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
 122 /* Display a path as leaf<parent<grandparent...
 123  * Returns the path string in a static buffer; it is NOT safe for
 124  * anything but debugging - in particular, it is NOT thread-safe! */
 125 char *
 126 path2sstr(path_t path, struct board *b)
 127 {
 128         /* Special case for pass and resign. */
 129         if (path < 0) return coord2sstr((coord_t)path, b);
 130
 131         static char buf[16][64];
 132         static int bi = 0;
 133         char *b2;
 134         b2 = buf[bi++ & 15];
 135         *b2 = '\0';
 136         char *s = b2;
 137         char *end = b2 + 64;
 138         coord_t leaf;
 139         while ((leaf = leaf_coord(path, b)) != 0) {
 140                 s += snprintf(s, end - s, "%s<", coord2sstr(leaf, b));
 141                 path = parent_path(path, b);
 142         }
 143         if (s != b2) s[-1] = '\0';
 144         return b2;
 145 }
 146
 147 /* Dispatch a new gtp command to all slaves.
 148  * The slave lock must not be held upon entry and is released upon return.
 149  * args is empty or ends with '\n' */
 150 static enum parse_code
 151 distributed_notify(struct engine *e, struct board *b, int id, char *cmd, char *args, char **reply)
 152 {
 153         struct distributed *dist = e->data;
 154
 155         /* Commands that should not be sent to slaves.
 156          * time_left will be part of next pachi-genmoves,
 157          * we reduce latency by not forwarding it here. */
 158         if ((!strcasecmp(cmd, "quit") && !dist->slaves_quit)
 159             || !strcasecmp(cmd, "uct_genbook")
 160             || !strcasecmp(cmd, "uct_dumpbook")
 161             || !strcasecmp(cmd, "kgs-chat")
 162             || !strcasecmp(cmd, "time_left")
 163
 164             /* and commands that will be sent to slaves later */
 165             || !strcasecmp(cmd, "genmove")
 166             || !strcasecmp(cmd, "kgs-genmove_cleanup")
 167             || !strcasecmp(cmd, "final_score")
 168             || !strcasecmp(cmd, "final_status_list"))
 169                 return P_OK;
 170
 171         protocol_lock();
 172
 173         // Create a new command to be sent by the slave threads.
 174         new_cmd(b, cmd, args);
 175
 176         /* Wait for replies here. If we don't wait, we run the
 177          * risk of getting out of sync with most slaves and
 178          * sending command history too frequently. */
 179         get_replies(time_now() + MAX_FAST_CMD_WAIT, active_slaves);
 180
 181         protocol_unlock();
 182         return P_OK;
 183 }
 184
 185 /* genmoves returns "=id played_own total_playouts threads keep_looking @size"
 186  * then a list of lines "coord playouts value" with absolute counts for
 187  * children of the root node, then a binary array of incr_stats structs.
 188  * To simplify the code, we assume that master and slave have the same architecture
 189  * (store values identically).
 190  * Return the move with most playouts, and additional stats.
 191  * Keep this code in sync with uct/slave.c:report_stats().
 192  * slave_lock is held on entry and on return. */
 193 static coord_t
 194 select_best_move(struct board *b, struct move_stats *stats, int *played,
 195                  int *total_playouts, int *total_threads, bool *keep_looking)
 196 {
 197         assert(reply_count > 0);
 198
 199         /* +2 for pass and resign */
 200         memset(stats-2, 0, (board_size2(b)+2) * sizeof(*stats));
 201
 202         coord_t best_move = pass;
 203         int best_playouts = -1;
 204         *played = 0;
 205         *total_playouts = 0;
 206         *total_threads = 0;
 207         int keep = 0;
 208
 209         for (int reply = 0; reply < reply_count; reply++) {
 210                 char *r = gtp_replies[reply];
 211                 int id, o, p, t, k;
 212                 if (sscanf(r, "=%d %d %d %d %d", &id, &o, &p, &t, &k) != 5) continue;
 213                 *played += o;
 214                 *total_playouts += p;
 215                 *total_threads += t;
 216                 keep += k;
 217                 // Skip the rest of the firt line in particular @size
 218                 r = strchr(r, '\n');
 219
 220                 char move[64];
 221                 struct move_stats s;
 222                 while (r && sscanf(++r, "%63s %d %f", move, &s.playouts, &s.value) == 3) {
 223                         coord_t c = str2scoord(move, board_size(b));
 224                         assert (c >= resign && c < board_size2(b) && s.playouts >= 0);
 225
 226                         stats_add_result(&stats[c], s.value, s.playouts);
 227
 228                         if (stats[c].playouts > best_playouts) {
 229                                 best_playouts = stats[c].playouts;
 230                                 best_move = c;
 231                         }
 232                         r = strchr(r, '\n');
 233                 }
 234         }
 235         *keep_looking = keep > reply_count / 2;
 236         return best_move;
 237 }
 238
 239 /* Set the args for the genmoves command. If binary_args is set,
 240  * each slave thred will add the correct binary size when sending
 241  * (see get_binary_arg()). args must have CMDS_SIZE bytes and
 242  * upon return ends with a single \n.
