distributed/distributed.c

   1 /* This is a master for the "distributed" engine. It receives connections
   2  * from slave machines, sends them gtp commands, then aggregates the
   3  * results. It can also act as a proxy for the logs of all slave machines.
   4  * The slave machines must run with engine "uct" (not "distributed").
   5  * The master sends pachi-genmoves gtp commands regularly to each slave,
   6  * gets as replies a list of nodes, their number of playouts
   7  * and their value. The master then picks the most popular move
   8  * among the top level nodes. */
   9
  10 /* With time control, the master waits for all slaves, except
  11  * when the allowed time is already passed. In this case the
  12  * master picks among the available replies, or waits for just
  13  * one reply if there is none yet.
  14  * Without time control, the master waits until the desired
  15  * number of games have been simulated. In this case the -t
  16  * parameter for the master should be the sum of the parameters
  17  * for all slaves. */
  18
  19 /* The master sends updated statistics for the best nodes in each
  20  * genmoves command. They are incremental updates from all other
  21  * slaves (so they exclude contributions from the target slave).
  22  * The slaves reply with just their own stats. So both master and
  23  * slave remember what was previously sent. A slave remembers in
  24  * the tree ("pu" field), which is stable across moves. The slave
  25  * also has a temporary hash table to map received coord paths
  26  * to tree nodes; the hash table is cleared at each new move.
  27  * The master remembers stats in a queue of received buffers that
  28  * are merged together, plus one hash table per slave. The master
  29  * queue and the hash tables are cleared at each new move. */
  30
  31 /* To allow the master to select the best move, slaves also send
  32  * absolute playout counts for the best top level nodes (children
  33  * of the root node), including contributions from other slaves.
  34  * The master sums these counts and picks the best sum, which is
  35  * equivalent to picking the best average. (The master cannot
  36  * use the incremental stats sent in binary form because they
  37  * are not maintained across moves, so playouts from previous
  38  * moves would be lost.) */
  39
  40 /* The master-slave protocol has fault tolerance. If a slave is
  41  * out of sync, the master sends it the appropriate command history. */
  42
  43 /* Pass me arguments like a=b,c=d,...
  44  * Supported arguments:
  45  * slave_port=SLAVE_PORT     slaves connect to this port; this parameter is mandatory.
  46  * max_slaves=MAX_SLAVES     default 24
  47  * shared_nodes=SHARED_NODES default 10K
  48  * stats_hbits=STATS_HBITS   default 21. 2^stats_bits = hash table size
  49  * slaves_quit=0|1           quit gtp command also sent to slaves, default false.
  50  * proxy_port=PROXY_PORT     slaves optionally send their logs to this port.
  51  *    Warning: with proxy_port, the master stderr mixes the logs of all
  52  *    machines but you can separate them again:
  53  *      slave logs:  sed -n '/< .*:/s/.*< /< /p' logfile
  54  *      master logs: perl -0777 -pe 's/<[ <].*:.*\n//g' logfile
  55  */
  56
  57 /* A configuration without proxy would have one master run on masterhost as:
  58  *    zzgo -e distributed slave_port=1234
  59  * and N slaves running as:
  60  *    zzgo -e uct -g masterhost:1234 slave
  61  * With log proxy:
  62  *    zzgo -e distributed slave_port=1234,proxy_port=1235
  63  *    zzgo -e uct -g masterhost:1234 -l masterhost:1235 slave
  64  * If the master itself runs on a machine other than that running gogui,
  65  * gogui-twogtp, kgsGtp or cgosGtp, it can redirect its gtp port:
  66  *    zzgo -e distributed -g 10000 slave_port=1234,proxy_port=1235
  67  */
  68
  69 #include <assert.h>
  70 #include <stdio.h>
  71 #include <stdlib.h>
  72 #include <string.h>
  73 #include <time.h>
  74 #include <alloca.h>
  75 #include <sys/types.h>
  76
  77 #define DEBUG
  78
  79 #include "engine.h"
  80 #include "move.h"
  81 #include "timeinfo.h"
  82 #include "playout.h"
  83 #include "stats.h"
  84 #include "mq.h"
  85 #include "debug.h"
  86 #include "distributed/distributed.h"
  87 #include "distributed/merge.h"
  88
  89 /* Internal engine state. */
  90 struct distributed {
  91         char *slave_port;
  92         char *proxy_port;
  93         int max_slaves;
  94         int shared_nodes;
  95         int stats_hbits;
  96         bool slaves_quit;
  97         struct move my_last_move;
  98         struct move_stats my_last_stats;
  99 };
 100
 101 /* Default number of simulations to perform per move.
