在Visual Studio中，int值'debug x86'和'release x64'有什么区别？

Question

#include "stdafx.h"
#include <stdio.h>
#include <iostream>
#include <algorithm>
#include <vector>

#define SIZE_X 256
#define SIZE_Y 256

using namespace std;


struct vectorStruct {
    int value;
    int point_x;
    int point_y;


    bool operator<(const vectorStruct& abc)const
        //compare function
    {
        return value < abc.value;
    }
};


int main()
{
    int c;
    int count = 1;

    vector<int> sortArr;
    int minSort;
    float imgWeight, imgHeight;
    // 0 ~ 255
    unsigned char arr[SIZE_X][SIZE_Y];
    unsigned char img[SIZE_X][SIZE_Y] = { { 0, }, };
    FILE *fp;

    std::vector<vectorStruct> minVector;
    vectorStruct vectorStruct;



    if ((fp = fopen("C:/Users/user/Documents/Visual Studio 2015/Projects/ExtracingImg/lena.raw", "rb")) == NULL) {
        perror;
        return -1;
    }
    fread(img, 1, SIZE_Y * SIZE_X, fp);
    fclose(fp);

    for (int y = 1; y < SIZE_Y - 1; y++) {
        for (int x = 1; x < SIZE_X - 1; x++) {
            imgWeight = img[y - 1][x + 1] // (0, 2) 
                - img[y - 1][x - 1] // (0, 0)
                + 2 * (img[y][x + 1] - img[y][x - 1]) // 2 * (1,2) - (1,0)
                + img[y + 1][x + 1] // (2, 2)
                - img[y + 1][x - 1]; // (2, 0)

            imgHeight = img[y - 1][x - 1]  // (0,0)
                - img[y + 1][x - 1] // (2, 0)
                + 2 * (img[y - 1][x] - img[y + 1][x]) // 2 * (0, 1) - (2, 0)
                + img[y - 1][x + 1] // (0 , 2)
                - img[y + 1][x + 1]; // (2, 2)

            c = (int)sqrt(imgHeight * imgHeight + imgWeight * imgWeight);

            if (c > 255) c = 255;
            arr[y][x] = c;
        }
    }

    for (int x = 1; x < SIZE_X - 1; x++)
    {
        for (int y = 1; y < SIZE_Y - 1; y++)
        {

            sortArr.push_back(arr[x - 1][y - 1]);
            sortArr.push_back(arr[x - 1][y]);
            sortArr.push_back(arr[x - 1][y + 1]);
            sortArr.push_back(arr[x][y - 1]);
            sortArr.push_back(arr[x][y]);
            sortArr.push_back(arr[x][y + 1]);
            sortArr.push_back(arr[x + 1][y - 1]);
            sortArr.push_back(arr[x + 1][y]);
            sortArr.push_back(arr[x + 1][y + 1]);

            auto iter = min_element(sortArr.begin(), sortArr.end());

            if (arr[x][y] == *iter)
            {
                vectorStruct.value = arr[x][y];
                vectorStruct.point_x = x;
                vectorStruct.point_y = y;
                minVector.push_back(vectorStruct);
            }
            sortArr.clear();
        }
    }
    cout << minVector.size() << endl;


    sort(minVector.begin(), minVector.end());


    // value, x, y  check
    for (auto i : minVector)
    {
        std::cout << i.value << " " << i.point_x << " " << i.point_y << std::endl;
        count++;
    }
    printf("%d", count);

    return 0;
}

我尝试在将“ lena.raw”图像文件导入为数组之后，并通过提取最小值按升序排序。

在这段代码中，我在“调试x86”模式下计算了“ for（auto i：minVector）循环”计数值为5785
但在“发布x64”模式下计数值为5609

在这种情况下有什么区别？
你能告诉我为什么我有不同的价值吗？