2); before you call dSquares.at(rand() % dSquares.size()).Partup();. This will allocate extra memory of the (potential) two new objects that are added so that when you call push_back within Partup a reallocation of the vector will not occur.
Another possibility is to erase the parent square first, then push the two new squares to the vector. When you push the first new square, it won't have to resize the vector (since there will be space for at least one element from removing the parent). Pushing the second element might result in a resize, so dereferencing this after that push could still crash.
I am trying to implement the rasterization method in cc+. I am trying to implement an interpolation function that handles the interpolation between the x,y and z vertices. That way I can save the inverse of z in a depth buffer.
At this point I get only the vertices drawn on the rendered image. Can someone see what is wrong with my code? I have posted the full code so you can see the whole program.
Many thanks in advance.
EDIT
I saw that I had made an error in vertexshader by writing pixel.zinv = 1 / vPrime.z instead of p.zinv = 1/ vPrime.z. Now nothing renders, just a black screen.
EDIT 2
My check to see if a pixel should be painted was wrong.
if (depthBuffer[row[i].x][row[i].y] < row[i].zinv)
is correct. Now I get little pieces of color.
#include <iostream>
#include <glm/glm.hpp>
#include <SDL.h>
#include "SDLauxiliary.h"
#include "TestModel.h"
using namespace std;
using glm::vec2;
using glm::vec3;
using glm::ivec2;
using glm::mat3;
using glm::max;
// ----------------------------------------------------------------------------
// GLOBAL VARIABLES
int cc = 0;
const int SCREEN_WIDTH = 500;
const int SCREEN_HEIGHT = 500;
SDL_Surface* screen;
int t;
vector<Triangle> triangles;
vec3 cameraPos(0, 0, -3.001);
float f = 500;
double yaw = 0;
vec3 c1(cos(yaw), 0, -sin(yaw));
vec3 c2(0, 1, 0);
vec3 c3(sin(yaw), 0, cos(yaw));
glm::mat3 R(c1, c2, c3);
float translation = 0.1; // use this to set translation increment
const float PI = 3.1415927;
vec3 currentColor;
float depthBuffer[SCREEN_HEIGHT][SCREEN_WIDTH];
// ----------------------------------------------------------------------------
// STUCTURES
struct Pixel
{
int x;
int y;
float zinv;
}pixel;
// ----------------------------------------------------------------------------
// FUNCTIONS
void Update();
void Draw();
void VertexShader(const vec3& v, Pixel& p);
void Interpolate(ivec2 a, ivec2 b, vector<ivec2>& result);
void DrawLineSDL(SDL_Surface* surface, ivec2 a, ivec2 b, vec3 color);
void DrawPolygonEdges(const vector<vec3>& vertices);
void ComputePolygonRows(const vector<Pixel>& vertexPixels, vector<Pixel>& leftPixels, vector<Pixel>& rightPixels);
void DrawPolygonRows(const vector<Pixel>& leftPixels, const vector<Pixel>& rightPixels);
void DrawPolygon(const vector<vec3>& vertices);
void Interpolate2(Pixel a, Pixel b, vector<Pixel>& result);
int main(int argc, char* argv[])
{
LoadTestModel(triangles);
screen = InitializeSDL(SCREEN_WIDTH, SCREEN_HEIGHT);
t = SDL_GetTicks(); // Set start value for timer.