 102  * Note that this is in total over all slaves! */
 103 #define DIST_GAMES      80000
 104 static const struct time_info default_ti = {
 105         .period = TT_MOVE,
 106         .dim = TD_GAMES,
 107         .len = { .games = DIST_GAMES },
 108 };
 109
 110 #define get_value(value, color) \
 111         ((color) == S_BLACK ? (value) : 1 - (value))
 112
 113
 114 /* Maximum time (seconds) to wait for answers to fast gtp commands
 115  * (all commands except pachi-genmoves and final_status_list). */
 116 #define MAX_FAST_CMD_WAIT 1.0
 117
 118 /* Maximum time (seconds) to wait for answers to genmoves. */
 119 #define MAX_GENMOVES_WAIT 0.1 /* 100 ms */
 120
 121 /* Minimum time (seconds) to wait before we stop early. This should
 122  * ensure that most slaves have replied at least once. */
 123 #define MIN_EARLY_STOP_WAIT 0.3 /* 300 ms */
 124
 125 /* Display a path as leaf<parent<grandparent...
 126  * Returns the path string in a static buffer; it is NOT safe for
 127  * anything but debugging - in particular, it is NOT thread-safe! */
 128 char *
 129 path2sstr(path_t path, struct board *b)
 130 {
 131         /* Special case for pass and resign. */
 132         if (path < 0) return coord2sstr((coord_t)path, b);
 133
 134         static char buf[16][64];
 135         static int bi = 0;
 136         char *b2;
 137         b2 = buf[bi++ & 15];
 138         *b2 = '\0';
 139         char *s = b2;
 140         char *end = b2 + 64;
 141         coord_t leaf;
 142         while ((leaf = leaf_coord(path, b)) != 0) {
 143                 s += snprintf(s, end - s, "%s<", coord2sstr(leaf, b));
 144                 path = parent_path(path, b);
 145         }
 146         if (s != b2) s[-1] = '\0';
 147         return b2;
 148 }
 149
 150 /* Dispatch a new gtp command to all slaves.
 151  * The slave lock must not be held upon entry and is released upon return.
 152  * args is empty or ends with '\n' */
 153 static enum parse_code
 154 distributed_notify(struct engine *e, struct board *b, int id, char *cmd, char *args, char **reply)
 155 {
 156         struct distributed *dist = e->data;
 157
 158         /* Commands that should not be sent to slaves.
 159          * time_left will be part of next pachi-genmoves,
 160          * we reduce latency by not forwarding it here. */
 161         if ((!strcasecmp(cmd, "quit") && !dist->slaves_quit)
 162             || !strcasecmp(cmd, "uct_gentbook")
 163             || !strcasecmp(cmd, "uct_dumptbook")
 164             || !strcasecmp(cmd, "kgs-chat")
 165             || !strcasecmp(cmd, "time_left")
 166
 167             /* and commands that will be sent to slaves later */
 168             || !strcasecmp(cmd, "genmove")
 169             || !strcasecmp(cmd, "kgs-genmove_cleanup")
 170             || !strcasecmp(cmd, "final_score")
 171             || !strcasecmp(cmd, "final_status_list"))
 172                 return P_OK;
 173
 174         protocol_lock();
 175
 176         // Create a new command to be sent by the slave threads.
