1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
#include "dzialka.h"
#include "message.h"

#include <algorithm>
#include <stdint.h>
#include <cstdio>
#include <vector>

using namespace std;

/*****************************************/

static uint32_t W, H, ID, N;
static uint32_t COL_CHUNKS, COL_CHUNK_SIZE;
static bool reversed;
static void init() {
    H = GetFieldHeight();
    W = GetFieldWidth();
    N = NumberOfNodes();
    ID = MyNodeId();
    reversed = (W<H);
    if (reversed) swap(W,H);
    COL_CHUNK_SIZE = 15000;
    COL_CHUNKS = (H + COL_CHUNK_SIZE - 1) / COL_CHUNK_SIZE;
}

static int getPixel(int Row, int Col) {
    if (reversed) return IsUsableCell(Col, Row);
    return IsUsableCell(Row, Col);
}

/*****************************************/

struct QElem {
    int32_t y;
    uint32_t h, value;
    uint64_t sum;
    QElem(int32_t Y=-1, uint32_t H=1, uint32_t VALUE=0, uint64_t SUM=0):y(Y),h(H),value(VALUE),sum(SUM){}
};

inline bool cmpQElem(const QElem &it, const uint32_t &y) { return it.y + it.h < y; }
uint64_t find_sum(vector<QElem> &Q, uint32_t y) {
    vector<QElem>::iterator it = lower_bound(Q.begin(), Q.end(), y, cmpQElem);
    if (it == Q.end()) return 0;
    return it->sum + (uint64_t)(y + 1 - it->y) * (uint64_t)it->value - (uint64_t)it->h * (uint64_t)it->value;
}

struct Column {
    uint64_t result;
    uint32_t width;
    vector<uint32_t> L, R;

    void send(int destID) {
        PutLL(destID, result);
        PutInt(destID, width);
        Send(destID);

        for (int i=0; i<COL_CHUNKS; ++i) {
            for (int j=0; j<COL_CHUNK_SIZE; ++j) {
                int idx = i*COL_CHUNK_SIZE + j;
                if (idx == L.size()) break;
                PutInt(destID, L[idx]);
            }
            Send(destID);
        }

        for (int i=0; i<COL_CHUNKS; ++i) {
            for (int j=0; j<COL_CHUNK_SIZE; ++j) {
                int idx = i*COL_CHUNK_SIZE + j;
                if (idx == R.size()) break;
                PutInt(destID, R[idx]);
            }
            Send(destID);
        }
    }

    void receive(int srcID) {
        Receive(srcID);
        result = GetLL(srcID);
        width = GetInt(srcID);

        for (int i=0; i<COL_CHUNKS; ++i) {
            Receive(srcID);
            for (int j=0; j<COL_CHUNK_SIZE; ++j) {
                int idx = i*COL_CHUNK_SIZE + j;
                if (idx == L.size()) break;
                L[idx] = GetInt(srcID);
            }
        }

        for (int i=0; i<COL_CHUNKS; ++i) {
            Receive(srcID);
            for (int j=0; j<COL_CHUNK_SIZE; ++j) {
                int idx = i*COL_CHUNK_SIZE + j;
                if (idx == R.size()) break;
                R[idx] = GetInt(srcID);
            }
        }
    }

    void merge(Column &c) {
        vector<QElem> QL, QR;
        vector<uint64_t> merge_result(L.size(), 0);
        QL.push_back(QElem()); QR.push_back(QElem());
        for(int32_t i=0; i<L.size(); ++i) {
            uint32_t Lvalue = R[i], Rvalue = c.L[i];
            if (Lvalue == 0 || Rvalue == 0) { QL.resize(1); QR.resize(1); QL[0].y = QR[0].y = i; continue; }

            uint32_t QLpos = QL.size() - 1;
            uint32_t QRpos = QR.size() - 1;
            while (QL[QLpos].value > Lvalue) --QLpos;
            while (QR[QRpos].value > Rvalue) --QRpos;