while (NoQuitMessageSDL())
{
Draw();
}
//Draw();
//cin.get();
SDL_SaveBMP(screen, "screenshot.bmp");
return 0;
}
void Draw()
{
SDL_FillRect(screen, 0, 0);
if (SDL_MUSTLOCK(screen))
SDL_LockSurface(screen);
for (int y = 0; y<SCREEN_HEIGHT; ++y)
for (int x = 0; x<SCREEN_WIDTH; ++x)
depthBuffer[y][x] = 0;
for (int i = 0; i<triangles.size(); ++i)
{
currentColor = triangles[i].color;
vector<vec3> vertices(3);
int aa = 24;
vertices[0] = triangles[i].v0;
vertices[1] = triangles[i].v1;
vertices[2] = triangles[i].v2;
DrawPolygon(vertices);
}
if (SDL_MUSTLOCK(screen))
SDL_UnlockSurface(screen);
SDL_UpdateRect(screen, 0, 0, 0, 0);
}
void VertexShader(const vec3& v, Pixel& p)
{
vec3 vPrime = (v - cameraPos)*R;
p.zinv = 1 / vPrime.z;
p.x = f * vPrime.x / vPrime.z + SCREEN_WIDTH / 2;
p.y = f * vPrime.y / vPrime.z + SCREEN_HEIGHT / 2;
//cout << p.x << " this is it " << p.y << endl;
depthBuffer[p.x][p.y] = pixel.zinv;
}
void ComputePolygonRows(const vector<Pixel>& vertexPixels,
vector<Pixel>& leftPixels, vector<Pixel>& rightPixels)
{
// Find y-min,max for the 3 vertices
vec3 vp(vertexPixels[0].y, vertexPixels[1].y, vertexPixels[2].y);
Pixel start; Pixel end; Pixel middle;
int yMin = 1000;
int yMax = -1000;
int w=0; int s=0;
for (int k = 0; k < vertexPixels.size(); ++k)
{
if (vp[k] <= yMin)
{
yMin = vp[k];
end = vertexPixels[k];
w = k;
}
}
for (int k = 0; k < vertexPixels.size(); ++k)
{
if (vp[k] >= yMax)
{
yMax = vp[k];
start = vertexPixels[k];
s = k;
}
}
for (int k = 0; k < vertexPixels.size(); ++k)
{
if (vertexPixels[k].y != start.y
&& vertexPixels[k].y != end.y)
{
middle = vertexPixels[k];
}
if (w!= k && s!= k)
{
middle = vertexPixels[k];
}
}
int ROWS = yMax - yMin + 1;
leftPixels.resize(ROWS);
rightPixels.resize(ROWS);
for (int i = 0; i<ROWS; ++i)
{
leftPixels[i].x = +numeric_limits<int>::max();
rightPixels[i].x = -numeric_limits<int>::max();
}
int pixels1 = glm::abs(start.y - end.y) + 1;
vector<Pixel> line1(pixels1);
Interpolate2(end, start, line1);
int pixels2 = glm::abs(end.y - middle.y) + 1;
vector<Pixel> line2(pixels2);
Interpolate2(end, middle, line2);
int pixels3 = glm::abs(middle.y - start.y) + 1;
vector<Pixel> line3(pixels3);
Interpolate2(middle, start, line3);
vector<Pixel> side1(ROWS);
for (int i = 0; i < line2.size(); ++i)
{
side1[i] = line2[i];
}
for (int i = 0; i < line3.size(); ++i)
{
side1[line2.size()+i-1] = line3[i];
}
for (int i = 0; i < ROWS; ++i)
{
if (line1[i].x < leftPixels[i].x)
{
leftPixels[i] = line1[i];
}
if (line1[i].x > rightPixels[i].x)
{
rightPixels[i] = line1[i];
}
if (side1[i].x < leftPixels[i].x)
{
leftPixels[i] = side1[i];
}
if (side1[i].x > rightPixels[i].x)
{
rightPixels[i] = side1[i];
}
}
}
void DrawPolygonRows(const vector<Pixel>& leftPixels, const vector<Pixel>& rightPixels)
{
//cout << cc++ << endl;
for (int k = 0; k < leftPixels.size(); ++k)
{
int pixels = glm::abs(leftPixels[k].x - rightPixels[k].x) + 1;
vector<Pixel> row(pixels);
Interpolate2(leftPixels[k], rightPixels[k], row);
for (int i = 0; i < pixels; ++i)
{
if (depthBuffer[row[i].x][row[i].y] < row[i].zinv)
{
PutPixelSDL(screen, row[i].x, row[i].y, currentColor);
depthBuffer[row[i].x][row[i].y] = row[i].zinv;
}
}
}
}
void DrawPolygon(const vector<vec3>& vertices)
{
int V = vertices.size();
vector<Pixel> vertexPixels(V);
for (int i = 0; i<V; ++i)
VertexShader(vertices[i], vertexPixels[i]);
vector<Pixel> leftPixels;
vector<Pixel> rightPixels;
ComputePolygonRows(vertexPixels, leftPixels, rightPixels);
DrawPolygonRows(leftPixels, rightPixels);
}
void Interpolate2(Pixel a, Pixel b, vector<Pixel>& result)
{
int N = result.size();
float stepx = (b.x - a.x) / float(glm::max(N - 1, 1));
