uct/uct.c

   1 #include <assert.h>
   2 #include <math.h>
   3 #include <stdio.h>
   4 #include <stdlib.h>
   5 #include <string.h>
   6 #include <time.h>
   7
   8 #define DEBUG
   9
  10 #include "debug.h"
  11 #include "board.h"
  12 #include "gtp.h"
  13 #include "chat.h"
  14 #include "move.h"
  15 #include "mq.h"
  16 #include "joseki/base.h"
  17 #include "playout.h"
  18 #include "playout/moggy.h"
  19 #include "playout/light.h"
  20 #include "tactics/util.h"
  21 #include "timeinfo.h"
  22 #include "uct/dynkomi.h"
  23 #include "uct/internal.h"
  24 #include "uct/plugins.h"
  25 #include "uct/prior.h"
  26 #include "uct/search.h"
  27 #include "uct/slave.h"
  28 #include "uct/tree.h"
  29 #include "uct/uct.h"
  30 #include "uct/walk.h"
  31 #include "dcnn.h"
  32
  33 struct uct_policy *policy_ucb1_init(struct uct *u, char *arg);
  34 struct uct_policy *policy_ucb1amaf_init(struct uct *u, char *arg, struct board *board);
  35 static void uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color);
  36
  37 /* Maximal simulation length. */
  38 #define MC_GAMELEN      MAX_GAMELEN
  39
  40
  41 static void
  42 setup_state(struct uct *u, struct board *b, enum stone color)
  43 {
  44         u->t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0,
  45                          u->max_pruned_size, u->pruning_threshold, u->local_tree_aging, u->stats_hbits);
  46         if (u->initial_extra_komi)
  47                 u->t->extra_komi = u->initial_extra_komi;
  48         if (u->force_seed)
  49                 fast_srandom(u->force_seed);
  50         if (UDEBUGL(3))
  51                 fprintf(stderr, "Fresh board with random seed %lu\n", fast_getseed());
  52         if (!u->no_tbook && b->moves == 0) {
  53                 if (color == S_BLACK) {
  54                         tree_load(u->t, b);
  55                 } else if (DEBUGL(0)) {
  56                         fprintf(stderr, "Warning: First move appears to be white\n");
  57                 }
  58         }
  59 }
  60
  61 static void
  62 reset_state(struct uct *u)
  63 {
  64         assert(u->t);
  65         tree_done(u->t); u->t = NULL;
  66 }
  67
  68 static void
  69 setup_dynkomi(struct uct *u, struct board *b, enum stone to_play)
  70 {
  71         if (u->t->use_extra_komi && !u->pondering && u->dynkomi->permove)
  72                 u->t->extra_komi = u->dynkomi->permove(u->dynkomi, b, u->t);
  73         else if (!u->t->use_extra_komi)
  74                 u->t->extra_komi = 0;
  75 }
  76
  77 void
  78 uct_prepare_move(struct uct *u, struct board *b, enum stone color)
  79 {
  80         if (u->t) {
  81                 /* Verify that we have sane state. */
  82                 assert(b->es == u);
  83                 assert(u->t && b->moves);
  84                 if (color != stone_other(u->t->root_color)) {
  85                         fprintf(stderr, "Fatal: Non-alternating play detected %d %d\n",
  86                                 color, u->t->root_color);
  87                         exit(1);
  88                 }
  89                 uct_htable_reset(u->t);
  90
  91         } else {
  92                 /* We need fresh state. */
  93                 b->es = u;
  94                 setup_state(u, b, color);
  95         }
  96
  97         u->ownermap.playouts = 0;
  98         memset(u->ownermap.map, 0, board_size2(b) * sizeof(u->ownermap.map[0]));
  99         u->played_own = u->played_all = 0;
 100 }
 101
 102 static void
 103 dead_group_list(struct uct *u, struct board *b, struct move_queue *mq, float thres)
 104 {
 105         enum gj_state gs_array[board_size2(b)];
 106         struct group_judgement gj = { .thres = thres, .gs = gs_array };
 107         board_ownermap_judge_groups(b, &u->ownermap, &gj);
 108         groups_of_status(b, &gj, GS_DEAD, mq);
 109 }
 110
 111 bool
 112 uct_pass_is_safe(struct uct *u, struct board *b, enum stone color, bool pass_all_alive)
 113 {
 114         /* Make sure enough playouts are simulated to get a reasonable dead group list. */
 115         while (u->ownermap.playouts < GJ_MINGAMES)
 116                 uct_playout(u, b, color, u->t);
 117
 118         struct move_queue mq = { .moves = 0 };
 119         dead_group_list(u, b, &mq, GJ_THRES);
 120         if (pass_all_alive) {
 121                 for (unsigned int i = 0; i < mq.moves; i++) {
 122                         if (board_at(b, mq.move[i]) == stone_other(color)) {
 123                                 return false; // We need to remove opponent dead groups first.
 124                         }
 125                 }
 126                 mq.moves = 0; // our dead stones are alive when pass_all_alive is true
 127         }
 128         if (u->allow_losing_pass) {
 129                 foreach_point(b) {
 130                         if (board_at(b, c) == S_OFFBOARD)
 131                                 continue;
 132                         if (board_ownermap_judge_point(&u->ownermap, c, GJ_THRES) == PJ_UNKNOWN) {
 133                                 if (UDEBUGL(3))
 134                                         fprintf(stderr, "uct_pass_is_safe fails at %s[%d]\n", coord2sstr(c, b), c);
 135                                 return false; // Unclear point, clarify first.