            uint32_t min_Ly = QL[QLpos].y + (QL[QLpos].value == Lvalue ? 0 : QL[QLpos].h);
            uint32_t min_Ry = QR[QRpos].y + (QR[QRpos].value == Rvalue ? 0 : QR[QRpos].h);
            uint32_t min_result_y = min(min_Ly, min_Ry);

            uint64_t min_sum;
            if (min_Ly < min_Ry) {
                min_sum = min_Ly > 0 ? find_sum(QR, min_Ly - 1) : 0;
            } else if (min_Ly > min_Ry) {
                min_sum = min_Ry > 0 ? find_sum(QL, min_Ry - 1) : 0;
            }

            QL.resize(QLpos+1); QR.resize(QRpos+1);
            uint32_t new_Ly = QL.back().y + QL.back().h;
            uint32_t delta_Lheight = i + 1 - new_Ly;
            if (QL.back().value == Lvalue) {
                QL.back().h += delta_Lheight;
                QL.back().sum += (uint64_t)delta_Lheight * (uint64_t)Lvalue;
            } else QL.push_back(QElem(new_Ly, delta_Lheight, Lvalue, QL.back().sum + (uint64_t)delta_Lheight * (uint64_t)Lvalue));

            uint32_t new_Ry = QR.back().y + QR.back().h;
            uint32_t delta_Rheight = i + 1 - new_Ry;
            if (QR.back().value == Rvalue) {
                QR.back().h += delta_Rheight;
                QR.back().sum += (uint64_t)delta_Rheight * (uint64_t)Rvalue;
            } else QR.push_back(QElem(new_Ry, delta_Rheight, Rvalue, QR.back().sum + (uint64_t)delta_Rheight * (uint64_t)Rvalue));

            merge_result[i] += (uint64_t)Lvalue * (uint64_t)Rvalue * (uint64_t)min(QL.back().h, QR.back().h);
            if (min_Ly < min_Ry) merge_result[i] += (uint64_t)Lvalue * (uint64_t)(QR[QR.size()-2].sum - min_sum);
            else if (min_Ly > min_Ry) merge_result[i] += (uint64_t)Rvalue * (uint64_t)(QL[QL.size()-2].sum - min_sum);
            if (min_result_y > 0) merge_result[i] += merge_result[min_result_y - 1];

            result += merge_result[i];
        }

        for(int i=0; i<L.size(); ++i) {
            L[i] = L[i] + (L[i]==width ? c.L[i] : 0);
            R[i] = c.R[i] + (c.R[i]==c.width ? R[i] : 0);
        }
        width += c.width;
        result += c.result;
    }

    Column(uint32_t H, uint32_t w, int32_t x = -1) {
        width = w;
        result = 0;
        L.resize(H, 0);
        R.resize(H, 0);
        if (width == 1) {
            int32_t last_usable_y = -1;
            for (int y=0; y<H; ++y) {
                L[y] = R[y] = getPixel(y, x);
                if (L[y] == 1) {
                    if (last_usable_y == -1) last_usable_y = y;
                    result += (last_usable_y == -1 ? 1 : y-last_usable_y+1);
                } else last_usable_y = -1;
            }
        }
    }
};

int main() {
    init();

    if (W < 500) {
        if (ID != 0) return 0;
        N = 1;
    }

    Column result(H, 0);
    int left = (W/N)*ID, right = (ID==N-1 ? W-1 : (W/N)*(ID+1) - 1);
    for(int i=left; i<=right; ++i) {
        Column d(H, 1, i);
        result.merge(d);
    }

    if (ID == 0) {
        for (int workerID=1; workerID<N; ++workerID) {
            Column workerResult(H, 0);
            workerResult.receive(workerID);
            result.merge(workerResult);
        }
        printf("%llu\n", result.result);
    } else {
        result.send(0);
    }

    return 0;
}