float stepy = (b.y - a.y) / float(glm::max(N - 1, 1));
float stepz = (b.zinv - a.zinv) / float(glm::max(N - 1, 1));
float currentx = a.x;
float currenty = a.y;
float currentz = a.zinv;
for (int i = 0; i<N; ++i)
{
result[i].x = currentx;
result[i].y = currenty;
result[i].zinv = currentz;
currentx = a.x;
currenty = a.y;
currentz = a.zinv;
currentx += stepx;
currenty += stepy;
currentz += stepz;
}
}
The last loop in the last function seems incorrect to me. You define currentx outside the loop. Then, define a local variable inside the loop with the same name and use it later in the loop. I'd suggest not using the same name for variable inside the loop and outside it to make it more readable. Also, using global variables make the code difficult to read too, since I prefer to look at a function as a separate entity for analysis.
I have created a program which creates a Mandelbrot set. Now I'm trying to make it multithreaded.
// mandelbrot.cpp
// compile with: g++ -std=c++11 mandelbrot.cpp -o mandelbrot
// view output with: eog mandelbrot.ppm
#include <fstream>
#include <complex> // if you make use of complex number facilities in C++
#include <iostream>
#include <cstdlib>
#include <thread>
#include <mutex>
#include <vector>
using namespace std;
template <class T> struct RGB { T r, g, b; };
template <class T>
class Matrix {
public:
Matrix(const size_t rows, const size_t cols) : _rows(rows), _cols(cols) {
_matrix = new T*[rows];
for (size_t i = 0; i < rows; ++i) {
_matrix[i] = new T[cols];
}
}
Matrix(const Matrix &m) : _rows(m._rows), _cols(m._cols) {
_matrix = new T*[m._rows];
for (size_t i = 0; i < m._rows; ++i) {
_matrix[i] = new T[m._cols];
for (size_t j = 0; j < m._cols; ++j) {
_matrix[i][j] = m._matrix[i][j];
}
}
}
~Matrix() {
for (size_t i = 0; i < _rows; ++i) {
delete [] _matrix[i];
}
delete [] _matrix;
}
T *operator[] (const size_t nIndex)
{
return _matrix[nIndex];
}
size_t width() const { return _cols; }
size_t height() const { return _rows; }
protected:
size_t _rows, _cols;
T **_matrix;
};
// Portable PixMap image
class PPMImage : public Matrix<RGB<unsigned char> >
{
public:
unsigned int size;
PPMImage(const size_t height, const size_t width) : Matrix(height, width) { }
void save(const std::string &filename)
{
std::ofstream out(filename, std::ios_base::binary);
out <<"P6" << std::endl << _cols << " " << _rows << std::endl << 255 << std::endl;
for (size_t y=0; y<_rows; y++)
for (size_t x=0; x<_cols; x++)
out << _matrix[y][x].r << _matrix[y][x].g << _matrix[y][x].b;
}
};
/*Draw mandelbrot according to the provided parameters*/
void draw_Mandelbrot(PPMImage & image, const unsigned width, const unsigned height, double cxmin, double cxmax, double cymin, double cymax,unsigned int max_iterations)
{
for (std::size_t ix = 0; ix < width; ++ix)
for (std::size_t iy = 0; iy < height; ++iy)
{
std::complex<double> c(cxmin + ix / (width - 1.0)*(cxmax - cxmin), cymin + iy / (height - 1.0)*(cymax - cymin));
std::complex<double> z = 0;
unsigned int iterations;
for (iterations = 0; iterations < max_iterations && std::abs(z) < 2.0; ++iterations)
z = z*z + c;
image[iy][ix].r = image[iy][ix].g = image[iy][ix].b = iterations;
}
}
int main()
{
const unsigned width = 1600;
const unsigned height = 1600;
PPMImage image(height, width);
int parts = 8;
std::vector<int>bnd (parts, image.size);
std::thread *tt = new std::thread[parts - 1];
time_t start, end;
time(&start);
//Lauch parts-1 threads
for (int i = 0; i < parts - 1; ++i) {
tt[i] = std::thread(draw_Mandelbrot,ref(image), width, height, -2.0, 0.5, -1.0, 1.0, 10);
}
//Use the main thread to do part of the work !!!