 136                         }
 137                 } foreach_point_end;
 138                 return true;
 139         }
 140         return pass_is_safe(b, color, &mq);
 141 }
 142
 143 static void
 144 uct_board_print(struct engine *e, struct board *b, FILE *f)
 145 {
 146         struct uct *u = b->es;
 147         board_print_ownermap(b, f, (u ? &u->ownermap : NULL));
 148 }
 149
 150 static float
 151 uct_owner_map(struct engine *e, struct board *b, coord_t c)
 152 {
 153         struct uct *u = b->es;
 154         return board_ownermap_estimate_point(&u->ownermap, c);
 155 }
 156
 157 static char *
 158 uct_notify_play(struct engine *e, struct board *b, struct move *m, char *enginearg)
 159 {
 160         struct uct *u = e->data;
 161         if (!u->t) {
 162                 /* No state, create one - this is probably game beginning
 163                  * and we need to load the opening tbook right now. */
 164                 uct_prepare_move(u, b, m->color);
 165                 assert(u->t);
 166         }
 167
 168         /* Stop pondering, required by tree_promote_at() */
 169         uct_pondering_stop(u);
 170         if (UDEBUGL(2) && u->slave)
 171                 tree_dump(u->t, u->dumpthres);
 172
 173         if (is_resign(m->coord)) {
 174                 /* Reset state. */
 175                 reset_state(u);
 176                 return NULL;
 177         }
 178
 179         /* Promote node of the appropriate move to the tree root. */
 180         assert(u->t->root);
 181         if (u->t->untrustworthy_tree | !tree_promote_at(u->t, b, m->coord)) {
 182                 if (UDEBUGL(3)) {
 183                         if (u->t->untrustworthy_tree)
 184                                 fprintf(stderr, "Not promoting move node in untrustworthy tree.\n");
 185                         else
 186                                 fprintf(stderr, "Warning: Cannot promote move node! Several play commands in row?\n");
 187                 }
 188                 /* Preserve dynamic komi information, though, that is important. */
 189                 u->initial_extra_komi = u->t->extra_komi;
 190                 reset_state(u);
 191                 return NULL;
 192         }
 193
 194         /* If we are a slave in a distributed engine, start pondering once
 195          * we know which move we actually played. See uct_genmove() about
 196          * the check for pass. */
 197         if (u->pondering_opt && u->slave && m->color == u->my_color && !is_pass(m->coord))
 198                 uct_pondering_start(u, b, u->t, stone_other(m->color));
 199
 200         return NULL;
 201 }
 202
 203 static char *
 204 uct_undo(struct engine *e, struct board *b)
 205 {
 206         struct uct *u = e->data;
 207
 208         if (!u->t) return NULL;
 209         uct_pondering_stop(u);
 210         u->initial_extra_komi = u->t->extra_komi;
 211         reset_state(u);
 212         return NULL;
 213 }
 214
 215 static char *
 216 uct_result(struct engine *e, struct board *b)
 217 {
 218         struct uct *u = e->data;
 219         static char reply[1024];
 220
 221         if (!u->t)
 222                 return NULL;
 223         enum stone color = u->t->root_color;
 224         struct tree_node *n = u->t->root;
 225         snprintf(reply, 1024, "%s %s %d %.2f %.1f",
 226                  stone2str(color), coord2sstr(node_coord(n), b),
 227                  n->u.playouts, tree_node_get_value(u->t, -1, n->u.value),
 228                  u->t->use_extra_komi ? u->t->extra_komi : 0);
 229         return reply;
 230 }
 231
 232 static char *
 233 uct_chat(struct engine *e, struct board *b, bool opponent, char *from, char *cmd)
 234 {
 235         struct uct *u = e->data;
 236
 237         if (!u->t)
 238                 return generic_chat(b, opponent, from, cmd, S_NONE, pass, 0, 1, u->threads, 0.0, 0.0);
 239
 240         struct tree_node *n = u->t->root;
 241         double winrate = tree_node_get_value(u->t, -1, n->u.value);
 242         double extra_komi = u->t->use_extra_komi && fabs(u->t->extra_komi) >= 0.5 ? u->t->extra_komi : 0;
 243
 244         return generic_chat(b, opponent, from, cmd, u->t->root_color, node_coord(n), n->u.playouts, 1,
 245                             u->threads, winrate, extra_komi);
 246 }
 247
 248 static void
 249 print_dead_groups(struct uct *u, struct board *b, struct move_queue *mq)
 250 {
 251         fprintf(stderr, "dead groups (playing %s)\n", (u->my_color ? stone2str(u->my_color) : "???"));
 252         if (!mq->moves)
 253                 fprintf(stderr, "  none\n");
 254         for (unsigned int i = 0; i < mq->moves; i++) {
 255                 fprintf(stderr, "  ");
 256                 foreach_in_group(b, mq->move[i]) {
 257                         fprintf(stderr, "%s ", coord2sstr(c, b));
 258                 } foreach_in_group_end;
 259                 fprintf(stderr, "\n");
 260         }
 261 }
 262
 263 static void
 264 print_extra_dead_group(struct board *b, group_t g, int found)
 265 {
 266         if (!found)
 267                 fprintf(stderr, "also adding\n");
 268         fprintf(stderr, "  ");
 269         foreach_in_group(b, g) {
 270                 fprintf(stderr, "%s ", coord2sstr(c, b));
 271         } foreach_in_group_end;
 272         fprintf(stderr, "\n");
 273 }
 274
 275 static void
 276 uct_dead_group_list(struct engine *e, struct board *b, struct move_queue *mq)
 277 {
 278         struct uct *u = e->data;
 279         bool unknown_color = !u->my_color;
 280
 281         /* This means the game is probably over, no use pondering on. */
 282         uct_pondering_stop(u);
 283
 284         if (u->pass_all_alive)
 285                 return; // no dead groups
 286
 287         /* Create mock state */
 288         if (u->t)  reset_state(u);
 289         // We need S_BLACK here, but don't clobber u->my_color with uct_genmove_setup() !
 290         uct_prepare_move(u, b, S_BLACK);
 291
 292         /* Make sure the ownermap is well-seeded. */
 293         while (u->ownermap.playouts < GJ_MINGAMES)
 294                 uct_playout(u, b, S_BLACK, u->t);
 295         /* Show the ownermap: */
 296         if (DEBUGL(2))
 297                 board_print_ownermap(b, stderr, &u->ownermap);
 298
 299         struct move_queue relaxed; relaxed.moves = 0;
 300         dead_group_list(u, b, mq, GJ_THRES);    // Strict
 301         dead_group_list(u, b, &relaxed, 0.55);  // Relaxed
 302         if (DEBUGL(2))  print_dead_groups(u, b, mq);
 303
 304         /* Add own unclear dead groups if it doesn't change the outcome
 305          * and spare opponent a genmove_cleanup phase... */
 306         if (!unknown_color) {
 307                 int found = 0;
 308                 bool result = pass_is_safe(b, u->my_color, mq);
 309                 for (unsigned int i = 0; i < relaxed.moves; i++) {
 310                         group_t g = relaxed.move[i];
 311                         if (board_at(b, g) != u->my_color || mq_has(mq, g))
 312                                 continue;
 313
 314                         struct move_queue tmp;  memcpy(&tmp, mq, sizeof(tmp));
 315                         mq_add(&tmp, g, 0);
 316                         if (result == pass_is_safe(b, u->my_color, &tmp)) {
 317                                 mq_add(mq, g, 0);
 318                                 if (DEBUGL(2))  print_extra_dead_group(b, g, found++);
 319                         }
 320                 }
 321         }
 322
 323         /* Clean up the mock state in case we will receive
 324          * a genmove; we could get a non-alternating-move
 325          * error from uct_prepare_move() in that case otherwise. */
 326         reset_state(u);
 327 }
 328
 329 static void
 330 uct_stop(struct engine *e)
 331 {
 332         /* This is called on game over notification. However, an undo
 333          * and game resume can follow, so don't panic yet and just
 334          * relax and stop thinking so that we don't waste CPU. */
 335         struct uct *u = e->data;
 336         uct_pondering_stop(u);
 337 }
 338
 339 static void
 340 uct_done(struct engine *e)
 341 {
 342         /* This is called on engine reset, especially when clear_board
 343          * is received and new game should begin. */
 344         free(e->comment);
 345
 346         struct uct *u = e->data;
 347         uct_pondering_stop(u);
 348         if (u->t) reset_state(u);
 349         if (u->dynkomi) u->dynkomi->done(u->dynkomi);
 350         free(u->ownermap.map);
 351
 352         if (u->policy) u->policy->done(u->policy);
 353         if (u->random_policy) u->random_policy->done(u->random_policy);
 354         playout_policy_done(u->playout);
 355         uct_prior_done(u->prior);
 356         joseki_done(u->jdict);
 357         pluginset_done(u->plugins);
 358 }
 359
 360
 361
 362 /* Run time-limited MCTS search on foreground. */
 363 static int
 364 uct_search(struct uct *u, struct board *b, struct time_info *ti, enum stone color, struct tree *t, bool print_progress)
 365 {
