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 "move.h"
  14 #include "mq.h"
  15 #include "playout.h"
  16 #include "playout/elo.h"
  17 #include "playout/moggy.h"
  18 #include "playout/light.h"
  19 #include "tactics.h"
  20 #include "timeinfo.h"
  21 #include "distributed/distributed.h"
  22 #include "uct/dynkomi.h"
  23 #include "uct/internal.h"
  24 #include "uct/prior.h"
  25 #include "uct/search.h"
  26 #include "uct/slave.h"
  27 #include "uct/tree.h"
  28 #include "uct/uct.h"
  29 #include "uct/walk.h"
  30
  31 struct uct_policy *policy_ucb1_init(struct uct *u, char *arg);
  32 struct uct_policy *policy_ucb1amaf_init(struct uct *u, char *arg);
  33 static void uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color);
  34
  35 /* Maximal simulation length. */
  36 #define MC_GAMELEN      MAX_GAMELEN
  37
  38
  39 static void
  40 setup_state(struct uct *u, struct board *b, enum stone color)
  41 {
  42         u->t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0, u->local_tree_aging);
  43         if (u->force_seed)
  44                 fast_srandom(u->force_seed);
  45         if (UDEBUGL(0))
  46                 fprintf(stderr, "Fresh board with random seed %lu\n", fast_getseed());
  47         //board_print(b, stderr);
  48         if (!u->no_book && b->moves == 0) {
  49                 assert(color == S_BLACK);
  50                 tree_load(u->t, b);
  51         }
  52 }
  53
  54 static void
  55 reset_state(struct uct *u)
  56 {
  57         assert(u->t);
  58         tree_done(u->t); u->t = NULL;
  59 }
  60
  61 static void
  62 setup_dynkomi(struct uct *u, struct board *b, enum stone to_play)
  63 {
  64         if (u->t->use_extra_komi && u->dynkomi->permove)
  65                 u->t->extra_komi = u->dynkomi->permove(u->dynkomi, b, u->t);
  66 }
  67
  68 void
  69 uct_prepare_move(struct uct *u, struct board *b, enum stone color)
  70 {
  71         if (u->t) {
  72                 /* Verify that we have sane state. */
  73                 assert(b->es == u);
  74                 assert(u->t && b->moves);
  75                 if (color != stone_other(u->t->root_color)) {
  76                         fprintf(stderr, "Fatal: Non-alternating play detected %d %d\n",
  77                                 color, u->t->root_color);
  78                         exit(1);
  79                 }
  80
  81         } else {
  82                 /* We need fresh state. */
  83                 b->es = u;
  84                 setup_state(u, b, color);
  85         }
  86
  87         u->ownermap.playouts = 0;
  88         memset(u->ownermap.map, 0, board_size2(b) * sizeof(u->ownermap.map[0]));
  89         memset(u->stats, 0, board_size2(b) * sizeof(u->stats[0]));
  90         u->played_own = u->played_all = 0;
  91 }
  92
  93 static void
  94 dead_group_list(struct uct *u, struct board *b, struct move_queue *mq)
  95 {
  96         struct group_judgement gj;
  97         gj.thres = GJ_THRES;
  98         gj.gs = alloca(board_size2(b) * sizeof(gj.gs[0]));
  99         board_ownermap_judge_group(b, &u->ownermap, &gj);
 100         groups_of_status(b, &gj, GS_DEAD, mq);
 101 }
 102
 103 bool
 104 uct_pass_is_safe(struct uct *u, struct board *b, enum stone color, bool pass_all_alive)
 105 {
 106         if (u->ownermap.playouts < GJ_MINGAMES)
 107                 return false;
 108
 109         struct move_queue mq = { .moves = 0 };
 110         dead_group_list(u, b, &mq);
 111         if (pass_all_alive && mq.moves > 0)
 112                 return false; // We need to remove some dead groups first.
 113         return pass_is_safe(b, color, &mq);
 114 }
 115
 116 static char *
 117 uct_printhook_ownermap(struct board *board, coord_t c, char *s, char *end)
 118 {
 119         struct uct *u = board->es;
 120         assert(u);
 121         const char chr[] = ":XO,"; // dame, black, white, unclear
 122         const char chm[] = ":xo,";
 123         char ch = chr[board_ownermap_judge_point(&u->ownermap, c, GJ_THRES)];
 124         if (ch == ',') { // less precise estimate then?