for (int i = parts - 1; i < parts; ++i) {
draw_Mandelbrot(ref(image), width, height, -2.0, 0.5, -1.0, 1.0, 10);
}
//Join parts-1 threads
for (int i = 0; i < parts - 1; ++i)
tt[i].join();
time(&end);
std::cout << difftime(end, start) << " seconds" << std::endl;
image.save("mandelbrot.ppm");
delete[] tt;
return 0;
}
Now every thread draws the complete fractal (look in main()). How can I let the threads draw different parts of the fractal?
You're making this (quite a lot) harder than it needs to be. This is the sort of task to which OpenMP is almost perfectly suited. For this task it gives almost perfect scaling with a bare minimum of effort.
I modified your draw_mandelbrot by inserting a pragma before the outer for loop:
#pragma omp parallel for
for (int ix = 0; ix < width; ++ix)
for (int iy = 0; iy < height; ++iy)
Then I simplified your main down to:
int main() {
const unsigned width = 1600;
const unsigned height = 1600;
PPMImage image(height, width);
clock_t start = clock();
draw_Mandelbrot(image, width, height, -2.0, 0.5, -1.0, 1.0, 10);
clock_t stop = clock();
std::cout << (double(stop - start) / CLOCKS_PER_SEC) << " seconds\n";
image.save("mandelbrot.ppm");
return 0;
}
On my (fairly slow) machine, your original code ran in 4.73 seconds. My modified code ran in 1.38 seconds. That's an improvement of 3.4x out of code that's nearly indistinguishable from a trivial single-threaded version.
Just for what it's worth, I did a bit more rewriting to get this:
// mandelbrot.cpp
// compile with: g++ -std=c++11 mandelbrot.cpp -o mandelbrot
// view output with: eog mandelbrot.ppm
#include <fstream>
#include <complex> // if you make use of complex number facilities in C++
#include <iostream>
#include <cstdlib>
#include <thread>
#include <mutex>
#include <vector>
using namespace std;
template <class T> struct RGB { T r, g, b; };
template <class T>
struct Matrix
{
std::vector<T> data;
size_t rows;
size_t cols;
class proxy {
Matrix &m;
size_t index_1;
public:
proxy(Matrix &m, size_t index_1) : m(m), index_1(index_1) { }
T &operator[](size_t index) { return m.data[index * m.rows + index_1]; }
};
class const_proxy {
Matrix const &m;
size_t index_1;
public:
const_proxy(Matrix const &m, size_t index_1) : m(m), index_1(index_1) { }
T const &operator[](size_t index) const { return m.data[index * m.rows + index_1]; }
};
public:
Matrix(size_t rows, size_t cols) : data(rows * cols), rows(rows), cols(cols) { }
proxy operator[](size_t index) { return proxy(*this, index); }
const_proxy operator[](size_t index) const { return const_proxy(*this, index); }
};
template <class T>
std::ostream &operator<<(std::ostream &out, Matrix<T> const &m) {
out << "P6" << std::endl << m.cols << " " << m.rows << std::endl << 255 << std::endl;
for (size_t y = 0; y < m.rows; y++)
for (size_t x = 0; x < m.cols; x++) {
T pixel = m[y][x];
out << pixel.r << pixel.g << pixel.b;
}
return out;
}
/*Draw Mandelbrot according to the provided parameters*/
template <class T>
void draw_Mandelbrot(T & image, const unsigned width, const unsigned height, double cxmin, double cxmax, double cymin, double cymax, unsigned int max_iterations) {
#pragma omp parallel for
for (int ix = 0; ix < width; ++ix)
for (int iy = 0; iy < height; ++iy)
{
std::complex<double> c(cxmin + ix / (width - 1.0)*(cxmax - cxmin), cymin + iy / (height - 1.0)*(cymax - cymin));
std::complex<double> z = 0;
unsigned int iterations;
for (iterations = 0; iterations < max_iterations && std::abs(z) < 2.0; ++iterations)
z = z*z + c;
image[iy][ix].r = image[iy][ix].g = image[iy][ix].b = iterations;
}
}
int main() {
const unsigned width = 1600;
const unsigned height = 1600;
Matrix<RGB<unsigned char>> image(height, width);
clock_t start = clock();
draw_Mandelbrot(image, width, height, -2.0, 0.5, -1.0, 1.0, 255);
clock_t stop = clock();
std::cout << (double(stop - start) / CLOCKS_PER_SEC) << " seconds\n";
std::ofstream out("mandelbrot.ppm", std::ios::binary);
out << image;
return 0;
}
On my machine, this code runs in about 0.5 to 0.6 seconds.