 366         struct uct_search_state s;
 367         uct_search_start(u, b, color, t, ti, &s);
 368         if (UDEBUGL(2) && s.base_playouts > 0)
 369                 fprintf(stderr, "<pre-simulated %d games>\n", s.base_playouts);
 370
 371         /* The search tree is ctx->t. This is currently == . It is important
 372          * to reference ctx->t directly since the
 373          * thread manager will swap the tree pointer asynchronously. */
 374
 375         /* Now, just periodically poll the search tree. */
 376         /* Note that in case of TD_GAMES, threads will not wait for
 377          * the uct_search_check_stop() signalization. */
 378         while (1) {
 379                 time_sleep(TREE_BUSYWAIT_INTERVAL);
 380                 /* TREE_BUSYWAIT_INTERVAL should never be less than desired time, or the
 381                  * time control is broken. But if it happens to be less, we still search
 382                  * at least 100ms otherwise the move is completely random. */
 383
 384                 int i = uct_search_games(&s);
 385                 /* Print notifications etc. */
 386                 uct_search_progress(u, b, color, t, ti, &s, i);
 387                 /* Check if we should stop the search. */
 388                 if (uct_search_check_stop(u, b, color, t, ti, &s, i))
 389                         break;
 390         }
 391
 392         struct uct_thread_ctx *ctx = uct_search_stop();
 393         if (UDEBUGL(2)) tree_dump(t, u->dumpthres);
 394         if (UDEBUGL(2))
 395                 fprintf(stderr, "(avg score %f/%d; dynkomi's %f/%d value %f/%d)\n",
 396                         t->avg_score.value, t->avg_score.playouts,
 397                         u->dynkomi->score.value, u->dynkomi->score.playouts,
 398                         u->dynkomi->value.value, u->dynkomi->value.playouts);
 399         if (print_progress)
 400                 uct_progress_status(u, t, color, ctx->games, NULL);
 401
 402         if (u->debug_after.playouts > 0) {
 403                 /* Now, start an additional run of playouts, single threaded. */
 404                 struct time_info debug_ti = {
 405                         .period = TT_MOVE,
 406                         .dim = TD_GAMES,
 407                 };
 408                 debug_ti.len.games = t->root->u.playouts + u->debug_after.playouts;
 409
 410                 board_print_ownermap(b, stderr, &u->ownermap);
 411                 fprintf(stderr, "--8<-- UCT debug post-run begin (%d:%d) --8<--\n", u->debug_after.level, u->debug_after.playouts);
 412
 413                 int debug_level_save = debug_level;
 414                 int u_debug_level_save = u->debug_level;
 415                 int p_debug_level_save = u->playout->debug_level;
 416                 debug_level = u->debug_after.level;
 417                 u->debug_level = u->debug_after.level;
 418                 u->playout->debug_level = u->debug_after.level;
 419                 uct_halt = false;
 420
 421                 uct_playouts(u, b, color, t, &debug_ti);
 422                 tree_dump(t, u->dumpthres);
 423
 424                 uct_halt = true;
 425                 debug_level = debug_level_save;
 426                 u->debug_level = u_debug_level_save;
 427                 u->playout->debug_level = p_debug_level_save;
 428
 429                 fprintf(stderr, "--8<-- UCT debug post-run finished --8<--\n");
 430         }
 431
 432         u->played_own += ctx->games;
 433         return ctx->games;
 434 }
 435
 436 /* Start pondering background with @color to play. */
 437 static void
 438 uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color)
 439 {
 440         if (UDEBUGL(1))
 441                 fprintf(stderr, "Starting to ponder with color %s\n", stone2str(stone_other(color)));
 442         u->pondering = true;
 443
 444         /* We need a local board copy to ponder upon. */
 445         struct board *b = malloc2(sizeof(*b)); board_copy(b, b0);
 446
 447         /* *b0 did not have the genmove'd move played yet. */
 448         struct move m = { node_coord(t->root), t->root_color };
 449         int res = board_play(b, &m);
 450         assert(res >= 0);
 451         setup_dynkomi(u, b, stone_other(m.color));
 452
 453         /* Start MCTS manager thread "headless". */
 454         static struct uct_search_state s;
 455         uct_search_start(u, b, color, t, NULL, &s);
 456 }
 457
 458 /* uct_search_stop() frontend for the pondering (non-genmove) mode, and
 459  * to stop the background search for a slave in the distributed engine. */
 460 void
 461 uct_pondering_stop(struct uct *u)
 462 {
 463         if (!thread_manager_running)
 464                 return;
 465
 466         /* Stop the thread manager. */
 467         struct uct_thread_ctx *ctx = uct_search_stop();
 468         if (UDEBUGL(1)) {
 469                 if (u->pondering) fprintf(stderr, "(pondering) ");
 470                 uct_progress_status(u, ctx->t, ctx->color, ctx->games, NULL);
 471         }
 472         if (u->pondering) {
 473                 free(ctx->b);
 474                 u->pondering = false;
 475         }
 476 }
 477
 478
 479 void
 480 uct_genmove_setup(struct uct *u, struct board *b, enum stone color)
 481 {
 482         if (b->superko_violation) {
 483                 fprintf(stderr, "!!! WARNING: SUPERKO VIOLATION OCCURED BEFORE THIS MOVE\n");
 484                 fprintf(stderr, "Maybe you play with situational instead of positional superko?\n");
 485                 fprintf(stderr, "I'm going to ignore the violation, but note that I may miss\n");
 486                 fprintf(stderr, "some moves valid under this ruleset because of this.\n");
 487                 b->superko_violation = false;
 488         }
 489
 490         uct_prepare_move(u, b, color);
 491
 492         assert(u->t);
 493         u->my_color = color;
 494
 495         /* How to decide whether to use dynkomi in this game? Since we use
 496          * pondering, it's not simple "who-to-play" matter. Decide based on
 497          * the last genmove issued. */
 498         u->t->use_extra_komi = !!(u->dynkomi_mask & color);
 499         setup_dynkomi(u, b, color);
 500
 501         if (b->rules == RULES_JAPANESE)
 502                 u->territory_scoring = true;
 503
 504         /* Make pessimistic assumption about komi for Japanese rules to
 505          * avoid losing by 0.5 when winning by 0.5 with Chinese rules.
 506          * The rules usually give the same winner if the integer part of komi
 507          * is odd so we adjust the komi only if it is even (for a board of
 508          * odd size). We are not trying  to get an exact evaluation for rare
 509          * cases of seki. For details see http://home.snafu.de/jasiek/parity.html */
 510         if (u->territory_scoring && (((int)floor(b->komi) + board_size(b)) & 1)) {
 511                 b->komi += (color == S_BLACK ? 1.0 : -1.0);
 512                 if (UDEBUGL(0))
 513                         fprintf(stderr, "Setting komi to %.1f assuming Japanese rules\n",
 514                                 b->komi);
 515         }
 516 }
 517
 518 static void
 519 uct_live_gfx_hook(struct engine *e)
 520 {
 521         struct uct *u = e->data;
 522         /* Hack: Override reportfreq to get decent update rates in GoGui */
 523         u->reportfreq = 1000;
 524 }
 525
 526 /* Kindof like uct_genmove() but just find the best candidates */
 527 static void
 528 uct_best_moves(struct engine *e, struct board *b, enum stone color)
 529 {
 530         struct time_info ti = { .period = TT_NULL };
 531         double start_time = time_now();
 532         struct uct *u = e->data;
 533         uct_pondering_stop(u);
 534         if (u->t)
 535                 reset_state(u);
 536         uct_genmove_setup(u, b, color);
 537
 538         /* Start the Monte Carlo Tree Search! */
 539         int base_playouts = u->t->root->u.playouts;
 540         int played_games = uct_search(u, b, &ti, color, u->t, false);
 541
 542         coord_t best_coord;
 543         uct_search_result(u, b, color, u->pass_all_alive, played_games, base_playouts, &best_coord);
 544
 545         if (UDEBUGL(2)) {
 546                 double time = time_now() - start_time + 0.000001; /* avoid divide by zero */
 547                 fprintf(stderr, "genmove in %0.2fs (%d games/s, %d games/s/thread)\n",
 548                         time, (int)(played_games/time), (int)(played_games/time/u->threads));
 549         }
 550
 551         uct_progress_status(u, u->t, color, played_games, &best_coord);
 552         reset_state(u);
 553 }
 554
 555 static coord_t *
 556 uct_genmove(struct engine *e, struct board *b, struct time_info *ti, enum stone color, bool pass_all_alive)
 557 {
 558         double start_time = time_now();
 559         struct uct *u = e->data;
 560         u->pass_all_alive |= pass_all_alive;
 561         uct_pondering_stop(u);
 562
 563         if (using_dcnn(b)) {
 564                 // dcnn hack: reset state to make dcnn priors kick in.