 177         new_cmd(b, cmd, args);
 178
 179         /* Wait for replies here. If we don't wait, we run the
 180          * risk of getting out of sync with most slaves and
 181          * sending command history too frequently. */
 182         get_replies(time_now() + MAX_FAST_CMD_WAIT, active_slaves);
 183
 184         protocol_unlock();
 185         return P_OK;
 186 }
 187
 188 /* The playouts sent by slaves for the children of the root node
 189  * include contributions from other slaves. To avoid 32-bit overflow on
 190  * large configurations with many slaves we must average the playouts. */
 191 struct large_stats {
 192         long playouts; // # of playouts
 193         floating_t value; // BLACK wins/playouts
 194 };
 195
 196 static void
 197 large_stats_add_result(struct large_stats *s, floating_t result, long playouts)
 198 {
 199         s->playouts += playouts;
 200         s->value += (result - s->value) * playouts / s->playouts;
 201 }
 202
 203 /* genmoves returns "=id played_own total_playouts threads keep_looking @size"
 204  * then a list of lines "coord playouts value" with absolute counts for
 205  * children of the root node, then a binary array of incr_stats structs.
 206  * To simplify the code, we assume that master and slave have the same architecture
 207  * (store values identically).
 208  * Return the move with most playouts, and additional stats.
 209  * keep_looking is set from a majority vote of the slaves seen so far for this
 210  * move but should not be trusted if too few slaves have been seen.
 211  * Keep this code in sync with uct/slave.c:report_stats().
 212  * slave_lock is held on entry and on return. */
 213 static coord_t
 214 select_best_move(struct board *b, struct large_stats *stats, int *played,
 215                  int *total_playouts, int *total_threads, bool *keep_looking)
 216 {
 217         assert(reply_count > 0);
 218
 219         /* +2 for pass and resign */
 220         memset(stats-2, 0, (board_size2(b)+2) * sizeof(*stats));
 221
 222         coord_t best_move = pass;
 223         long best_playouts = -1;
 224         *played = 0;
 225         *total_playouts = 0;
 226         *total_threads = 0;
 227         int keep = 0;
 228
 229         for (int reply = 0; reply < reply_count; reply++) {
 230                 char *r = gtp_replies[reply];
 231                 int id, o, p, t, k;
 232                 if (sscanf(r, "=%d %d %d %d %d", &id, &o, &p, &t, &k) != 5) continue;
 233                 *played += o;
 234                 *total_playouts += p;
 235                 *total_threads += t;
 236                 keep += k;
 237                 // Skip the rest of the firt line in particular @size
 238                 r = strchr(r, '\n');
 239
 240                 char move[64];
 241                 struct move_stats s;
 242                 while (r && sscanf(++r, "%63s %d " PRIfloating, move, &s.playouts, &s.value) == 3) {
 243                         coord_t c = str2scoord(move, board_size(b));
 244                         assert (c >= resign && c < board_size2(b) && s.playouts >= 0);
 245
 246                         large_stats_add_result(&stats[c], s.value, (long)s.playouts);
 247
 248                         if (stats[c].playouts > best_playouts) {
 249                                 best_playouts = stats[c].playouts;
 250                                 best_move = c;
 251                         }
 252                         r = strchr(r, '\n');
 253                 }
 254         }
 255         for (coord_t c = resign; c < board_size2(b); c++)
 256                 stats[c].playouts /= reply_count;
 257         *keep_looking = keep > reply_count / 2;
 258         return best_move;
 259 }
 260
 261 /* Set the args for the genmoves command. If binary_args is set,
 262  * each slave thred will add the correct binary size when sending
 263  * (see get_binary_arg()). args must have CMDS_SIZE bytes and
 264  * upon return ends with a single \n.