 125                 ch = chm[board_ownermap_judge_point(&u->ownermap, c, 0.67)];
 126         }
 127         s += snprintf(s, end - s, "%c ", ch);
 128         return s;
 129 }
 130
 131 static char *
 132 uct_notify_play(struct engine *e, struct board *b, struct move *m)
 133 {
 134         struct uct *u = e->data;
 135         if (!u->t) {
 136                 /* No state, create one - this is probably game beginning
 137                  * and we need to load the opening book right now. */
 138                 uct_prepare_move(u, b, m->color);
 139                 assert(u->t);
 140         }
 141
 142         /* Stop pondering, required by tree_promote_at() */
 143         uct_pondering_stop(u);
 144         if (UDEBUGL(2) && u->slave)
 145                 tree_dump(u->t, u->dumpthres);
 146
 147         if (is_resign(m->coord)) {
 148                 /* Reset state. */
 149                 reset_state(u);
 150                 return NULL;
 151         }
 152
 153         /* Promote node of the appropriate move to the tree root. */
 154         assert(u->t->root);
 155         if (!tree_promote_at(u->t, b, m->coord)) {
 156                 if (UDEBUGL(0))
 157                         fprintf(stderr, "Warning: Cannot promote move node! Several play commands in row?\n");
 158                 reset_state(u);
 159                 return NULL;
 160         }
 161
 162         /* If we are a slave in a distributed engine, start pondering once
 163          * we know which move we actually played. See uct_genmove() about
 164          * the check for pass. */
 165         if (u->pondering_opt && u->slave && m->color == u->my_color && !is_pass(m->coord))
 166                 uct_pondering_start(u, b, u->t, stone_other(m->color));
 167
 168         return NULL;
 169 }
 170
 171 static char *
 172 uct_chat(struct engine *e, struct board *b, char *cmd)
 173 {
 174         struct uct *u = e->data;
 175         static char reply[1024];
 176
 177         cmd += strspn(cmd, " \n\t");
 178         if (!strncasecmp(cmd, "winrate", 7)) {
 179                 if (!u->t)
 180                         return "no game context (yet?)";
 181                 enum stone color = u->t->root_color;
 182                 struct tree_node *n = u->t->root;
 183                 snprintf(reply, 1024, "In %d playouts at %d threads, %s %s can win with %.2f%% probability",
 184                          n->u.playouts, u->threads, stone2str(color), coord2sstr(n->coord, b),
 185                          tree_node_get_value(u->t, -1, n->u.value) * 100);
 186                 if (u->t->use_extra_komi && abs(u->t->extra_komi) >= 0.5) {
 187                         sprintf(reply + strlen(reply), ", while self-imposing extra komi %.1f",
 188                                 u->t->extra_komi);
 189                 }
 190                 strcat(reply, ".");
 191                 return reply;
 192         }
 193         return NULL;
 194 }
 195
 196 static void
 197 uct_dead_group_list(struct engine *e, struct board *b, struct move_queue *mq)
 198 {
 199         struct uct *u = e->data;
 200
 201         /* This means the game is probably over, no use pondering on. */
 202         uct_pondering_stop(u);
 203
 204         if (u->pass_all_alive)
 205                 return; // no dead groups
 206
 207         bool mock_state = false;
 208
 209         if (!u->t) {
 210                 /* No state, but we cannot just back out - we might
 211                  * have passed earlier, only assuming some stones are
 212                  * dead, and then re-connected, only to lose counting
 213                  * when all stones are assumed alive. */
 214                 /* Mock up some state and seed the ownermap by few
 215                  * simulations. */
 216                 uct_prepare_move(u, b, S_BLACK); assert(u->t);
 217                 for (int i = 0; i < GJ_MINGAMES; i++)
 218                         uct_playout(u, b, S_BLACK, u->t);
 219                 mock_state = true;
 220         }
 221
 222         dead_group_list(u, b, mq);
 223
 224         if (mock_state) {
 225                 /* Clean up the mock state in case we will receive
 226                  * a genmove; we could get a non-alternating-move
 227                  * error from uct_prepare_move() in that case otherwise. */
 228                 reset_state(u);
 229         }
 230 }
 231
 232 static void
 233 playout_policy_done(struct playout_policy *p)
 234 {
 235         if (p->done) p->done(p);
 236         if (p->data) free(p->data);
 237         free(p);
 238 }
 239
 240 static void
 241 uct_done(struct engine *e)
 242 {
 243         /* This is called on engine reset, especially when clear_board
 244          * is received and new game should begin. */
 245         struct uct *u = e->data;
 246         uct_pondering_stop(u);
 247         if (u->t) reset_state(u);
 248         free(u->ownermap.map);
 249         free(u->stats);
 250
 251         free(u->policy);
 252         free(u->random_policy);
 253         playout_policy_done(u->playout);
 254         uct_prior_done(u->prior);
 255 }
 256
 257
 258
 259 /* Run time-limited MCTS search on foreground. */
 260 static int