 565                 // FIXME this makes pondering useless when using dcnn ...
 566                 if (u->t) {
 567                         u->initial_extra_komi = u->t->extra_komi;
 568                         reset_state(u);
 569                 }
 570         }
 571
 572         uct_genmove_setup(u, b, color);
 573
 574         /* Start the Monte Carlo Tree Search! */
 575         int base_playouts = u->t->root->u.playouts;
 576         int played_games = uct_search(u, b, ti, color, u->t, false);
 577
 578         coord_t best_coord;
 579         struct tree_node *best;
 580         best = uct_search_result(u, b, color, u->pass_all_alive, played_games, base_playouts, &best_coord);
 581
 582         if (UDEBUGL(2)) {
 583                 double time = time_now() - start_time + 0.000001; /* avoid divide by zero */
 584                 fprintf(stderr, "genmove in %0.2fs (%d games/s, %d games/s/thread)\n",
 585                         time, (int)(played_games/time), (int)(played_games/time/u->threads));
 586         }
 587
 588         uct_progress_status(u, u->t, color, played_games, &best_coord);
 589
 590         if (!best) {
 591                 /* Pass or resign. */
 592                 if (is_pass(best_coord))
 593                         u->initial_extra_komi = u->t->extra_komi;
 594                 reset_state(u);
 595                 return coord_copy(best_coord);
 596         }
 597
 598         if (!u->t->untrustworthy_tree) {
 599                 tree_promote_node(u->t, &best);
 600         } else {
 601                 /* Throw away an untrustworthy tree. */
 602                 /* Preserve dynamic komi information, though, that is important. */
 603                 u->initial_extra_komi = u->t->extra_komi;
 604                 reset_state(u);
 605         }
 606
 607         /* After a pass, pondering is harmful for two reasons:
 608          * (i) We might keep pondering even when the game is over.
 609          * Of course this is the case for opponent resign as well.
 610          * (ii) More importantly, the ownermap will get skewed since
 611          * the UCT will start cutting off any playouts. */
 612         if (u->pondering_opt && u->t && !is_pass(node_coord(best))) {
 613                 uct_pondering_start(u, b, u->t, stone_other(color));
 614         }
 615         return coord_copy(best_coord);
 616 }
 617
 618
 619 bool
 620 uct_gentbook(struct engine *e, struct board *b, struct time_info *ti, enum stone color)
 621 {
 622         struct uct *u = e->data;
 623         if (!u->t) uct_prepare_move(u, b, color);
 624         assert(u->t);
 625
 626         if (ti->dim == TD_GAMES) {
 627                 /* Don't count in games that already went into the tbook. */
 628                 ti->len.games += u->t->root->u.playouts;
 629         }
 630         uct_search(u, b, ti, color, u->t, true);
 631
 632         assert(ti->dim == TD_GAMES);
 633         tree_save(u->t, b, ti->len.games / 100);
 634
 635         return true;
 636 }
 637
 638 void
 639 uct_dumptbook(struct engine *e, struct board *b, enum stone color)
 640 {
 641         struct uct *u = e->data;
 642         struct tree *t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0,
 643                          u->max_pruned_size, u->pruning_threshold, u->local_tree_aging, 0);
 644         tree_load(t, b);
 645         tree_dump(t, 0);
 646         tree_done(t);
 647 }
 648
 649
 650 floating_t
 651 uct_evaluate_one(struct engine *e, struct board *b, struct time_info *ti, coord_t c, enum stone color)
 652 {
 653         struct uct *u = e->data;
 654
 655         struct board b2;
 656         board_copy(&b2, b);
 657         struct move m = { c, color };
 658         int res = board_play(&b2, &m);
 659         if (res < 0)
 660                 return NAN;
 661         color = stone_other(color);
 662
 663         if (u->t) reset_state(u);
 664         uct_prepare_move(u, &b2, color);
 665         assert(u->t);
 666
 667         floating_t bestval;
 668         uct_search(u, &b2, ti, color, u->t, true);
 669         struct tree_node *best = u->policy->choose(u->policy, u->t->root, &b2, color, resign);
 670         if (!best) {
 671                 bestval = NAN; // the opponent has no reply!
 672         } else {
 673                 bestval = tree_node_get_value(u->t, 1, best->u.value);
 674         }
 675
 676         reset_state(u); // clean our junk
 677
 678         return isnan(bestval) ? NAN : 1.0f - bestval;
 679 }
 680
 681 void
 682 uct_evaluate(struct engine *e, struct board *b, struct time_info *ti, floating_t *vals, enum stone color)
 683 {
 684         for (int i = 0; i < b->flen; i++) {
 685                 if (is_pass(b->f[i]))
 686                         vals[i] = NAN;
 687                 else
 688                         vals[i] = uct_evaluate_one(e, b, ti, b->f[i], color);
 689         }
 690 }
 691
 692
 693 struct uct *
 694 uct_state_init(char *arg, struct board *b)
 695 {
 696         struct uct *u = calloc2(1, sizeof(struct uct));
 697         bool pat_setup = false;
 698
 699         u->debug_level = debug_level;
 700         u->reportfreq = 10000;
 701         u->gamelen = MC_GAMELEN;
 702         u->resign_threshold = 0.2;
 703         u->sure_win_threshold = 0.95;
 704         u->mercymin = 0;
 705         u->significant_threshold = 50;
 706         u->expand_p = 8;
 707         u->dumpthres = 0.01;
 708         u->playout_amaf = true;
 709         u->amaf_prior = false;
 710         u->max_tree_size = 1408ULL * 1048576;
 711         u->fast_alloc = true;
 712         u->pruning_threshold = 0;
 713
 714         u->threads = 1;
 715         u->thread_model = TM_TREEVL;
 716         u->virtual_loss = 1;
 717
 718         u->pondering_opt = true;
 719
 720         u->fuseki_end = 20; // max time at 361*20% = 72 moves (our 36th move, still 99 to play)
 721         u->yose_start = 40; // (100-40-25)*361/100/2 = 63 moves still to play by us then
 722         u->bestr_ratio = 0.02;
 723         // 2.5 is clearly too much, but seems to compensate well for overly stern time allocations.
 724         // TODO: Further tuning and experiments with better time allocation schemes.
 725         u->best2_ratio = 2.5;
 726         // Higher values of max_maintime_ratio sometimes cause severe time trouble in tournaments
 727         // It might be necessary to reduce it to 1.5 on large board, but more tuning is needed.
 728         u->max_maintime_ratio = 2.0;
 729
 730         u->val_scale = 0; u->val_points = 40;
 731         u->dynkomi_interval = 1000;
 732         u->dynkomi_mask = S_BLACK | S_WHITE;
 733
 734         u->tenuki_d = 4;
 735         u->local_tree_aging = 80;
 736         u->local_tree_depth_decay = 1.5;
 737         u->local_tree_eval = LTE_ROOT;
 738         u->local_tree_neival = true;
 739
 740         u->max_slaves = -1;
 741         u->slave_index = -1;
 742         u->stats_delay = 0.01; // 10 ms
 743         u->shared_levels = 1;
 744
 745         u->plugins = pluginset_init(b);
 746
 747         u->jdict = joseki_load(b->size);
 748
 749         if (arg) {
 750                 char *optspec, *next = arg;
 751                 while (*next) {
 752                         optspec = next;
 753                         next += strcspn(next, ",");
 754                         if (*next) { *next++ = 0; } else { *next = 0; }
 755
 756                         char *optname = optspec;
 757                         char *optval = strchr(optspec, '=');
 758                         if (optval) *optval++ = 0;
 759
 760                         /** Basic options */
 761
 762                         if (!strcasecmp(optname, "debug")) {
 763                                 if (optval)
 764                                         u->debug_level = atoi(optval);
 765                                 else
 766                                         u->debug_level++;
 767                         } else if (!strcasecmp(optname, "reporting") && optval) {
 768                                 /* The format of output for detailed progress
 769                                  * information (such as current best move and
 770                                  * its value, etc.). */
 771                                 if (!strcasecmp(optval, "text")) {
 772                                         /* Plaintext traditional output. */
 773                                         u->reporting = UR_TEXT;
 774                                 } else if (!strcasecmp(optval, "json")) {
 775                                         /* JSON output. Implies debug=0. */
 776                                         u->reporting = UR_JSON;
 777                                         u->debug_level = 0;
 778                                 } else if (!strcasecmp(optval, "jsonbig")) {
 779                                         /* JSON output, but much more detailed.
