// clang-format off // It's hashing time! https://youtu.be/bS5P_LAqiVg?t=613 // hash of a_0, a_1, ..., a_n = a_0 + a_1 * Prime + a_2 * Prime^2 + ... + a_n * Prime^n // clang-format on #include <algorithm> #include <cassert> #include <cstdlib> #include <iostream> #include <string> #include <vector> #include "message.h" #include "poszukiwania.h" using namespace std; #define COLOR_BLACK "\x1b[30m" #define COLOR_RED "\x1b[31m" #define COLOR_GREEN "\x1b[32m" #define COLOR_YELLOW "\x1b[33m" #define COLOR_BLUE "\x1b[34m" #define COLOR_MAGENTA "\x1b[35m" #define COLOR_CYAN "\x1b[36m" #define COLOR_WHITE "\x1b[37m" #define COLOR_RESET "\x1b[0m" #define COLOR_BOLDBLACK "\033[1m\033[30m" #define COLOR_BOLDRED "\033[1m\033[31m" #define COLOR_BOLDGREEN "\033[1m\033[32m" #define COLOR_BOLDYELLOW "\033[1m\033[33m" #define COLOR_BOLDBLUE "\033[1m\033[34m" #define COLOR_BOLDMAGENTA "\033[1m\033[35m" #define COLOR_BOLDCYAN "\033[1m\033[36m" #define COLOR_BOLDWHITE "\033[1m\033[37m" const std::vector<const char*> colors{ COLOR_RED, COLOR_GREEN, COLOR_YELLOW, COLOR_BLUE, COLOR_MAGENTA, COLOR_CYAN, COLOR_BOLDRED, COLOR_BOLDGREEN, COLOR_BOLDYELLOW, COLOR_BOLDBLUE, COLOR_BOLDMAGENTA, COLOR_BOLDCYAN}; #define NDEBUG #ifndef NDEBUG #include <fstream> #include <cstring> struct MemChecker { ~MemChecker() { std::ifstream infile("/proc/self/status"); std::string line; while (std::getline(infile, line)) { if (line.compare(0, 6, "VmPeak") == 0) { int rv = 0; for (int i = 0; i < line.size(); ++i) { if (line[i] >= '0' && line[i] <= '9') { rv = 10 * rv + (line[i] - '0'); } } std::cerr << "memory: " << rv << " kB" << std::endl; return; } } std::cerr << "memory stats not found" << std::endl; } } __mem_checker; #define cdebug \ cerr << colors[MyNodeId() % colors.size()] << "node" << COLOR_RESET << " " #else class DebugStream {}; template <typename T> DebugStream& operator<<(DebugStream& s, T) { return s; } DebugStream cdebug; #endif typedef long long LL; constexpr LL kPrime = 1100000227; constexpr LL kPrimeMod = 1600002223; constexpr LL kInitHash = 1; int P; // number of instances int N; // text length int M; // pattern length int CS; // chunk length (size) int Pid; // current instance id LL Occurences; // occurences count int BufSizes[101]; int TotalSent[101]; void _PutLL(int target, long long x) { // BufSizes[target] += 8; // TotalSent[target] += 8; PutLL(target, x); } void _PutInt(int target, int x) { // BufSizes[target] += 4; // TotalSent[target] += 4; PutInt(target, x); } void _PutChar(int target, char x) { // BufSizes[target] += 1; // TotalSent[target] += 1; PutChar(target, x); } void _Send(int target) { // std::cerr << "sending " << BufSizes[target] << " bytes\n"; // BufSizes[target] = 0; Send(target); } // indexes start at 0 LL Text(int i) { assert(i >= 0 && i < N); return i < SeqLength() ? SeqAt(i + 1) : 1000000010; } LL Pattern(int i) { assert(i >= 0 && i < M); return SignalAt(i + 1); } LL TextBegin() { return CS * Pid; } LL TextEnd() { return CS * (Pid + 1); } vector<LL> ChunkHashes; LL PatternPrefixHash; // hash of the prefix of maximum len divisible by CS LL hashsh(const vector<LL>& vec) { LL hash = 0; LL p = kInitHash; for (LL i : vec) { hash = (hash + p * i) % kPrimeMod; p = p * kPrime % kPrimeMod; } return hash; } // Init. void MyInit() { N = SeqLength(); M = SignalLength(); P = min(NumberOfNodes(), N / 3); Pid = MyNodeId(); if (Pid >= P) { exit(0); } if (N % P) { N += P - N % P; } CS = N / P; ChunkHashes.resize(P); for (int i = CS * Pid; i < CS * (Pid + 1); ++i) { cdebug << "text[" << i << "] " << Text(i) << "\n"; } if (Pid == 0) { for (int i = 0; i < M; ++i) { cdebug << "pattern[" << i << "] " << Pattern(i) << "\n"; } } } // Compute chunk hashes and send them everyone (and receive from everyone). void ComputeChunkHashes() { LL my_hash = 0; LL p = kInitHash; for (LL i = TextBegin(); i < TextEnd(); ++i) { my_hash = (my_hash + p * Text(i)) % kPrimeMod; p = p * kPrime % kPrimeMod; } for (int pid = 0; pid < P; ++pid) { if (pid == Pid) { continue; } _PutLL(pid, my_hash); _Send(pid); } ChunkHashes[Pid] = my_hash; for (int pid = 0; pid < P; ++pid) { if (pid == Pid) { continue; } Receive(pid); ChunkHashes[pid] = GetLL(pid); } if (Pid == 0) { for (int i = 0; i < P; ++i) { cdebug << "hash " << i << " = " << ChunkHashes[i] << "\n"; } } } // a ^ x mod kPrimeMod LL Exp(LL a, LL x) { LL res = 1; while (x > 0) { if (x & 1) { res = res * a % kPrimeMod; } a = a * a % kPrimeMod; x >>= 1; } return res; } // Compute PatternPrefixHash. void ComputePatternPrefixHash() { if (M < CS) { return; } LL prefix_len = M - M % CS; LL my_part_start = prefix_len * Pid / P; LL my_part_end = prefix_len * (Pid + 1) / P; LL my_part_hash = 0; LL p = kInitHash; for (LL i = my_part_start; i < my_part_end; ++i) { my_part_hash = (my_part_hash + p * Pattern(i)) % kPrimeMod; p = p * kPrime % kPrimeMod; } PatternPrefixHash = 0; cdebug << " MY pattern prefix hash part " << my_part_start << ".." << my_part_end << " : " << my_part_hash << "\n"; // Send it to the first instance. _PutLL(0, my_part_hash); _Send(0); if (Pid == 0) { // Merge it. for (int pid = 0; pid < P; ++pid) { Receive(pid); LL received = GetLL(pid); PatternPrefixHash += received * Exp(kPrime, prefix_len * pid / P); PatternPrefixHash %= kPrimeMod; } // Send it back to all other instances. for (int i = 1; i < P; ++i) { _PutLL(i, PatternPrefixHash); _Send(i); } cdebug << "pattern prefix hash " << PatternPrefixHash << "\n"; } else { // Receive. Receive(0); PatternPrefixHash = GetLL(0); } } LL ArrTail(int pos) { if (pos < CS) { return Text(Pid * CS + pos); } else if (pos == CS) { return 1000000020; } else { return Pattern(max(0, M - CS) + pos - CS - 1); } } // One needs to send to at most to others. <how many back, how many bits> pair<pair<int, int>, pair<int, int>> _CalcSendTo() { if (M == 1) { return make_pair(make_pair(-1, -1), make_pair(-1, -1)); } if (M <= CS + 1) { return make_pair(make_pair(1, M - 1), make_pair(-1, -1)); } if ((M - 1) % CS == 0) { return make_pair(make_pair(M / CS + 1, CS), make_pair(-1, -1)); } int first_sent = (M - 2) / CS + 1; int bits_sent = (M - 2) % CS + 1; return make_pair(make_pair(first_sent, bits_sent), make_pair(first_sent - 1, CS - bits_sent)); } pair<pair<int, int>, pair<int, int>> CalcSendTo() { auto rv = _CalcSendTo(); if (rv.second.second == -1) { cdebug << "send/receive: " << rv.first.first << " " << rv.first.second << "\n"; } else { cdebug << "send/receive: " << rv.first.first << " " << rv.first.second << " " << rv.second.first << " " << rv.second.second << "\n"; } return rv; } // 250000 bytes at once at most const int kBytesAtOnce = 250000; void SendBits(int target, int start, int end, const vector<bool>& bits) { if (target < 0 || target >= P) { return; } cdebug << "SendBits to " << target << " " << start << ".." << end << "\n"; // _PutInt(target, end - start); while (start < end) { for (int i = 0; i < kBytesAtOnce && start < end; ++i) { char x = 0; int j = 0; for (j = 0; j < 8 && start < end; ++j) { x <<= 1; x |= (int)bits[start]; ++start; } x <<= (8 - j); cdebug << " sending 0x" << std::hex << (uint)x << std::dec << "\n"; _PutChar(target, x); } _Send(target); } cdebug << "done sending\n"; } void ReceiveBits(int target, int start, int end, vector<bool>& bits) { if (target < 0 || target >= P) { return; } cdebug << "ReceiveBits from " << target << " " << start << ".." << end << "\n"; // if (len != end - start) { // cdebug << "len " << len << " start " << start << " end " << end << " (" // << end - start << ")\n"; // assert(0); // } while (start < end) { Receive(target); for (int i = 0; i < kBytesAtOnce && start < end; ++i) { char x = GetChar(target); cdebug << " received 0x" << std::hex << (uint)x << std::dec << "\n"; for (int j = 7; j >= 0 && start < end; --j) { bits[start] = (x & (1 << j)) > 0; ++start; } } } cdebug << "done receiving\n"; } // Sent positions where the pattern can end (checking only data in this chunk). void SendBackMyStats() { // Pref-suf array for <chunk>#<pattern tail> (don't keep the second part of // the array T). // 2 * 10 ^ 9 / 100 * 4 B = 80 MB. Phew! // indexed from 1, sorry. vector<int> T(CS + 2); // for Chunk int k = T[0] = -1; for (int i = 0; i < CS; ++i) { while (k >= 0 && ArrTail(k) != ArrTail(i)) { k = T[k]; } ++k; T[i + 1] = k; // cdebug << "k at i + 1 = " << i + 1 << " = " << k << "\n"; } // No match for #. k = 0; // Let's try'em patterns. for (int i = 0; i < min(M, CS); ++i) { while (k >= 0 && ArrTail(k) != ArrTail(CS + 1 + i)) { k = T[k]; } ++k; // cdebug << " __ k at i + 1 = " << i + 1 << " = " << k << "\n"; } // Iterate over pref-sufs. cdebug << "final k: " << k << "\n"; // < 3 MB // Where a match can end? vector<bool> matches(CS); while (k > 0) { cdebug << "output : " << k - 1 << "\n"; matches[k - 1] = true; k = T[k]; } for (int i = 0; i < CS; ++i) { cdebug << " sending matches: " << matches[i] << "\n"; } // figure out where and what data to send and do it auto send_to = CalcSendTo(); if (send_to.first.second != -1) { SendBits(Pid - send_to.first.first, 0, send_to.first.second, matches); if (send_to.second.second != -1) { SendBits(Pid - send_to.second.first, send_to.first.second, CS, matches); } } } LL ArrHead(int pos) { if (pos < M) { return Pattern(pos); } else if (pos == M) { return 1000000030; } else { return Text(Pid * CS + pos - CS - 1); } } // iterate over the positions in this chunk and count pattern occurences void CalculateResultForChunk() { if ((P - Pid) * CS < M) { cdebug << "nothing fits\n"; return; // nothing fits } vector<bool> matches(CS); // receive the sent data. auto receive_from = CalcSendTo(); if (receive_from.first.second != -1) { ReceiveBits(Pid + receive_from.first.first, CS - receive_from.first.second, CS, matches); if (receive_from.second.second != -1) { ReceiveBits(Pid + receive_from.second.first, 0, CS - receive_from.first.second, matches); } } for (int i = 0; i < CS; ++i) { cdebug << " rec matches: " << matches[i] << "\n"; } if (M < CS) { // 1. M < CS // We won't fit any full chunks hashes - KMP this shit. // Pref-suf array for <pattern>#<chunk> (don't keep the second part of // the array T). // indexed from 1, as above. // WE'RE LOOKING FOR WHOLE MATCHES TOO vector<int> T(M + 2); // for pattern int k = T[0] = -1; for (int i = 0; i < M; ++i) { while (k >= 0 && ArrHead(k) != ArrHead(i)) { k = T[k]; } ++k; T[i + 1] = k; cdebug << "k at i + 1 = " << i + 1 << " = " << k << "\n"; } // No match for #. k = 0; // Let's try'em patterns. for (int i = 0; i < CS; ++i) { while (k >= 0 && ArrHead(k) != ArrHead(CS + 1 + i)) { k = T[k]; } ++k; if (k == M) { cdebug << "---------OCC k == M\n"; ++Occurences; } cdebug << " __ k at i + 1 = " << i + 1 << " = " << k << "\n"; } cdebug << "final k: " << k << "\n"; while (k > 0) { cdebug << " matches : " << k - 1 << " " << matches[k - 1] << "\n"; if (matches[CS - k]) { cdebug << " ---------OCC matches[CS - k]\n"; ++Occurences; } k = T[k]; } return; } // 2. M >= CS. // Get the hash of the chunks that fit in the pattern and move it move it! LL hash = 0; for (int i = 0; CS * (i + 1) <= M; ++i) { hash = (hash + ChunkHashes[Pid + i] * Exp(kPrime, CS * i)) % kPrimeMod; cdebug << "added chunk " << Pid + i << " to hash " << hash << "\n"; } LL p = 1; for (int i = 0; i < CS; ++i) { LL moved_pattern_hash = PatternPrefixHash * p % kPrimeMod; cdebug << "hash at " << i << " " << hash << " PatternHash " << moved_pattern_hash << " orig " << PatternPrefixHash << "\n"; if (moved_pattern_hash == hash && (matches[i] || (i == 0 && M % CS == 0))) { cdebug << " ---------OCC match at " << i << "\n"; ++Occurences; } // chop the first hash and the last from PatternPrefixHash and move hash // (multiply the multiplier by Prime) hash -= p * Text(CS * Pid + i) % kPrimeMod; cdebug << " hash tmp 1 " << hash << " (Text " << Text(CS * Pid + i) << ")\n"; hash = (hash + kPrimeMod) % kPrimeMod; cdebug << " hash tmp 2 " << hash << "\n"; p = p * kPrime % kPrimeMod; PatternPrefixHash -= Pattern(M - M % CS - i - 1) * Exp(kPrime, M - M % CS - i - 1) % kPrimeMod; cdebug << " pattern hash tmp 1 " << PatternPrefixHash << " (" << Pattern(M - M % CS - i - 1) << " " << Pattern(M - M % CS - i - 1) * Exp(kPrime, M - M % CS - i - 1) % kPrimeMod << ")\n"; PatternPrefixHash = (PatternPrefixHash + kPrimeMod) % kPrimeMod; cdebug << " pattern hash tmp 2 " << PatternPrefixHash << "\n"; } } void Final() { if (Pid == 0) { for (int pid = 1; pid < P; ++pid) { Receive(pid); Occurences += GetLL(pid); } cout << Occurences << "\n"; } else { _PutLL(0, Occurences); _Send(0); } } int main() { MyInit(); ComputeChunkHashes(); ComputePatternPrefixHash(); SendBackMyStats(); CalculateResultForChunk(); Final(); // for (int i = 0; i < P; ++i) { // std::cerr << " total sent: " << TotalSent[i] << "\n"; // } }