 261 uct_search(struct uct *u, struct board *b, struct time_info *ti, enum stone color, struct tree *t)
 262 {
 263         struct uct_search_state s;
 264         uct_search_start(u, b, color, t, ti, &s);
 265         if (UDEBUGL(2) && s.base_playouts > 0)
 266                 fprintf(stderr, "<pre-simulated %d games>\n", s.base_playouts);
 267
 268         /* The search tree is ctx->t. This is normally == t, but in case of
 269          * TM_ROOT, it is one of the trees belonging to the independent
 270          * workers. It is important to reference ctx->t directly since the
 271          * thread manager will swap the tree pointer asynchronously. */
 272         /* XXX: This means TM_ROOT support is suboptimal since single stalled
 273          * thread can stall the others in case of limiting the search by game
 274          * count. However, TM_ROOT just does not deserve any more extra code
 275          * right now. */
 276
 277         /* Now, just periodically poll the search tree. */
 278         while (1) {
 279                 time_sleep(TREE_BUSYWAIT_INTERVAL);
 280                 /* TREE_BUSYWAIT_INTERVAL should never be less than desired time, or the
 281                  * time control is broken. But if it happens to be less, we still search
 282                  * at least 100ms otherwise the move is completely random. */
 283
 284                 int i = uct_search_games(&s);
 285                 /* Print notifications etc. */
 286                 uct_search_progress(u, b, color, t, ti, &s, i);
 287                 /* Check if we should stop the search. */
 288                 if (uct_search_check_stop(u, b, color, t, ti, &s, i))
 289                         break;
 290         }
 291
 292         struct uct_thread_ctx *ctx = uct_search_stop();
 293         if (UDEBUGL(2)) tree_dump(t, u->dumpthres);
 294         if (UDEBUGL(2))
 295                 fprintf(stderr, "(avg score %f/%d value %f/%d)\n",
 296                         u->dynkomi->score.value, u->dynkomi->score.playouts,
 297                         u->dynkomi->value.value, u->dynkomi->value.playouts);
 298         if (UDEBUGL(0))
 299                 uct_progress_status(u, t, color, ctx->games);
 300
 301         u->played_own += ctx->games;
 302         return ctx->games;
 303 }
 304
 305 /* Start pondering background with @color to play. */
 306 static void
 307 uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color)
 308 {
 309         if (UDEBUGL(1))
 310                 fprintf(stderr, "Starting to ponder with color %s\n", stone2str(stone_other(color)));
 311         u->pondering = true;
 312
 313         /* We need a local board copy to ponder upon. */
 314         struct board *b = malloc2(sizeof(*b)); board_copy(b, b0);
 315
 316         /* *b0 did not have the genmove'd move played yet. */
 317         struct move m = { t->root->coord, t->root_color };
 318         int res = board_play(b, &m);
 319         assert(res >= 0);
 320         setup_dynkomi(u, b, stone_other(m.color));
 321
 322         /* Start MCTS manager thread "headless". */
 323         static struct uct_search_state s;
 324         uct_search_start(u, b, color, t, NULL, &s);
 325 }
 326
 327 /* uct_search_stop() frontend for the pondering (non-genmove) mode, and
 328  * to stop the background search for a slave in the distributed engine. */
 329 void
 330 uct_pondering_stop(struct uct *u)
 331 {
 332         if (!thread_manager_running)
 333                 return;
 334
 335         /* Stop the thread manager. */
 336         struct uct_thread_ctx *ctx = uct_search_stop();
 337         if (UDEBUGL(1)) {
 338                 if (u->pondering) fprintf(stderr, "(pondering) ");
 339                 uct_progress_status(u, ctx->t, ctx->color, ctx->games);
 340         }
 341         if (u->pondering) {
 342                 free(ctx->b);
 343                 u->pondering = false;
 344         }
 345 }
 346
 347
 348 void
 349 uct_genmove_setup(struct uct *u, struct board *b, enum stone color)
 350 {
 351         if (b->superko_violation) {
 352                 fprintf(stderr, "!!! WARNING: SUPERKO VIOLATION OCCURED BEFORE THIS MOVE\n");
 353                 fprintf(stderr, "Maybe you play with situational instead of positional superko?\n");
 354                 fprintf(stderr, "I'm going to ignore the violation, but note that I may miss\n");
 355                 fprintf(stderr, "some moves valid under this ruleset because of this.\n");
 356                 b->superko_violation = false;
 357         }
 358
 359         uct_prepare_move(u, b, color);
 360
 361         assert(u->t);
 362         u->my_color = color;
 363
 364         /* How to decide whether to use dynkomi in this game? Since we use
 365          * pondering, it's not simple "who-to-play" matter. Decide based on
 366          * the last genmove issued. */
 367         u->t->use_extra_komi = !!(u->dynkomi_mask & color);
 368         setup_dynkomi(u, b, color);
 369
 370         if (b->rules == RULES_JAPANESE)
 371                 u->territory_scoring = true;
 372
 373         /* Make pessimistic assumption about komi for Japanese rules to
 374          * avoid losing by 0.5 when winning by 0.5 with Chinese rules.