 780                                          * Implies debug=0. */
 781                                         u->reporting = UR_JSON_BIG;
 782                                         u->debug_level = 0;
 783                                 } else {
 784                                         fprintf(stderr, "UCT: Invalid reporting format %s\n", optval);
 785                                         exit(1);
 786                                 }
 787                         } else if (!strcasecmp(optname, "reportfreq") && optval) {
 788                                 /* The progress information line will be shown
 789                                  * every <reportfreq> simulations. */
 790                                 u->reportfreq = atoi(optval);
 791                         } else if (!strcasecmp(optname, "dumpthres") && optval) {
 792                                 /* When dumping the UCT tree on output, include
 793                                  * nodes with at least this many playouts.
 794                                  * (A fraction of the total # of playouts at the
 795                                  * tree root.) */
 796                                 /* Use 0 to list all nodes with at least one
 797                                  * simulation, and -1 to list _all_ nodes. */
 798                                 u->dumpthres = atof(optval);
 799                         } else if (!strcasecmp(optname, "resign_threshold") && optval) {
 800                                 /* Resign when this ratio of games is lost
 801                                  * after GJ_MINGAMES sample is taken. */
 802                                 u->resign_threshold = atof(optval);
 803                         } else if (!strcasecmp(optname, "sure_win_threshold") && optval) {
 804                                 /* Stop reading when this ratio of games is won
 805                                  * after PLAYOUT_EARLY_BREAK_MIN sample is
 806                                  * taken. (Prevents stupid time losses,
 807                                  * friendly to human opponents.) */
 808                                 u->sure_win_threshold = atof(optval);
 809                         } else if (!strcasecmp(optname, "force_seed") && optval) {
 810                                 /* Set RNG seed at the tree setup. */
 811                                 u->force_seed = atoi(optval);
 812                         } else if (!strcasecmp(optname, "no_tbook")) {
 813                                 /* Disable UCT opening tbook. */
 814                                 u->no_tbook = true;
 815                         } else if (!strcasecmp(optname, "pass_all_alive")) {
 816                                 /* Whether to consider passing only after all
 817                                  * dead groups were removed from the board;
 818                                  * this is like all genmoves are in fact
 819                                  * kgs-genmove_cleanup. */
 820                                 u->pass_all_alive = !optval || atoi(optval);
 821                         } else if (!strcasecmp(optname, "allow_losing_pass")) {
 822                                 /* Whether to consider passing in a clear
 823                                  * but losing situation, to be scored as a loss
 824                                  * for us. */
 825                                 u->allow_losing_pass = !optval || atoi(optval);
 826                         } else if (!strcasecmp(optname, "territory_scoring")) {
 827                                 /* Use territory scoring (default is area scoring).
 828                                  * An explicit kgs-rules command overrides this. */
 829                                 u->territory_scoring = !optval || atoi(optval);
 830                         } else if (!strcasecmp(optname, "stones_only")) {
 831                                 /* Do not count eyes. Nice to teach go to kids.
 832                                  * http://strasbourg.jeudego.org/regle_strasbourgeoise.htm */
 833                                 b->rules = RULES_STONES_ONLY;
 834                                 u->pass_all_alive = true;
 835                         } else if (!strcasecmp(optname, "debug_after")) {
 836                                 /* debug_after=9:1000 will make Pachi think under
 837                                  * the normal conditions, but at the point when
 838                                  * a move is to be chosen, the tree is dumped and
 839                                  * another 1000 simulations are run single-threaded
 840                                  * with debug level 9, allowing inspection of Pachi's
 841                                  * behavior after it has thought a lot. */
 842                                 if (optval) {
 843                                         u->debug_after.level = atoi(optval);
 844                                         char *playouts = strchr(optval, ':');
 845                                         if (playouts)
 846                                                 u->debug_after.playouts = atoi(playouts+1);
 847                                         else
 848                                                 u->debug_after.playouts = 1000;
 849                                 } else {
 850                                         u->debug_after.level = 9;
 851                                         u->debug_after.playouts = 1000;
 852                                 }
 853                         } else if (!strcasecmp(optname, "banner") && optval) {
 854                                 /* Additional banner string. This must come as the
 855                                  * last engine parameter. You can use '+' instead
 856                                  * of ' ' if you are wrestling with kgsGtp. */
 857                                 if (*next) *--next = ',';
 858                                 u->banner = strdup(optval);
 859                                 for (char *b = u->banner; *b; b++) {
 860                                         if (*b == '+') *b = ' ';
 861                                 }
 862                                 break;
 863                         } else if (!strcasecmp(optname, "plugin") && optval) {
 864                                 /* Load an external plugin; filename goes before the colon,
 865                                  * extra arguments after the colon. */
 866                                 char *pluginarg = strchr(optval, ':');
 867                                 if (pluginarg)
 868                                         *pluginarg++ = 0;
 869                                 plugin_load(u->plugins, optval, pluginarg);
 870
 871                         /** UCT behavior and policies */
 872
 873                         } else if ((!strcasecmp(optname, "policy")
 874                                 /* Node selection policy. ucb1amaf is the
 875                                  * default policy implementing RAVE, while
 876                                  * ucb1 is the simple exploration/exploitation
 877                                  * policy. Policies can take further extra
 878                                  * options. */
 879                                     || !strcasecmp(optname, "random_policy")) && optval) {
 880                                 /* A policy to be used randomly with small
 881                                  * chance instead of the default policy. */
 882                                 char *policyarg = strchr(optval, ':');
 883                                 struct uct_policy **p = !strcasecmp(optname, "policy") ? &u->policy : &u->random_policy;
 884                                 if (policyarg)
 885                                         *policyarg++ = 0;
 886                                 if (!strcasecmp(optval, "ucb1")) {
 887                                         *p = policy_ucb1_init(u, policyarg);
 888                                 } else if (!strcasecmp(optval, "ucb1amaf")) {
 889                                         *p = policy_ucb1amaf_init(u, policyarg, b);
 890                                 } else {
 891                                         fprintf(stderr, "UCT: Invalid tree policy %s\n", optval);
 892                                         exit(1);
 893                                 }
 894                         } else if (!strcasecmp(optname, "playout") && optval) {
 895                                 /* Random simulation (playout) policy.
 896                                  * moggy is the default policy with large
 897                                  * amount of domain-specific knowledge and
 898                                  * heuristics. light is a simple uniformly
 899                                  * random move selection policy. */
 900                                 char *playoutarg = strchr(optval, ':');
 901                                 if (playoutarg)
 902                                         *playoutarg++ = 0;
 903                                 if (!strcasecmp(optval, "moggy")) {
 904                                         u->playout = playout_moggy_init(playoutarg, b, u->jdict);
 905                                 } else if (!strcasecmp(optval, "light")) {
 906                                         u->playout = playout_light_init(playoutarg, b);
 907                                 } else {
 908                                         fprintf(stderr, "UCT: Invalid playout policy %s\n", optval);
 909                                         exit(1);
 910                                 }
 911                         } else if (!strcasecmp(optname, "prior") && optval) {
 912                                 /* Node priors policy. When expanding a node,
 913                                  * it will seed node values heuristically
 914                                  * (most importantly, based on playout policy
 915                                  * opinion, but also with regard to other
 916                                  * things). See uct/prior.c for details.