 265  * Keep this code in sync with uct/slave.c:uct_genmoves().
 266  * slave_lock is held on entry and on return but we don't
 267  * rely on the lock here. */
 268 static void
 269 genmoves_args(char *args, enum stone color, int played,
 270               struct time_info *ti, bool binary_args)
 271 {
 272         char *end = args + CMDS_SIZE;
 273         char *s = args + snprintf(args, CMDS_SIZE, "%s %d", stone2str(color), played);
 274
 275         if (ti->dim == TD_WALLTIME) {
 276                 s += snprintf(s, end - s, " %.3f %.3f %d %d",
 277                               ti->len.t.main_time, ti->len.t.byoyomi_time,
 278                               ti->len.t.byoyomi_periods, ti->len.t.byoyomi_stones);
 279         }
 280         s += snprintf(s, end - s, binary_args ? " @0\n" : "\n");
 281 }
 282
 283 /* Time control is mostly done by the slaves, so we use default values here. */
 284 #define FUSEKI_END 20
 285 #define YOSE_START 40
 286 #define MAX_MAINTIME_RATIO 3.0
 287
 288 /* Regularly send genmoves command to the slaves, and select the best move. */
 289 static coord_t *
 290 distributed_genmove(struct engine *e, struct board *b, struct time_info *ti,
 291                     enum stone color, bool pass_all_alive)
 292 {
 293         struct distributed *dist = e->data;
 294         double now = time_now();
 295         double first = now;
 296         char buf[BSIZE]; // debug only
 297
 298         char *cmd = pass_all_alive ? "pachi-genmoves_cleanup" : "pachi-genmoves";
 299         char args[CMDS_SIZE];
 300
 301         coord_t best;
 302         int played, playouts, threads;
 303
 304         if (ti->period == TT_NULL) *ti = default_ti;
 305         struct time_stop stop;
 306         time_stop_conditions(ti, b, FUSEKI_END, YOSE_START, MAX_MAINTIME_RATIO, &stop);
 307         struct time_info saved_ti = *ti;
 308
 309         /* Combined move stats from all slaves, only for children
 310          * of the root node, plus 2 for pass and resign. */
 311         struct large_stats *stats = alloca((board_size2(b)+2) * sizeof(struct large_stats));
 312         stats += 2;
 313
 314         protocol_lock();
 315         clear_receive_queue();
 316
 317         /* Send the first genmoves without stats. */
 318         genmoves_args(args, color, 0, ti, false);
 319         new_cmd(b, cmd, args);
 320
 321         /* Loop until most slaves want to quit or time elapsed. */
 322         int iterations;
 323         for (iterations = 1; ; iterations++) {
 324                 double start = now;
 325                 /* Wait for just one slave to get stats as fresh as possible,
 326                  * or at most 100ms to check if we run out of time. */
 327                 get_replies(now + MAX_GENMOVES_WAIT, 1);
 328                 now = time_now();
 329                 if (ti->dim == TD_WALLTIME)
 330                         time_sub(ti, now - start, false);
 331
 332                 bool keep_looking;
 333                 best = select_best_move(b, stats, &played, &playouts, &threads, &keep_looking);
 334
 335                 if (ti->dim == TD_WALLTIME) {
 336                         if (now - ti->len.t.timer_start >= stop.worst.time) break;
 337                         if (!keep_looking && now - first >= MIN_EARLY_STOP_WAIT) break;
 338                 } else {
 339                         if (!keep_looking || played >= stop.worst.playouts) break;
 340                 }
 341                 if (DEBUGVV(2)) {
 342                         char *coord = coord2sstr(best, b);
 343                         snprintf(buf, sizeof(buf),
 344                                  "temp winner is %s %s with score %1.4f (%d/%d games)"
 345                                  " %d slaves %d threads\n",
 346                                  stone2str(color), coord, get_value(stats[best].value, color),
 347                                  (int)stats[best].playouts, playouts, reply_count, threads);
 348                         logline(NULL, "* ", buf);
 349                 }