As to why I made these changes: mostly to make it faster, cleaner, and simpler. Your Matrix class allocated a separate block of memory for each row (or perhaps column--didn't pay very close of attention). This allocates one contiguous block of the entire matrix instead. This eliminates a level of indirection to get to the data, and increases locality of reference, thus improving cache usage. It also reduces the total amount of data used.
Changing from using time to using clock to do the timing was to measure CPU time instead of wall time (and typically improve precision substantially as well).
Getting rid of the PPMImage class was done simply because (IMO) having a PPImage class that derives from a Matrix class just doesn't make much (if any) sense. I suppose it works (for a sufficiently loose definition of "work") but it doesn't strike me as good design. If you insist on doing it at all, it should at least be private derivation, because you're just using the Matrix as a way of implementing your PPMImage class, not (at least I certainly hope not) trying to make assertions about properties of PPM images.
If, for whatever, reason, you decide to handle the threading manually, the obvious way of dividing the work up between threads would still be by looking at the loops inside of draw_mandelbrot. The obvious one would be to leave your outer loop alone, but send the computation for each iteration off to a thread pool:
for (int ix = 0; ix < width; ++ix)
compute_thread(ix);
where the body of compute_thread is basically this chunk of code:
for (int iy = 0; iy < height; ++iy)
{
std::complex<double> c(cxmin + ix / (width - 1.0)*(cxmax - cxmin), cymin + iy / (height - 1.0)*(cymax - cymin));
std::complex<double> z = 0;
unsigned int iterations;
for (iterations = 0; iterations < max_iterations && std::abs(z) < 2.0; ++iterations)
z = z*z + c;
image[iy][ix].r = image[iy][ix].g = image[iy][ix].b = iterations;
}
There would obviously be a little work involved in passing the correct data to the compute thread (each thread should be pass a reference to a slice of the resulting picture), but that would be an obvious and fairly clean place to divide things up. In particular it divides the job up into enough tasks that you semi-automatically get pretty good load balancing (i.e., you can keep all the cores busy) but large enough that you don't waste massive amounts of time on communication and synchronization between the threads.
As to the result, with the number of iterations set to 255, I get the following (scaled to 25%):
...which is pretty much as I'd expect.
One of the big issues with this approach is that different regions take different amounts of time to calculate.
A more general approach is.
Start 1 source thread.
Start N worker threads.
Start 1 sink thread.
Create 2 thread safe queues (call them the source queue and the sink queue).
Divide the image into M (many more than N) pieces.
The source thread pushes pieces into the source queue
The workers pull piecse from the source queue, convert the pieces into result fragments, and pushes those fragments into the sink queue.
The sink thread takes fragments from the sink queue and combines them into the final image.
By dividing up the work this way, all the worker threads will be busy all the time.
You can divide the fractal into pieces by divide the start and end of the fractal with the screen dimension:
$this->stepsRe = (double)((($this->startRe * -1) + ($this->endeRe)) / ($this->size_x-1));
$this->stepsIm = (double)((($this->startIm * -1) + ($this->endeIm)) / ($this->size_y-1));