 917                                  * Use prior=eqex=0 to disable priors. */
 918                                 u->prior = uct_prior_init(optval, b, u);
 919                         } else if (!strcasecmp(optname, "mercy") && optval) {
 920                                 /* Minimal difference of black/white captures
 921                                  * to stop playout - "Mercy Rule". Speeds up
 922                                  * hopeless playouts at the expense of some
 923                                  * accuracy. */
 924                                 u->mercymin = atoi(optval);
 925                         } else if (!strcasecmp(optname, "gamelen") && optval) {
 926                                 /* Maximum length of single simulation
 927                                  * in moves. */
 928                                 u->gamelen = atoi(optval);
 929                         } else if (!strcasecmp(optname, "expand_p") && optval) {
 930                                 /* Expand UCT nodes after it has been
 931                                  * visited this many times. */
 932                                 u->expand_p = atoi(optval);
 933                         } else if (!strcasecmp(optname, "random_policy_chance") && optval) {
 934                                 /* If specified (N), with probability 1/N, random_policy policy
 935                                  * descend is used instead of main policy descend; useful
 936                                  * if specified policy (e.g. UCB1AMAF) can make unduly biased
 937                                  * choices sometimes, you can fall back to e.g.
 938                                  * random_policy=UCB1. */
 939                                 u->random_policy_chance = atoi(optval);
 940
 941                         /** General AMAF behavior */
 942                         /* (Only relevant if the policy supports AMAF.
 943                          * More variables can be tuned as policy
 944                          * parameters.) */
 945
 946                         } else if (!strcasecmp(optname, "playout_amaf")) {
 947                                 /* Whether to include random playout moves in
 948                                  * AMAF as well. (Otherwise, only tree moves
 949                                  * are included in AMAF. Of course makes sense
 950                                  * only in connection with an AMAF policy.) */
 951                                 /* with-without: 55.5% (+-4.1) */
 952                                 if (optval && *optval == '0')
 953                                         u->playout_amaf = false;
 954                                 else
 955                                         u->playout_amaf = true;
 956                         } else if (!strcasecmp(optname, "playout_amaf_cutoff") && optval) {
 957                                 /* Keep only first N% of playout stage AMAF
 958                                  * information. */
 959                                 u->playout_amaf_cutoff = atoi(optval);
 960                         } else if (!strcasecmp(optname, "amaf_prior") && optval) {
 961                                 /* In node policy, consider prior values
 962                                  * part of the real result term or part
 963                                  * of the AMAF term? */
 964                                 u->amaf_prior = atoi(optval);
 965
 966                         /** Performance and memory management */
 967
 968                         } else if (!strcasecmp(optname, "threads") && optval) {
 969                                 /* By default, Pachi will run with only single
 970                                  * tree search thread! */
 971                                 u->threads = atoi(optval);
 972                         } else if (!strcasecmp(optname, "thread_model") && optval) {
 973                                 if (!strcasecmp(optval, "tree")) {
 974                                         /* Tree parallelization - all threads
 975                                          * grind on the same tree. */
 976                                         u->thread_model = TM_TREE;
 977                                         u->virtual_loss = 0;
 978                                 } else if (!strcasecmp(optval, "treevl")) {
 979                                         /* Tree parallelization, but also
 980                                          * with virtual losses - this discou-
 981                                          * rages most threads choosing the
 982                                          * same tree branches to read. */
 983                                         u->thread_model = TM_TREEVL;
 984                                 } else {
 985                                         fprintf(stderr, "UCT: Invalid thread model %s\n", optval);
 986                                         exit(1);
 987                                 }
 988                         } else if (!strcasecmp(optname, "virtual_loss") && optval) {
 989                                 /* Number of virtual losses added before evaluating a node. */
 990                                 u->virtual_loss = atoi(optval);
 991                         } else if (!strcasecmp(optname, "pondering")) {
 992                                 /* Keep searching even during opponent's turn. */
 993                                 u->pondering_opt = !optval || atoi(optval);
 994                         } else if (!strcasecmp(optname, "max_tree_size") && optval) {
 995                                 /* Maximum amount of memory [MiB] consumed by the move tree.
 996                                  * For fast_alloc it includes the temp tree used for pruning.
 997                                  * Default is 3072 (3 GiB). */
 998                                 u->max_tree_size = atol(optval) * 1048576;
 999                         } else if (!strcasecmp(optname, "fast_alloc")) {
1000                                 u->fast_alloc = !optval || atoi(optval);
1001                         } else if (!strcasecmp(optname, "pruning_threshold") && optval) {
1002                                 /* Force pruning at beginning of a move if the tree consumes
1003                                  * more than this [MiB]. Default is 10% of max_tree_size.
1004                                  * Increase to reduce pruning time overhead if memory is plentiful.
1005                                  * This option is meaningful only for fast_alloc. */
1006                                 u->pruning_threshold = atol(optval) * 1048576;
1007
1008                         /** Time control */
1009
1010                         } else if (!strcasecmp(optname, "best2_ratio") && optval) {
1011                                 /* If set, prolong simulating while
1012                                  * first_best/second_best playouts ratio
1013                                  * is less than best2_ratio. */
1014                                 u->best2_ratio = atof(optval);
1015                         } else if (!strcasecmp(optname, "bestr_ratio") && optval) {
1016                                 /* If set, prolong simulating while
1017                                  * best,best_best_child values delta
1018                                  * is more than bestr_ratio. */
1019                                 u->bestr_ratio = atof(optval);
1020                         } else if (!strcasecmp(optname, "max_maintime_ratio") && optval) {
1021                                 /* If set and while not in byoyomi, prolong simulating no more than
1022                                  * max_maintime_ratio times the normal desired thinking time. */
1023                                 u->max_maintime_ratio = atof(optval);
1024                         } else if (!strcasecmp(optname, "fuseki_end") && optval) {
1025                                 /* At the very beginning it's not worth thinking
1026                                  * too long because the playout evaluations are
1027                                  * very noisy. So gradually increase the thinking
1028                                  * time up to maximum when fuseki_end percent
1029                                  * of the board has been played.
1030                                  * This only applies if we are not in byoyomi. */
1031                                 u->fuseki_end = atoi(optval);
1032                         } else if (!strcasecmp(optname, "yose_start") && optval) {
1033                                 /* When yose_start percent of the board has been
1034                                  * played, or if we are in byoyomi, stop spending
1035                                  * more time and spread the remaining time
1036                                  * uniformly.