 350                 /* Send the command with the same gtp id, to avoid discarding
 351                  * a reply to a previous genmoves at the same move. */
 352                 genmoves_args(args, color, played, ti, true);
 353                 update_cmd(b, cmd, args, false);
 354         }
 355         int replies = reply_count;
 356
 357         /* Do not subtract time spent twice (see gtp_parse). */
 358         *ti = saved_ti;
 359
 360         dist->my_last_move.color = color;
 361         dist->my_last_move.coord = best;
 362         dist->my_last_stats.value = stats[best].value;
 363         dist->my_last_stats.playouts = (int)stats[best].playouts;
 364
 365         /* Tell the slaves to commit to the selected move, overwriting
 366          * the last "pachi-genmoves" in the command history. */
 367         clear_receive_queue();
 368         char coordbuf[4];
 369         char *coord = coord2bstr(coordbuf, best, b);
 370         snprintf(args, sizeof(args), "%s %s\n", stone2str(color), coord);
 371         update_cmd(b, "play", args, true);
 372         protocol_unlock();
 373
 374         if (DEBUGL(1)) {
 375                 double time = now - first + 0.000001; /* avoid divide by zero */
 376                 snprintf(buf, sizeof(buf),
 377                          "GLOBAL WINNER is %s %s with score %1.4f (%d/%d games)\n"
 378                          "genmove %d games in %0.2fs %d slaves %d threads (%d games/s,"
 379                          " %d games/s/slave, %d games/s/thread, %.3f ms/iter)\n",
 380                          stone2str(color), coord, get_value(stats[best].value, color),
 381                          (int)stats[best].playouts, playouts, played, time, replies, threads,
 382                          (int)(played/time), (int)(played/time/replies),
 383                          (int)(played/time/threads), 1000*time/iterations);
 384                 logline(NULL, "* ", buf);
 385         }
 386         if (DEBUGL(3)) {
 387                 int total_hnodes = replies * (1 << dist->stats_hbits);
 388                 merge_print_stats(total_hnodes);
 389         }
 390         return coord_copy(best);
 391 }
 392
 393 static char *
 394 distributed_chat(struct engine *e, struct board *b, char *cmd)
 395 {
 396         struct distributed *dist = e->data;
 397         static char reply[BSIZE];
 398
 399         cmd += strspn(cmd, " \n\t");
 400         if (!strncasecmp(cmd, "winrate", 7)) {
 401                 enum stone color = dist->my_last_move.color;
 402                 snprintf(reply, BSIZE, "In %d playouts at %d machines, %s %s can win with %.2f%% probability.",
 403                          dist->my_last_stats.playouts, active_slaves, stone2str(color),
 404                          coord2sstr(dist->my_last_move.coord, b),
 405                          100 * get_value(dist->my_last_stats.value, color));
 406                 return reply;
 407         }
 408         return NULL;
 409 }
 410
 411 static int
 412 scmp(const void *p1, const void *p2)
 413 {
 414         return strcasecmp(*(char * const *)p1, *(char * const *)p2);
 415 }
 416
 417 static void
 418 distributed_dead_group_list(struct engine *e, struct board *b, struct move_queue *mq)
 419 {
 420         protocol_lock();
 421
 422         new_cmd(b, "final_status_list", "dead\n");
 423         get_replies(time_now() + MAX_FAST_CMD_WAIT, active_slaves);
 424
 425         /* Find the most popular reply. */
 426         qsort(gtp_replies, reply_count, sizeof(char *), scmp);
 427         int best_reply = 0;
 428         int best_count = 1;
 429         int count = 1;
 430         for (int reply = 1; reply < reply_count; reply++) {
 431                 if (!strcmp(gtp_replies[reply], gtp_replies[reply-1])) {
 432                         count++;
 433                 } else {
 434                         count = 1;
 435                 }
 436                 if (count > best_count) {
 437                         best_count = count;
 438                         best_reply = reply;
 439                 }
 440         }
 441