 375          * The rules usually give the same winner if the integer part of komi
 376          * is odd so we adjust the komi only if it is even (for a board of
 377          * odd size). We are not trying  to get an exact evaluation for rare
 378          * cases of seki. For details see http://home.snafu.de/jasiek/parity.html */
 379         if (u->territory_scoring && (((int)floor(b->komi) + board_size(b)) & 1)) {
 380                 b->komi += (color == S_BLACK ? 1.0 : -1.0);
 381                 if (UDEBUGL(0))
 382                         fprintf(stderr, "Setting komi to %.1f assuming Japanese rules\n",
 383                                 b->komi);
 384         }
 385 }
 386
 387 static coord_t *
 388 uct_genmove(struct engine *e, struct board *b, struct time_info *ti, enum stone color, bool pass_all_alive)
 389 {
 390         double start_time = time_now();
 391         struct uct *u = e->data;
 392         uct_pondering_stop(u);
 393         uct_genmove_setup(u, b, color);
 394
 395         /* Start the Monte Carlo Tree Search! */
 396         int base_playouts = u->t->root->u.playouts;
 397         int played_games = uct_search(u, b, ti, color, u->t);
 398
 399         coord_t best_coord;
 400         struct tree_node *best;
 401         best = uct_search_result(u, b, color, pass_all_alive, played_games, base_playouts, &best_coord);
 402
 403         if (UDEBUGL(2)) {
 404                 double time = time_now() - start_time + 0.000001; /* avoid divide by zero */
 405                 fprintf(stderr, "genmove in %0.2fs (%d games/s, %d games/s/thread)\n",
 406                         time, (int)(played_games/time), (int)(played_games/time/u->threads));
 407         }
 408
 409         if (!best) {
 410                 /* Pass or resign. */
 411                 reset_state(u);
 412                 return coord_copy(best_coord);
 413         }
 414         tree_promote_node(u->t, &best);
 415
 416         /* After a pass, pondering is harmful for two reasons:
 417          * (i) We might keep pondering even when the game is over.
 418          * Of course this is the case for opponent resign as well.
 419          * (ii) More importantly, the ownermap will get skewed since
 420          * the UCT will start cutting off any playouts. */
 421         if (u->pondering_opt && !is_pass(best->coord)) {
 422                 uct_pondering_start(u, b, u->t, stone_other(color));
 423         }
 424         return coord_copy(best_coord);
 425 }
 426
 427
 428 bool
 429 uct_genbook(struct engine *e, struct board *b, struct time_info *ti, enum stone color)
 430 {
 431         struct uct *u = e->data;
 432         if (!u->t) uct_prepare_move(u, b, color);
 433         assert(u->t);
 434
 435         if (ti->dim == TD_GAMES) {
 436                 /* Don't count in games that already went into the book. */
 437                 ti->len.games += u->t->root->u.playouts;
 438         }
 439         uct_search(u, b, ti, color, u->t);
 440
 441         assert(ti->dim == TD_GAMES);
 442         tree_save(u->t, b, ti->len.games / 100);
 443
 444         return true;
 445 }
 446
 447 void
 448 uct_dumpbook(struct engine *e, struct board *b, enum stone color)
 449 {
 450         struct uct *u = e->data;
 451         struct tree *t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0, u->local_tree_aging);
 452         tree_load(t, b);
 453         tree_dump(t, 0);
 454         tree_done(t);
 455 }
 456
 457
 458 struct uct *
 459 uct_state_init(char *arg, struct board *b)
 460 {
 461         struct uct *u = calloc2(1, sizeof(struct uct));
 462         bool using_elo = false;
 463
 464         u->debug_level = debug_level;
 465         u->gamelen = MC_GAMELEN;
 466         u->mercymin = 0;
 467         u->expand_p = 2;
 468         u->dumpthres = 1000;
 469         u->playout_amaf = true;
 470         u->playout_amaf_nakade = false;
 471         u->amaf_prior = false;
 472         u->max_tree_size = 3072ULL * 1048576;
 473
 474         u->dynkomi_mask = S_BLACK;
 475
 476         u->threads = 1;
 477         u->thread_model = TM_TREEVL;
 478         u->parallel_tree = true;
 479         u->virtual_loss = true;
 480
 481         u->fuseki_end = 20; // max time at 361*20% = 72 moves (our 36th move, still 99 to play)
 482         u->yose_start = 40; // (100-40-25)*361/100/2 = 63 moves still to play by us then
 483         u->bestr_ratio = 0.02;
 484         // 2.5 is clearly too much, but seems to compensate well for overly stern time allocations.