1037                                  * Between fuseki_end and yose_start, we spend
1038                                  * a constant proportion of the remaining time
1039                                  * on each move. (yose_start should actually
1040                                  * be much earlier than when real yose start,
1041                                  * but "yose" is a good short name to convey
1042                                  * the idea.) */
1043                                 u->yose_start = atoi(optval);
1044
1045                         /** Dynamic komi */
1046
1047                         } else if (!strcasecmp(optname, "dynkomi") && optval) {
1048                                 /* Dynamic komi approach; there are multiple
1049                                  * ways to adjust komi dynamically throughout
1050                                  * play. We currently support two: */
1051                                 char *dynkomiarg = strchr(optval, ':');
1052                                 if (dynkomiarg)
1053                                         *dynkomiarg++ = 0;
1054                                 if (!strcasecmp(optval, "none")) {
1055                                         u->dynkomi = uct_dynkomi_init_none(u, dynkomiarg, b);
1056                                 } else if (!strcasecmp(optval, "linear")) {
1057                                         /* You should set dynkomi_mask=1 or a very low
1058                                          * handicap_value for white. */
1059                                         u->dynkomi = uct_dynkomi_init_linear(u, dynkomiarg, b);
1060                                 } else if (!strcasecmp(optval, "adaptive")) {
1061                                         /* There are many more knobs to
1062                                          * crank - see uct/dynkomi.c. */
1063                                         u->dynkomi = uct_dynkomi_init_adaptive(u, dynkomiarg, b);
1064                                 } else {
1065                                         fprintf(stderr, "UCT: Invalid dynkomi mode %s\n", optval);
1066                                         exit(1);
1067                                 }
1068                         } else if (!strcasecmp(optname, "dynkomi_mask") && optval) {
1069                                 /* Bitmask of colors the player must be
1070                                  * for dynkomi be applied; the default dynkomi_mask=3 allows
1071                                  * dynkomi even in games where Pachi is white. */
1072                                 u->dynkomi_mask = atoi(optval);
1073                         } else if (!strcasecmp(optname, "dynkomi_interval") && optval) {
1074                                 /* If non-zero, re-adjust dynamic komi
1075                                  * throughout a single genmove reading,
1076                                  * roughly every N simulations. */
1077                                 /* XXX: Does not work with tree
1078                                  * parallelization. */
1079                                 u->dynkomi_interval = atoi(optval);
1080                         } else if (!strcasecmp(optname, "extra_komi") && optval) {
1081                                 /* Initial dynamic komi settings. This
1082                                  * is useful for the adaptive dynkomi
1083                                  * policy as the value to start with
1084                                  * (this is NOT kept fixed) in case
1085                                  * there is not enough time in the search
1086                                  * to adjust the value properly (e.g. the
1087                                  * game was interrupted). */
1088                                 u->initial_extra_komi = atof(optval);
1089
1090                         /** Node value result scaling */
1091
1092                         } else if (!strcasecmp(optname, "val_scale") && optval) {
1093                                 /* How much of the game result value should be
1094                                  * influenced by win size. Zero means it isn't. */
1095                                 u->val_scale = atof(optval);
1096                         } else if (!strcasecmp(optname, "val_points") && optval) {
1097                                 /* Maximum size of win to be scaled into game
1098                                  * result value. Zero means boardsize^2. */
1099                                 u->val_points = atoi(optval) * 2; // result values are doubled
1100                         } else if (!strcasecmp(optname, "val_extra")) {
1101                                 /* If false, the score coefficient will be simply
1102                                  * added to the value, instead of scaling the result
1103                                  * coefficient because of it. */
1104                                 u->val_extra = !optval || atoi(optval);
1105                         } else if (!strcasecmp(optname, "val_byavg")) {
1106                                 /* If true, the score included in the value will
1107                                  * be relative to average score in the current
1108                                  * search episode inst. of jigo. */
1109                                 u->val_byavg = !optval || atoi(optval);
1110                         } else if (!strcasecmp(optname, "val_bytemp")) {
1111                                 /* If true, the value scaling coefficient
1112                                  * is different based on value extremity
1113                                  * (dist. from 0.5), linear between
1114                                  * val_bytemp_min, val_scale. */
1115                                 u->val_bytemp = !optval || atoi(optval);
1116                         } else if (!strcasecmp(optname, "val_bytemp_min") && optval) {
1117                                 /* Minimum val_scale in case of val_bytemp. */
1118                                 u->val_bytemp_min = atof(optval);
1119
1120                         /** Local trees */
1121                         /* (Purely experimental. Does not work - yet!) */
1122
1123                         } else if (!strcasecmp(optname, "local_tree")) {
1124                                 /* Whether to bias exploration by local tree values. */
1125                                 u->local_tree = !optval || atoi(optval);
1126                         } else if (!strcasecmp(optname, "tenuki_d") && optval) {
1127                                 /* Tenuki distance at which to break the local tree. */
1128                                 u->tenuki_d = atoi(optval);
1129                                 if (u->tenuki_d > TREE_NODE_D_MAX + 1) {
1130                                         fprintf(stderr, "uct: tenuki_d must not be larger than TREE_NODE_D_MAX+1 %d\n", TREE_NODE_D_MAX + 1);
1131                                         exit(1);
1132                                 }
1133                         } else if (!strcasecmp(optname, "local_tree_aging") && optval) {
1134                                 /* How much to reduce local tree values between moves. */
1135                                 u->local_tree_aging = atof(optval);
1136                         } else if (!strcasecmp(optname, "local_tree_depth_decay") && optval) {
1137                                 /* With value x>0, during the descent the node
1138                                  * contributes 1/x^depth playouts in
1139                                  * the local tree. I.e., with x>1, nodes more
1140                                  * distant from local situation contribute more
1141                                  * than nodes near the root. */
1142                                 u->local_tree_depth_decay = atof(optval);
1143                         } else if (!strcasecmp(optname, "local_tree_allseq")) {
1144                                 /* If disabled, only complete sequences are stored
1145                                  * in the local tree. If this is on, also
1146                                  * subsequences starting at each move are stored. */
1147                                 u->local_tree_allseq = !optval || atoi(optval);
1148                         } else if (!strcasecmp(optname, "local_tree_neival")) {
1149                                 /* If disabled, local node value is not
1150                                  * computed just based on terminal status
1151                                  * of the coordinate, but also its neighbors. */
1152                                 u->local_tree_neival = !optval || atoi(optval);
1153                         } else if (!strcasecmp(optname, "local_tree_eval")) {
1154                                 /* How is the value inserted in the local tree
1155                                  * determined. */
1156                                 if (!strcasecmp(optval, "root"))
1157                                         /* All moves within a tree branch are
1158                                          * considered wrt. their merit
1159                                          * reaching tachtical goal of making
1160                                          * the first move in the branch
1161                                          * survive. */
1162                                         u->local_tree_eval = LTE_ROOT;
1163                                 else if (!strcasecmp(optval, "each"))
1164                                         /* Each move is considered wrt.
1165                                          * its own survival. */
1166                                         u->local_tree_eval = LTE_EACH;
1167                                 else if (!strcasecmp(optval, "total"))
1168                                         /* The tactical goal is the survival
1169                                          * of all the moves of my color and
1170                                          * non-survival of all the opponent
1171                                          * moves. Local values (and their
1172                                          * inverses) are averaged. */
1173                                         u->local_tree_eval = LTE_TOTAL;
1174                                 else {
1175                                         fprintf(stderr, "uct: unknown local_tree_eval %s\n", optval);
1176                                         exit(1);
1177                                 }
1178                         } else if (!strcasecmp(optname, "local_tree_rootchoose")) {
1179                                 /* If disabled, only moves within the local
1180                                  * tree branch are considered; the values
1181                                  * of the branch roots (i.e. root children)
1182                                  * are ignored. This may make sense together
1183                                  * with eval!=each, we consider only moves
1184                                  * that influence the goal, not the "rating"
1185                                  * of the goal itself. (The real solution
1186                                  * will be probably using criticality to pick
1187                                  * local tree branches.) */
1188                                 u->local_tree_rootchoose = !optval || atoi(optval);
1189
1190                         /** Other heuristics */
1191                         } else if (!strcasecmp(optname, "patterns")) {
1192                                 /* Load pattern database. Various modules
1193                                  * (priors, policies etc.) may make use
1194                                  * of this database. They will request
1195                                  * it automatically in that case, but you
1196                                  * can use this option to tweak the pattern
1197                                  * parameters. */
1198                                 patterns_init(&u->pat, optval, false, true);
1199                                 u->want_pat = pat_setup = true;
1200                         } else if (!strcasecmp(optname, "significant_threshold") && optval) {
1201                                 /* Some heuristics (XXX: none in mainline) rely
1202                                  * on the knowledge of the last "significant"
1203                                  * node in the descent. Such a node is
1204                                  * considered reasonably trustworthy to carry
1205                                  * some meaningful information in the values
1206                                  * of the node and its children. */
1207                                 u->significant_threshold = atoi(optval);
1208
1209                         /** Distributed engine slaves setup */
1210
1211                         } else if (!strcasecmp(optname, "slave")) {
1212                                 /* Act as slave for the distributed engine. */
1213                                 u->slave = !optval || atoi(optval);
1214                         } else if (!strcasecmp(optname, "slave_index") && optval) {
1215                                 /* Optional index if per-slave behavior is desired.