 442         /* Pick the first move of each line as group. */
 443         char *dead = gtp_replies[best_reply];
 444         dead = strchr(dead, ' '); // skip "id "
 445         while (dead && *++dead != '\n') {
 446                 mq_add(mq, str2scoord(dead, board_size(b)), 0);
 447                 dead = strchr(dead, '\n');
 448         }
 449         protocol_unlock();
 450 }
 451
 452 static struct distributed *
 453 distributed_state_init(char *arg, struct board *b)
 454 {
 455         struct distributed *dist = calloc2(1, sizeof(struct distributed));
 456
 457         dist->stats_hbits = DEFAULT_STATS_HBITS;
 458         dist->max_slaves = DEFAULT_MAX_SLAVES;
 459         dist->shared_nodes = DEFAULT_SHARED_NODES;
 460         if (arg) {
 461                 char *optspec, *next = arg;
 462                 while (*next) {
 463                         optspec = next;
 464                         next += strcspn(next, ",");
 465                         if (*next) { *next++ = 0; } else { *next = 0; }
 466
 467                         char *optname = optspec;
 468                         char *optval = strchr(optspec, '=');
 469                         if (optval) *optval++ = 0;
 470
 471                         if (!strcasecmp(optname, "slave_port") && optval) {
 472                                 dist->slave_port = strdup(optval);
 473                         } else if (!strcasecmp(optname, "proxy_port") && optval) {
 474                                 dist->proxy_port = strdup(optval);
 475                         } else if (!strcasecmp(optname, "max_slaves") && optval) {
 476                                 dist->max_slaves = atoi(optval);
 477                         } else if (!strcasecmp(optname, "shared_nodes") && optval) {
 478                                 /* Share at most shared_nodes between master and slave at each genmoves.
 479                                  * Must use the same value in master and slaves. */
 480                                 dist->shared_nodes = atoi(optval);
 481                         } else if (!strcasecmp(optname, "stats_hbits") && optval) {
 482                                 /* Set hash table size to 2^stats_hbits for the shared stats. */
 483                                 dist->stats_hbits = atoi(optval);
 484                         } else if (!strcasecmp(optname, "slaves_quit")) {
 485                                 dist->slaves_quit = !optval || atoi(optval);
 486                         } else {
 487                                 fprintf(stderr, "distributed: Invalid engine argument %s or missing value\n", optname);
 488                         }
 489                 }
 490         }
 491
 492         gtp_replies = calloc2(dist->max_slaves, sizeof(char *));
 493
 494         if (!dist->slave_port) {
 495                 fprintf(stderr, "distributed: missing slave_port\n");
 496                 exit(1);
 497         }
 498
 499         merge_init(&default_sstate, dist->shared_nodes, dist->stats_hbits, dist->max_slaves);
 500         protocol_init(dist->slave_port, dist->proxy_port, dist->max_slaves);
 501
 502         return dist;
 503 }
 504
 505 struct engine *
 506 engine_distributed_init(char *arg, struct board *b)
 507 {
 508         struct distributed *dist = distributed_state_init(arg, b);
 509         struct engine *e = calloc2(1, sizeof(struct engine));
 510         e->name = "Distributed Engine";
 511         e->comment = "I'm playing the distributed engine. When I'm losing, I will resign, "
 512                 "if I think I win, I play until you pass. "
 513                 "Anyone can send me 'winrate' in private chat to get my assessment of the position.";
 514         e->notify = distributed_notify;
 515         e->genmove = distributed_genmove;
 516         e->dead_group_list = distributed_dead_group_list;
 517         e->chat = distributed_chat;
 518         e->data = dist;
 519         // Keep the threads and the open socket connections:
 520         e->keep_on_clear = true;
 521
 522         return e;
 523 }