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 | // clang-format off // It's hashing time! https://youtu.be/bS5P_LAqiVg?t=613 // hash of a_0, a_1, ..., a_n = a_0 + a_1 * Prime + a_2 * Prime^2 + ... + a_n * Prime^n // clang-format on #include <algorithm> #include <cassert> #include <cstdlib> #include <iostream> #include <string> #include <vector> #include "message.h" #include "poszukiwania.h" using namespace std; #define COLOR_BLACK "\x1b[30m" #define COLOR_RED "\x1b[31m" #define COLOR_GREEN "\x1b[32m" #define COLOR_YELLOW "\x1b[33m" #define COLOR_BLUE "\x1b[34m" #define COLOR_MAGENTA "\x1b[35m" #define COLOR_CYAN "\x1b[36m" #define COLOR_WHITE "\x1b[37m" #define COLOR_RESET "\x1b[0m" #define COLOR_BOLDBLACK "\033[1m\033[30m" #define COLOR_BOLDRED "\033[1m\033[31m" #define COLOR_BOLDGREEN "\033[1m\033[32m" #define COLOR_BOLDYELLOW "\033[1m\033[33m" #define COLOR_BOLDBLUE "\033[1m\033[34m" #define COLOR_BOLDMAGENTA "\033[1m\033[35m" #define COLOR_BOLDCYAN "\033[1m\033[36m" #define COLOR_BOLDWHITE "\033[1m\033[37m" const std::vector<const char*> colors{ COLOR_RED, COLOR_GREEN, COLOR_YELLOW, COLOR_BLUE, COLOR_MAGENTA, COLOR_CYAN, COLOR_BOLDRED, COLOR_BOLDGREEN, COLOR_BOLDYELLOW, COLOR_BOLDBLUE, COLOR_BOLDMAGENTA, COLOR_BOLDCYAN}; #define NDEBUG #ifndef NDEBUG #include <fstream> #include <cstring> struct MemChecker { ~MemChecker() { std::ifstream infile("/proc/self/status"); std::string line; while (std::getline(infile, line)) { if (line.compare(0, 6, "VmPeak") == 0) { int rv = 0; for (int i = 0; i < line.size(); ++i) { if (line[i] >= '0' && line[i] <= '9') { rv = 10 * rv + (line[i] - '0'); } } std::cerr << "memory: " << rv << " kB" << std::endl; return; } } std::cerr << "memory stats not found" << std::endl; } } __mem_checker; #define cdebug \ cerr << colors[MyNodeId() % colors.size()] << "node" << COLOR_RESET << " " #else class DebugStream {}; template <typename T> DebugStream& operator<<(DebugStream& s, T) { return s; } DebugStream cdebug; #endif typedef long long LL; constexpr LL kPrime = 1100000227; constexpr LL kPrimeMod = 1600002223; constexpr LL kInitHash = 1; int P; // number of instances int N; // text length int M; // pattern length int CS; // chunk length (size) int Pid; // current instance id LL Occurences; // occurences count int BufSizes[101]; int TotalSent[101]; void _PutLL(int target, long long x) { // BufSizes[target] += 8; // TotalSent[target] += 8; PutLL(target, x); } void _PutInt(int target, int x) { // BufSizes[target] += 4; // TotalSent[target] += 4; PutInt(target, x); } void _PutChar(int target, char x) { // BufSizes[target] += 1; // TotalSent[target] += 1; PutChar(target, x); } void _Send(int target) { // std::cerr << "sending " << BufSizes[target] << " bytes\n"; // BufSizes[target] = 0; Send(target); } // indexes start at 0 LL Text(int i) { assert(i >= 0 && i < N); return i < SeqLength() ? SeqAt(i + 1) : 1000000010; } LL Pattern(int i) { assert(i >= 0 && i < M); return SignalAt(i + 1); } LL TextBegin() { return CS * Pid; } LL TextEnd() { return CS * (Pid + 1); } vector<LL> ChunkHashes; LL PatternPrefixHash; // hash of the prefix of maximum len divisible by CS LL hashsh(const vector<LL>& vec) { LL hash = 0; LL p = kInitHash; for (LL i : vec) { hash = (hash + p * i) % kPrimeMod; p = p * kPrime % kPrimeMod; } return hash; } // Init. void MyInit() { N = SeqLength(); M = SignalLength(); P = min(NumberOfNodes(), N / 3); Pid = MyNodeId(); if (Pid >= P) { exit(0); } if (N % P) { N += P - N % P; } CS = N / P; ChunkHashes.resize(P); for (int i = CS * Pid; i < CS * (Pid + 1); ++i) { cdebug << "text[" << i << "] " << Text(i) << "\n"; } if (Pid == 0) { for (int i = 0; i < M; ++i) { cdebug << "pattern[" << i << "] " << Pattern(i) << "\n"; } } } // Compute chunk hashes and send them everyone (and receive from everyone). void ComputeChunkHashes() { LL my_hash = 0; LL p = kInitHash; for (LL i = TextBegin(); i < TextEnd(); ++i) { my_hash = (my_hash + p * Text(i)) % kPrimeMod; p = p * kPrime % kPrimeMod; } for (int pid = 0; pid < P; ++pid) { if (pid == Pid) { continue; } _PutLL(pid, my_hash); _Send(pid); } ChunkHashes[Pid] = my_hash; for (int pid = 0; pid < P; ++pid) { if (pid == Pid) { continue; } Receive(pid); ChunkHashes[pid] = GetLL(pid); } if (Pid == 0) { for (int i = 0; i < P; ++i) { cdebug << "hash " << i << " = " << ChunkHashes[i] << "\n"; } } } // a ^ x mod kPrimeMod LL Exp(LL a, LL x) { LL res = 1; while (x > 0) { if (x & 1) { res = res * a % kPrimeMod; } a = a * a % kPrimeMod; x >>= 1; } return res; } // Compute PatternPrefixHash. void ComputePatternPrefixHash() { if (M < CS) { return; } LL prefix_len = M - M % CS; LL my_part_start = prefix_len * Pid / P; LL my_part_end = prefix_len * (Pid + 1) / P; LL my_part_hash = 0; LL p = kInitHash; for (LL i = my_part_start; i < my_part_end; ++i) { my_part_hash = (my_part_hash + p * Pattern(i)) % kPrimeMod; p = p * kPrime % kPrimeMod; } PatternPrefixHash = 0; cdebug << " MY pattern prefix hash part " << my_part_start << ".." << my_part_end << " : " << my_part_hash << "\n"; // Send it to the first instance. _PutLL(0, my_part_hash); _Send(0); if (Pid == 0) { // Merge it. for (int pid = 0; pid < P; ++pid) { Receive(pid); LL received = GetLL(pid); PatternPrefixHash += received * Exp(kPrime, prefix_len * pid / P); PatternPrefixHash %= kPrimeMod; } // Send it back to all other instances. for (int i = 1; i < P; ++i) { _PutLL(i, PatternPrefixHash); _Send(i); } cdebug << "pattern prefix hash " << PatternPrefixHash << "\n"; } else { // Receive. Receive(0); PatternPrefixHash = GetLL(0); } } LL ArrTail(int pos) { if (pos < CS) { return Text(Pid * CS + pos); } else if (pos == CS) { return 1000000020; } else { return Pattern(max(0, M - CS) + pos - CS - 1); } } // One needs to send to at most to others. <how many back, how many bits> pair<pair<int, int>, pair<int, int>> _CalcSendTo() { if (M == 1) { return make_pair(make_pair(-1, -1), make_pair(-1, -1)); } if (M <= CS + 1) { return make_pair(make_pair(1, M - 1), make_pair(-1, -1)); } if ((M - 1) % CS == 0) { return make_pair(make_pair(M / CS + 1, CS), make_pair(-1, -1)); } int first_sent = (M - 2) / CS + 1; int bits_sent = (M - 2) % CS + 1; return make_pair(make_pair(first_sent, bits_sent), make_pair(first_sent - 1, CS - bits_sent)); } pair<pair<int, int>, pair<int, int>> CalcSendTo() { auto rv = _CalcSendTo(); if (rv.second.second == -1) { cdebug << "send/receive: " << rv.first.first << " " << rv.first.second << "\n"; } else { cdebug << "send/receive: " << rv.first.first << " " << rv.first.second << " " << rv.second.first << " " << rv.second.second << "\n"; } return rv; } // 250000 bytes at once at most const int kBytesAtOnce = 250000; void SendBits(int target, int start, int end, const vector<bool>& bits) { if (target < 0 || target >= P) { return; } cdebug << "SendBits to " << target << " " << start << ".." << end << "\n"; // _PutInt(target, end - start); while (start < end) { for (int i = 0; i < kBytesAtOnce && start < end; ++i) { char x = 0; int j = 0; for (j = 0; j < 8 && start < end; ++j) { x <<= 1; x |= (int)bits[start]; ++start; } x <<= (8 - j); cdebug << " sending 0x" << std::hex << (uint)x << std::dec << "\n"; _PutChar(target, x); } _Send(target); } cdebug << "done sending\n"; } void ReceiveBits(int target, int start, int end, vector<bool>& bits) { if (target < 0 || target >= P) { return; } cdebug << "ReceiveBits from " << target << " " << start << ".." << end << "\n"; // if (len != end - start) { // cdebug << "len " << len << " start " << start << " end " << end << " (" // << end - start << ")\n"; // assert(0); // } while (start < end) { Receive(target); for (int i = 0; i < kBytesAtOnce && start < end; ++i) { char x = GetChar(target); cdebug << " received 0x" << std::hex << (uint)x << std::dec << "\n"; for (int j = 7; j >= 0 && start < end; --j) { bits[start] = (x & (1 << j)) > 0; ++start; } } } cdebug << "done receiving\n"; } // Sent positions where the pattern can end (checking only data in this chunk). void SendBackMyStats() { // Pref-suf array for <chunk>#<pattern tail> (don't keep the second part of // the array T). // 2 * 10 ^ 9 / 100 * 4 B = 80 MB. Phew! // indexed from 1, sorry. vector<int> T(CS + 2); // for Chunk int k = T[0] = -1; for (int i = 0; i < CS; ++i) { while (k >= 0 && ArrTail(k) != ArrTail(i)) { k = T[k]; } ++k; T[i + 1] = k; // cdebug << "k at i + 1 = " << i + 1 << " = " << k << "\n"; } // No match for #. k = 0; // Let's try'em patterns. for (int i = 0; i < min(M, CS); ++i) { while (k >= 0 && ArrTail(k) != ArrTail(CS + 1 + i)) { k = T[k]; } ++k; // cdebug << " __ k at i + 1 = " << i + 1 << " = " << k << "\n"; } // Iterate over pref-sufs. cdebug << "final k: " << k << "\n"; // < 3 MB // Where a match can end? vector<bool> matches(CS); while (k > 0) { cdebug << "output : " << k - 1 << "\n"; matches[k - 1] = true; k = T[k]; } for (int i = 0; i < CS; ++i) { cdebug << " sending matches: " << matches[i] << "\n"; } // figure out where and what data to send and do it auto send_to = CalcSendTo(); if (send_to.first.second != -1) { SendBits(Pid - send_to.first.first, 0, send_to.first.second, matches); if (send_to.second.second != -1) { SendBits(Pid - send_to.second.first, send_to.first.second, CS, matches); } } } LL ArrHead(int pos) { if (pos < M) { return Pattern(pos); } else if (pos == M) { return 1000000030; } else { return Text(Pid * CS + pos - CS - 1); } } // iterate over the positions in this chunk and count pattern occurences void CalculateResultForChunk() { if ((P - Pid) * CS < M) { cdebug << "nothing fits\n"; return; // nothing fits } vector<bool> matches(CS); // receive