 485         // TODO: Further tuning and experiments with better time allocation schemes.
 486         u->best2_ratio = 2.5;
 487
 488         u->val_scale = 0.04; u->val_points = 40;
 489
 490         u->tenuki_d = 4;
 491         u->local_tree_aging = 2;
 492
 493         if (arg) {
 494                 char *optspec, *next = arg;
 495                 while (*next) {
 496                         optspec = next;
 497                         next += strcspn(next, ",");
 498                         if (*next) { *next++ = 0; } else { *next = 0; }
 499
 500                         char *optname = optspec;
 501                         char *optval = strchr(optspec, '=');
 502                         if (optval) *optval++ = 0;
 503
 504                         if (!strcasecmp(optname, "debug")) {
 505                                 if (optval)
 506                                         u->debug_level = atoi(optval);
 507                                 else
 508                                         u->debug_level++;
 509                         } else if (!strcasecmp(optname, "mercy") && optval) {
 510                                 /* Minimal difference of black/white captures
 511                                  * to stop playout - "Mercy Rule". Speeds up
 512                                  * hopeless playouts at the expense of some
 513                                  * accuracy. */
 514                                 u->mercymin = atoi(optval);
 515                         } else if (!strcasecmp(optname, "gamelen") && optval) {
 516                                 u->gamelen = atoi(optval);
 517                         } else if (!strcasecmp(optname, "expand_p") && optval) {
 518                                 u->expand_p = atoi(optval);
 519                         } else if (!strcasecmp(optname, "dumpthres") && optval) {
 520                                 u->dumpthres = atoi(optval);
 521                         } else if (!strcasecmp(optname, "best2_ratio") && optval) {
 522                                 /* If set, prolong simulating while
 523                                  * first_best/second_best playouts ratio
 524                                  * is less than best2_ratio. */
 525                                 u->best2_ratio = atof(optval);
 526                         } else if (!strcasecmp(optname, "bestr_ratio") && optval) {
 527                                 /* If set, prolong simulating while
 528                                  * best,best_best_child values delta
 529                                  * is more than bestr_ratio. */
 530                                 u->bestr_ratio = atof(optval);
 531                         } else if (!strcasecmp(optname, "playout_amaf")) {
 532                                 /* Whether to include random playout moves in
 533                                  * AMAF as well. (Otherwise, only tree moves
 534                                  * are included in AMAF. Of course makes sense
 535                                  * only in connection with an AMAF policy.) */
 536                                 /* with-without: 55.5% (+-4.1) */
 537                                 if (optval && *optval == '0')
 538                                         u->playout_amaf = false;
 539                                 else
 540                                         u->playout_amaf = true;
 541                         } else if (!strcasecmp(optname, "playout_amaf_nakade")) {
 542                                 /* Whether to include nakade moves from playouts
 543                                  * in the AMAF statistics; this tends to nullify
 544                                  * the playout_amaf effect by adding too much
 545                                  * noise. */
 546                                 if (optval && *optval == '0')
 547                                         u->playout_amaf_nakade = false;
 548                                 else
 549                                         u->playout_amaf_nakade = true;
 550                         } else if (!strcasecmp(optname, "playout_amaf_cutoff") && optval) {
 551                                 /* Keep only first N% of playout stage AMAF
 552                                  * information. */
 553                                 u->playout_amaf_cutoff = atoi(optval);
 554                         } else if ((!strcasecmp(optname, "policy") || !strcasecmp(optname, "random_policy")) && optval) {
 555                                 char *policyarg = strchr(optval, ':');
 556                                 struct uct_policy **p = !strcasecmp(optname, "policy") ? &u->policy : &u->random_policy;
 557                                 if (policyarg)
 558                                         *policyarg++ = 0;
 559                                 if (!strcasecmp(optval, "ucb1")) {
 560                                         *p = policy_ucb1_init(u, policyarg);
 561                                 } else if (!strcasecmp(optval, "ucb1amaf")) {
 562                                         *p = policy_ucb1amaf_init(u, policyarg);
 563                                 } else {
 564                                         fprintf(stderr, "UCT: Invalid tree policy %s\n", optval);
 565                                         exit(1);
 566                                 }
 567                         } else if (!strcasecmp(optname, "playout") && optval) {
 568                                 char *playoutarg = strchr(optval, ':');
 569                                 if (playoutarg)
 570                                         *playoutarg++ = 0;
 571                                 if (!strcasecmp(optval, "moggy")) {
 572                                         u->playout = playout_moggy_init(playoutarg, b);
 573                                 } else if (!strcasecmp(optval, "light")) {
 574                                         u->playout = playout_light_init(playoutarg, b);
 575                                 } else if (!strcasecmp(optval, "elo")) {
 576                                         u->playout = playout_elo_init(playoutarg, b);
 577                                         using_elo = true;