 243  * Keep this code in sync with uct/slave.c:uct_genmoves().
 244  * slave_lock is held on entry and on return but we don't
 245  * rely on the lock here. */
 246 static void
 247 genmoves_args(char *args, enum stone color, int played,
 248               struct time_info *ti, bool binary_args)
 249 {
 250         char *end = args + CMDS_SIZE;
 251         char *s = args + snprintf(args, CMDS_SIZE, "%s %d", stone2str(color), played);
 252
 253         if (ti->dim == TD_WALLTIME) {
 254                 s += snprintf(s, end - s, " %.3f %.3f %d %d",
 255                               ti->len.t.main_time, ti->len.t.byoyomi_time,
 256                               ti->len.t.byoyomi_periods, ti->len.t.byoyomi_stones);
 257         }
 258         s += snprintf(s, end - s, binary_args ? " @0\n" : "\n");
 259 }
 260
 261 /* Time control is mostly done by the slaves, so we use default values here. */
 262 #define FUSEKI_END 20
 263 #define YOSE_START 40
 264
 265 /* In the ascii reply to genmoves, each slave sends absolute counts
 266  * including contributions from other slaves. For human display
 267  * reduce the sum to an average. */
 268 #define stats_average(playouts) ((playouts) / reply_count)
 269
 270 /* Regularly send genmoves command to the slaves, and select the best move. */
 271 static coord_t *
 272 distributed_genmove(struct engine *e, struct board *b, struct time_info *ti,
 273                     enum stone color, bool pass_all_alive)
 274 {
 275         struct distributed *dist = e->data;
 276         double now = time_now();
 277         double first = now;
 278         char buf[BSIZE]; // debug only
 279
 280         char *cmd = pass_all_alive ? "pachi-genmoves_cleanup" : "pachi-genmoves";
 281         char args[CMDS_SIZE];
 282
 283         coord_t best;
 284         int played, playouts, threads;
 285
 286         if (ti->period == TT_NULL) *ti = default_ti;
 287         struct time_stop stop;
 288         time_stop_conditions(ti, b, FUSEKI_END, YOSE_START, &stop);
 289         struct time_info saved_ti = *ti;
 290
 291         /* Combined move stats from all slaves, only for children
 292          * of the root node, plus 2 for pass and resign. */
 293         struct move_stats *stats = alloca((board_size2(b)+2) * sizeof(struct move_stats));
 294         stats += 2;
 295
 296         protocol_lock();
 297         clear_receive_queue();
 298
 299         /* Send the first genmoves without stats. */
 300         genmoves_args(args, color, 0, ti, false);
 301         new_cmd(b, cmd, args);
 302
 303         /* Loop until most slaves want to quit or time elapsed. */
 304         int iterations;
 305         for (iterations = 1; ; iterations++) {
 306                 double start = now;
 307                 /* Wait for just one slave to get stats as fresh as possible,
 308                  * or at most 100ms to check if we run out of time. */
 309                 get_replies(now + MAX_GENMOVES_WAIT, 1);
 310                 now = time_now();
 311                 if (ti->dim == TD_WALLTIME)
 312                         time_sub(ti, now - start, false);
 313
 314                 bool keep_looking;
 315                 best = select_best_move(b, stats, &played, &playouts, &threads, &keep_looking);
 316
 317                 if (!keep_looking) break;
 318                 if (ti->dim == TD_WALLTIME) {
 319                         if (now - ti->len.t.timer_start >= stop.worst.time) break;
 320                 } else {
 321                         if (played >= stop.worst.playouts) break;
 322                 }
 323                 if (DEBUGVV(2)) {
 324                         char *coord = coord2sstr(best, b);
 325                         snprintf(buf, sizeof(buf),
 326                                  "temp winner is %s %s with score %1.4f (%d/%d games)"
 327                                  " %d slaves %d threads\n",
 328                                  stone2str(color), coord, get_value(stats[best].value, color),