1216                                  * Must be given as index/max */
1217                                 u->slave_index = atoi(optval);
1218                                 char *p = strchr(optval, '/');
1219                                 if (p) u->max_slaves = atoi(++p);
1220                         } else if (!strcasecmp(optname, "shared_nodes") && optval) {
1221                                 /* Share at most shared_nodes between master and slave at each genmoves.
1222                                  * Must use the same value in master and slaves. */
1223                                 u->shared_nodes = atoi(optval);
1224                         } else if (!strcasecmp(optname, "shared_levels") && optval) {
1225                                 /* Share only nodes of level <= shared_levels. */
1226                                 u->shared_levels = atoi(optval);
1227                         } else if (!strcasecmp(optname, "stats_hbits") && optval) {
1228                                 /* Set hash table size to 2^stats_hbits for the shared stats. */
1229                                 u->stats_hbits = atoi(optval);
1230                         } else if (!strcasecmp(optname, "stats_delay") && optval) {
1231                                 /* How long to wait in slave for initial stats to build up before
1232                                  * replying to the genmoves command (in ms) */
1233                                 u->stats_delay = 0.001 * atof(optval);
1234
1235                         /** Presets */
1236
1237                         } else if (!strcasecmp(optname, "maximize_score")) {
1238                                 /* A combination of settings that will make
1239                                  * Pachi try to maximize his points (instead
1240                                  * of playing slack yose) or minimize his loss
1241                                  * (and proceed to counting even when losing). */
1242                                 /* Please note that this preset might be
1243                                  * somewhat weaker than normal Pachi, and the
1244                                  * score maximization is approximate; point size
1245                                  * of win/loss still should not be used to judge
1246                                  * strength of Pachi or the opponent. */
1247                                 /* See README for some further notes. */
1248                                 if (!optval || atoi(optval)) {
1249                                         /* Allow scoring a lost game. */
1250                                         u->allow_losing_pass = true;
1251                                         /* Make Pachi keep his calm when losing
1252                                          * and/or maintain winning marging. */
1253                                         /* Do not play games that are losing
1254                                          * by too much. */
1255                                         /* XXX: komi_ratchet_age=40000 is necessary
1256                                          * with losing_komi_ratchet, but 40000
1257                                          * is somewhat arbitrary value. */
1258                                         char dynkomi_args[] = "losing_komi_ratchet:komi_ratchet_age=60000:no_komi_at_game_end=0:max_losing_komi=30";
1259                                         u->dynkomi = uct_dynkomi_init_adaptive(u, dynkomi_args, b);
1260                                         /* XXX: Values arbitrary so far. */
1261                                         /* XXX: Also, is bytemp sensible when
1262                                          * combined with dynamic komi?! */
1263                                         u->val_scale = 0.01;
1264                                         u->val_bytemp = true;
1265                                         u->val_bytemp_min = 0.001;
1266                                         u->val_byavg = true;
1267                                 }
1268
1269                         } else {
1270                                 fprintf(stderr, "uct: Invalid engine argument %s or missing value\n", optname);
1271                                 exit(1);
1272                         }
1273                 }
1274         }
1275
1276         if (!u->policy)
1277                 u->policy = policy_ucb1amaf_init(u, NULL, b);
1278
1279         if (!!u->random_policy_chance ^ !!u->random_policy) {
1280                 fprintf(stderr, "uct: Only one of random_policy and random_policy_chance is set\n");
1281                 exit(1);
1282         }
1283
1284         if (!u->local_tree) {
1285                 /* No ltree aging. */
1286                 u->local_tree_aging = 1.0f;
1287         }
1288
1289         if (u->fast_alloc) {
1290                 if (u->pruning_threshold < u->max_tree_size / 10)
1291                         u->pruning_threshold = u->max_tree_size / 10;
1292                 if (u->pruning_threshold > u->max_tree_size / 2)
1293                         u->pruning_threshold = u->max_tree_size / 2;
1294
1295                 /* Limit pruning temp space to 20% of memory. Beyond this we discard
1296                  * the nodes and recompute them at the next move if necessary. */
1297                 u->max_pruned_size = u->max_tree_size / 5;
1298                 u->max_tree_size -= u->max_pruned_size;
1299         } else {
1300                 /* Reserve 5% memory in case the background free() are slower
1301                  * than the concurrent allocations. */
1302                 u->max_tree_size -= u->max_tree_size / 20;
1303         }
1304
1305         if (!u->prior)
1306                 u->prior = uct_prior_init(NULL, b, u);
1307
1308         if (!u->playout)
1309                 u->playout = playout_moggy_init(NULL, b, u->jdict);
1310         if (!u->playout->debug_level)
1311                 u->playout->debug_level = u->debug_level;
1312
1313         if (u->want_pat && !pat_setup)
1314                 patterns_init(&u->pat, NULL, false, true);
1315         dcnn_init();
1316
1317         u->ownermap.map = malloc2(board_size2(b) * sizeof(u->ownermap.map[0]));
1318
1319         if (u->slave) {
1320                 if (!u->stats_hbits) u->stats_hbits = DEFAULT_STATS_HBITS;
1321                 if (!u->shared_nodes) u->shared_nodes = DEFAULT_SHARED_NODES;
1322                 assert(u->shared_levels * board_bits2(b) <= 8 * (int)sizeof(path_t));
1323         }
1324
1325         if (!u->dynkomi)
1326                 u->dynkomi = board_small(b) ? uct_dynkomi_init_none(u, NULL, b)
1327                         : uct_dynkomi_init_linear(u, NULL, b);
1328
1329         /* Some things remain uninitialized for now - the opening tbook
1330          * is not loaded and the tree not set up. */
1331         /* This will be initialized in setup_state() at the first move
1332          * received/requested. This is because right now we are not aware
1333          * about any komi or handicap setup and such. */
1334
1335         return u;
1336 }
1337
1338 struct engine *
1339 engine_uct_init(char *arg, struct board *b)
1340 {
1341         struct uct *u = uct_state_init(arg, b);
1342         struct engine *e = calloc2(1, sizeof(struct engine));
1343         e->name = "UCT";
1344         e->board_print = uct_board_print;
1345         e->notify_play = uct_notify_play;
1346         e->chat = uct_chat;
1347         e->undo = uct_undo;
1348         e->result = uct_result;
1349         e->genmove = uct_genmove;
1350         e->genmoves = uct_genmoves;
1351         e->evaluate = uct_evaluate;
1352         e->dead_group_list = uct_dead_group_list;
1353         e->stop = uct_stop;
1354         e->done = uct_done;
1355         e->owner_map = uct_owner_map;
1356         e->best_moves = uct_best_moves;
1357         e->live_gfx_hook = uct_live_gfx_hook;
1358         e->data = u;
1359         if (u->slave)
1360                 e->notify = uct_notify;
1361
1362         const char banner[] = "If you believe you have won but I am still playing, "
1363                 "please help me understand by capturing all dead stones. "
1364                 "Anyone can send me 'winrate' in private chat to get my assessment of the position.";
1365         if (!u->banner) u->banner = "";
1366         e->comment = malloc2(sizeof(banner) + strlen(u->banner) + 1);
1367         sprintf(e->comment, "%s %s", banner, u->banner);
1368
1369         return e;
1370 }