 578                                 } else {
 579                                         fprintf(stderr, "UCT: Invalid playout policy %s\n", optval);
 580                                         exit(1);
 581                                 }
 582                         } else if (!strcasecmp(optname, "prior") && optval) {
 583                                 u->prior = uct_prior_init(optval, b);
 584                         } else if (!strcasecmp(optname, "amaf_prior") && optval) {
 585                                 u->amaf_prior = atoi(optval);
 586                         } else if (!strcasecmp(optname, "threads") && optval) {
 587                                 /* By default, Pachi will run with only single
 588                                  * tree search thread! */
 589                                 u->threads = atoi(optval);
 590                         } else if (!strcasecmp(optname, "thread_model") && optval) {
 591                                 if (!strcasecmp(optval, "root")) {
 592                                         /* Root parallelization - each thread
 593                                          * does independent search, trees are
 594                                          * merged at the end. */
 595                                         u->thread_model = TM_ROOT;
 596                                         u->parallel_tree = false;
 597                                         u->virtual_loss = false;
 598                                 } else if (!strcasecmp(optval, "tree")) {
 599                                         /* Tree parallelization - all threads
 600                                          * grind on the same tree. */
 601                                         u->thread_model = TM_TREE;
 602                                         u->parallel_tree = true;
 603                                         u->virtual_loss = false;
 604                                 } else if (!strcasecmp(optval, "treevl")) {
 605                                         /* Tree parallelization, but also
 606                                          * with virtual losses - this discou-
 607                                          * rages most threads choosing the
 608                                          * same tree branches to read. */
 609                                         u->thread_model = TM_TREEVL;
 610                                         u->parallel_tree = true;
 611                                         u->virtual_loss = true;
 612                                 } else {
 613                                         fprintf(stderr, "UCT: Invalid thread model %s\n", optval);
 614                                         exit(1);
 615                                 }
 616                         } else if (!strcasecmp(optname, "pondering")) {
 617                                 /* Keep searching even during opponent's turn. */
 618                                 u->pondering_opt = !optval || atoi(optval);
 619                         } else if (!strcasecmp(optname, "fuseki_end") && optval) {
 620                                 /* At the very beginning it's not worth thinking
 621                                  * too long because the playout evaluations are
 622                                  * very noisy. So gradually increase the thinking
 623                                  * time up to maximum when fuseki_end percent
 624                                  * of the board has been played.
 625                                  * This only applies if we are not in byoyomi. */
 626                                 u->fuseki_end = atoi(optval);
 627                         } else if (!strcasecmp(optname, "yose_start") && optval) {
 628                                 /* When yose_start percent of the board has been
 629                                  * played, or if we are in byoyomi, stop spending
 630                                  * more time and spread the remaining time
 631                                  * uniformly.
 632                                  * Between fuseki_end and yose_start, we spend
 633                                  * a constant proportion of the remaining time
 634                                  * on each move. (yose_start should actually
 635                                  * be much earlier than when real yose start,
 636                                  * but "yose" is a good short name to convey
 637                                  * the idea.) */
 638                                 u->yose_start = atoi(optval);
 639                         } else if (!strcasecmp(optname, "force_seed") && optval) {
 640                                 u->force_seed = atoi(optval);
 641                         } else if (!strcasecmp(optname, "no_book")) {
 642                                 u->no_book = true;
 643                         } else if (!strcasecmp(optname, "dynkomi") && optval) {
 644                                 /* Dynamic komi approach; there are multiple
 645                                  * ways to adjust komi dynamically throughout
 646                                  * play. We currently support two: */
 647                                 char *dynkomiarg = strchr(optval, ':');
 648                                 if (dynkomiarg)
 649                                         *dynkomiarg++ = 0;
 650                                 if (!strcasecmp(optval, "none")) {
 651                                         u->dynkomi = uct_dynkomi_init_none(u, dynkomiarg, b);
 652                                 } else if (!strcasecmp(optval, "linear")) {
 653                                         u->dynkomi = uct_dynkomi_init_linear(u, dynkomiarg, b);
 654                                 } else if (!strcasecmp(optval, "adaptive")) {
 655                                         u->dynkomi = uct_dynkomi_init_adaptive(u, dynkomiarg, b);
 656                                 } else {
 657                                         fprintf(stderr, "UCT: Invalid dynkomi mode %s\n", optval);
 658                                         exit(1);
 659                                 }