the sent data. auto receive_from = CalcSendTo(); if (receive_from.first.second != -1) { ReceiveBits(Pid + receive_from.first.first, CS - receive_from.first.second, CS, matches); if (receive_from.second.second != -1) { ReceiveBits(Pid + receive_from.second.first, 0, CS - receive_from.first.second, matches); } } for (int i = 0; i < CS; ++i) { cdebug << " rec matches: " << matches[i] << "\n"; } if (M < CS) { // 1. M < CS // We won't fit any full chunks hashes - KMP this shit. // Pref-suf array for <pattern>#<chunk> (don't keep the second part of // the array T). // indexed from 1, as above. // WE'RE LOOKING FOR WHOLE MATCHES TOO vector<int> T(M + 2); // for pattern int k = T[0] = -1; for (int i = 0; i < M; ++i) { while (k >= 0 && ArrHead(k) != ArrHead(i)) { k = T[k]; } ++k; T[i + 1] = k; cdebug << "k at i + 1 = " << i + 1 << " = " << k << "\n"; } // No match for #. k = 0; // Let's try'em patterns. for (int i = 0; i < CS; ++i) { while (k >= 0 && ArrHead(k) != ArrHead(CS + 1 + i)) { k = T[k]; } ++k; if (k == M) { cdebug << "---------OCC k == M\n"; ++Occurences; } cdebug << " __ k at i + 1 = " << i + 1 << " = " << k << "\n"; } cdebug << "final k: " << k << "\n"; while (k > 0) { cdebug << " matches : " << k - 1 << " " << matches[k - 1] << "\n"; if (matches[CS - k]) { cdebug << " ---------OCC matches[CS - k]\n"; ++Occurences; } k = T[k]; } return; } // 2. M >= CS. // Get the hash of the chunks that fit in the pattern and move it move it! LL hash = 0; for (int i = 0; CS * (i + 1) <= M; ++i) { hash = (hash + ChunkHashes[Pid + i] * Exp(kPrime, CS * i)) % kPrimeMod; cdebug << "added chunk " << Pid + i << " to hash " << hash << "\n"; } LL p = 1; for (int i = 0; i < CS; ++i) { LL moved_pattern_hash = PatternPrefixHash * p % kPrimeMod; cdebug << "hash at " << i << " " << hash << " PatternHash " << moved_pattern_hash << " orig " << PatternPrefixHash << "\n"; if (moved_pattern_hash == hash && (matches[i] || (i == 0 && M % CS == 0))) { cdebug << " ---------OCC match at " << i << "\n"; ++Occurences; } // chop the first hash and the last from PatternPrefixHash and move hash // (multiply the multiplier by Prime) hash -= p * Text(CS * Pid + i) % kPrimeMod; cdebug << " hash tmp 1 " << hash << " (Text " << Text(CS * Pid + i) << ")\n"; hash = (hash + kPrimeMod) % kPrimeMod; cdebug << " hash tmp 2 " << hash << "\n"; p = p * kPrime % kPrimeMod; PatternPrefixHash -= Pattern(M - M % CS - i - 1) * Exp(kPrime, M - M % CS - i - 1) % kPrimeMod; cdebug << " pattern hash tmp 1 " << PatternPrefixHash << " (" << Pattern(M - M % CS - i - 1) << " " << Pattern(M - M % CS - i - 1) * Exp(kPrime, M - M % CS - i - 1) % kPrimeMod << ")\n"; PatternPrefixHash = (PatternPrefixHash + kPrimeMod) % kPrimeMod; cdebug << " pattern hash tmp 2 " << PatternPrefixHash << "\n"; } } void Final() { if (Pid == 0) { for (int pid = 1; pid < P; ++pid) { Receive(pid); Occurences += GetLL(pid); } cout << Occurences << "\n"; } else { _PutLL(0, Occurences); _Send(0); } } int main() { MyInit(); ComputeChunkHashes(); ComputePatternPrefixHash(); SendBackMyStats(); CalculateResultForChunk(); Final(); // for (int i = 0; i < P; ++i) { // std::cerr << " total sent: " << TotalSent[i] << "\n"; // } } |