 660                         } else if (!strcasecmp(optname, "dynkomi_mask") && optval) {
 661                                 /* Bitmask of colors the player must be
 662                                  * for dynkomi be applied; you may want
 663                                  * to use dynkomi_mask=3 to allow dynkomi
 664                                  * even in games where Pachi is white. */
 665                                 u->dynkomi_mask = atoi(optval);
 666                         } else if (!strcasecmp(optname, "dynkomi_interval") && optval) {
 667                                 /* If non-zero, re-adjust dynamic komi
 668                                  * throughout a single genmove reading,
 669                                  * roughly every N simulations. */
 670                                 /* XXX: Does not work with tree
 671                                  * parallelization. */
 672                                 u->dynkomi_interval = atoi(optval);
 673                         } else if (!strcasecmp(optname, "val_scale") && optval) {
 674                                 /* How much of the game result value should be
 675                                  * influenced by win size. Zero means it isn't. */
 676                                 u->val_scale = atof(optval);
 677                         } else if (!strcasecmp(optname, "val_points") && optval) {
 678                                 /* Maximum size of win to be scaled into game
 679                                  * result value. Zero means boardsize^2. */
 680                                 u->val_points = atoi(optval) * 2; // result values are doubled
 681                         } else if (!strcasecmp(optname, "val_extra")) {
 682                                 /* If false, the score coefficient will be simply
 683                                  * added to the value, instead of scaling the result
 684                                  * coefficient because of it. */
 685                                 u->val_extra = !optval || atoi(optval);
 686                         } else if (!strcasecmp(optname, "local_tree") && optval) {
 687                                 /* Whether to bias exploration by local tree values
 688                                  * (must be supported by the used policy).
 689                                  * 0: Don't.
 690                                  * 1: Do, value = result.
 691                                  * Try to temper the result:
 692                                  * 2: Do, value = 0.5+(result-expected)/2.
 693                                  * 3: Do, value = 0.5+bzz((result-expected)^2).
 694                                  * 4: Do, value = 0.5+sqrt(result-expected)/2. */
 695                                 u->local_tree = atoi(optval);
 696                         } else if (!strcasecmp(optname, "tenuki_d") && optval) {
 697                                 /* Tenuki distance at which to break the local tree. */
 698                                 u->tenuki_d = atoi(optval);
 699                                 if (u->tenuki_d > TREE_NODE_D_MAX + 1) {
 700                                         fprintf(stderr, "uct: tenuki_d must not be larger than TREE_NODE_D_MAX+1 %d\n", TREE_NODE_D_MAX + 1);
 701                                         exit(1);
 702                                 }
 703                         } else if (!strcasecmp(optname, "local_tree_aging") && optval) {
 704                                 /* How much to reduce local tree values between moves. */
 705                                 u->local_tree_aging = atof(optval);
 706                         } else if (!strcasecmp(optname, "local_tree_allseq")) {
 707                                 /* By default, only complete sequences are stored
 708                                  * in the local tree. If this is on, also
 709                                  * subsequences starting at each move are stored. */
 710                                 u->local_tree_allseq = !optval || atoi(optval);
 711                         } else if (!strcasecmp(optname, "local_tree_playout")) {
 712                                 /* Whether to adjust ELO playout probability
 713                                  * distributions according to matched localtree
 714                                  * information. */
 715                                 u->local_tree_playout = !optval || atoi(optval);
 716                         } else if (!strcasecmp(optname, "local_tree_pseqroot")) {
 717                                 /* By default, when we have no sequence move
 718                                  * to suggest in-playout, we give up. If this
 719                                  * is on, we make probability distribution from
 720                                  * sequences first moves instead. */
 721                                 u->local_tree_pseqroot = !optval || atoi(optval);
 722                         } else if (!strcasecmp(optname, "pass_all_alive")) {
 723                                 /* Whether to consider passing only after all
 724                                  * dead groups were removed from the board;
 725                                  * this is like all genmoves are in fact
 726                                  * kgs-genmove_cleanup. */
 727                                 u->pass_all_alive = !optval || atoi(optval);
 728                         } else if (!strcasecmp(optname, "territory_scoring")) {
 729                                 /* Use territory scoring (default is area scoring).
 730                                  * An explicit kgs-rules command overrides this. */
 731                                 u->territory_scoring = !optval || atoi(optval);
 732                         } else if (!strcasecmp(optname, "random_policy_chance") && optval) {
 733                                 /* If specified (N), with probability 1/N, random_policy policy
 734                                  * descend is used instead of main policy descend; useful
 735                                  * if specified policy (e.g. UCB1AMAF) can make unduly biased
 736                                  * choices sometimes, you can fall back to e.g.
 737                                  * random_policy=UCB1. */
 738                                 u->random_policy_chance = atoi(optval);
 739                         } else if (!strcasecmp(optname, "max_tree_size") && optval) {
 740                                 /* Maximum amount of memory [MiB] consumed by the move tree.
 741                                  * For fast_alloc it includes the temp tree used for pruning.
 742                                  * Default is 3072 (3 GiB). Note that if you use TM_ROOT,
 743                                  * this limits size of only one of the trees, not all of them
 744                                  * together. */
 745                                 u->max_tree_size = atol(optval) * 1048576;
 746                         } else if (!strcasecmp(optname, "fast_alloc")) {
 747                                 u->fast_alloc = !optval || atoi(optval);
 748                         } else if (!strcasecmp(optname, "slave")) {
 749                                 /* Act as slave for the distributed engine. */
 750                                 u->slave = !optval || atoi(optval);
 751                         } else if (!strcasecmp(optname, "banner") && optval) {
 752                                 /* Additional banner string. This must come as the
 753                                  * last engine parameter. */
 754                                 if (*next) *--next = ',';
 755                                 u->banner = strdup(optval);
 756                                 break;
 757                         } else {
 758                                 fprintf(stderr, "uct: Invalid engine argument %s or missing value\n", optname);
 759                                 exit(1);
 760                         }
 761                 }
 762         }
 763
 764         u->resign_ratio = 0.2; /* Resign when most games are lost. */
 765         u->loss_threshold = 0.85; /* Stop reading if after at least 2000 playouts this is best value. */
 766         if (!u->policy)
 767                 u->policy = policy_ucb1amaf_init(u, NULL);
 768
 769         if (!!u->random_policy_chance ^ !!u->random_policy) {
 770                 fprintf(stderr, "uct: Only one of random_policy and random_policy_chance is set\n");
 771                 exit(1);
 772         }
 773
 774         if (!u->local_tree) {
 775                 /* No ltree aging. */
 776                 u->local_tree_aging = 1.0f;
 777         }
 778         if (!using_elo)
 779                 u->local_tree_playout = false;
 780
 781         if (u->fast_alloc && !u->parallel_tree) {
 782                 fprintf(stderr, "fast_alloc not supported with root parallelization.\n");
 783                 exit(1);
 784         }
 785         if (u->fast_alloc)
 786                 u->max_tree_size = (100ULL * u->max_tree_size) / (100 + MIN_FREE_MEM_PERCENT);
 787
 788         if (!u->prior)
 789                 u->prior = uct_prior_init(NULL, b);
 790
 791         if (!u->playout)
 792                 u->playout = playout_moggy_init(NULL, b);
 793         u->playout->debug_level = u->debug_level;
 794
 795         u->ownermap.map = malloc2(board_size2(b) * sizeof(u->ownermap.map[0]));
 796         u->stats = malloc2(board_size2(b) * sizeof(u->stats[0]));
 797
 798         if (!u->dynkomi)
 799                 u->dynkomi = uct_dynkomi_init_linear(u, NULL, b);
 800
 801         /* Some things remain uninitialized for now - the opening book
 802          * is not loaded and the tree not set up. */
 803         /* This will be initialized in setup_state() at the first move
 804          * received/requested. This is because right now we are not aware
 805          * about any komi or handicap setup and such. */
 806
 807         return u;
 808 }
 809
 810 struct engine *
 811 engine_uct_init(char *arg, struct board *b)
 812 {
 813         struct uct *u = uct_state_init(arg, b);
 814         struct engine *e = calloc2(1, sizeof(struct engine));
 815         e->name = "UCT Engine";
 816         e->printhook = uct_printhook_ownermap;
 817         e->notify_play = uct_notify_play;
 818         e->chat = uct_chat;
 819         e->genmove = uct_genmove;
 820         e->genmoves = uct_genmoves;
 821         e->dead_group_list = uct_dead_group_list;
 822         e->done = uct_done;
 823         e->data = u;
 824         if (u->slave)
 825                 e->notify = uct_notify;
 826
 827         const char banner[] = "I'm playing UCT. When I'm losing, I will resign, "
 828                 "if I think I win, I play until you pass. "
 829                 "Anyone can send me 'winrate' in private chat to get my assessment of the position.";
 830         if (!u->banner) u->banner = "";
 831         e->comment = malloc2(sizeof(banner) + strlen(u->banner) + 1);
 832         sprintf(e->comment, "%s %s", banner, u->banner);
 833
 834         return e;
 835 }