 329                                  stats_average(stats[best].playouts), playouts, reply_count, threads);
 330                         logline(NULL, "* ", buf);
 331                 }
 332                 /* Send the command with the same gtp id, to avoid discarding
 333                  * a reply to a previous genmoves at the same move. */
 334                 genmoves_args(args, color, played, ti, true);
 335                 update_cmd(b, cmd, args, false);
 336         }
 337         int replies = reply_count;
 338
 339         /* Do not subtract time spent twice (see gtp_parse). */
 340         *ti = saved_ti;
 341
 342         dist->my_last_move.color = color;
 343         dist->my_last_move.coord = best;
 344         dist->my_last_stats.value = stats[best].value;
 345         dist->my_last_stats.playouts = stats_average(stats[best].playouts);
 346
 347         /* Tell the slaves to commit to the selected move, overwriting
 348          * the last "pachi-genmoves" in the command history. */
 349         clear_receive_queue();
 350         char coordbuf[4];
 351         char *coord = coord2bstr(coordbuf, best, b);
 352         snprintf(args, sizeof(args), "%s %s\n", stone2str(color), coord);
 353         update_cmd(b, "play", args, true);
 354         protocol_unlock();
 355
 356         if (DEBUGL(1)) {
 357                 double time = now - first + 0.000001; /* avoid divide by zero */
 358                 snprintf(buf, sizeof(buf),
 359                          "GLOBAL WINNER is %s %s with score %1.4f (%d/%d games)\n"
 360                          "genmove %d games in %0.2fs %d slaves %d threads (%d games/s,"
 361                          " %d games/s/slave, %d games/s/thread, %.3f ms/iter)\n",
 362                          stone2str(color), coord, get_value(stats[best].value, color),
 363                          stats[best].playouts, playouts, played, time, replies, threads,
 364                          (int)(played/time), (int)(played/time/replies),
 365                          (int)(played/time/threads), 1000*time/iterations);
 366                 logline(NULL, "* ", buf);
 367         }
 368         if (DEBUGL(3)) {
 369                 int total_hnodes = replies * (1 << dist->stats_hbits);
 370                 merge_print_stats(total_hnodes);
 371         }
 372         return coord_copy(best);
 373 }
 374
 375 static char *
 376 distributed_chat(struct engine *e, struct board *b, char *cmd)
 377 {
 378         struct distributed *dist = e->data;
 379         static char reply[BSIZE];
 380
 381         cmd += strspn(cmd, " \n\t");
 382         if (!strncasecmp(cmd, "winrate", 7)) {
 383                 enum stone color = dist->my_last_move.color;
 384                 snprintf(reply, BSIZE, "In %d playouts at %d machines, %s %s can win with %.2f%% probability.",
 385                          dist->my_last_stats.playouts, active_slaves, stone2str(color),
 386                          coord2sstr(dist->my_last_move.coord, b),
 387                          100 * get_value(dist->my_last_stats.value, color));
 388                 return reply;
 389         }
 390         return NULL;
 391 }
 392
 393 static int
 394 scmp(const void *p1, const void *p2)
 395 {
 396         return strcasecmp(*(char * const *)p1, *(char * const *)p2);
 397 }
 398
 399 static void
 400 distributed_dead_group_list(struct engine *e, struct board *b, struct move_queue *mq)
 401 {
 402         protocol_lock();
 403
 404         new_cmd(b, "final_status_list", "dead\n");
 405         get_replies(time_now() + MAX_FAST_CMD_WAIT, active_slaves);
 406
 407         /* Find the most popular reply. */
 408         qsort(gtp_replies, reply_count, sizeof(char *), scmp);
 409         int best_reply = 0;
 410         int best_count = 1;
 411         int count = 1;
 412         for (int reply = 1; reply < reply_count; reply++) {
 413                 if (!strcmp(gtp_replies[reply], gtp_replies[reply-1])) {
 414                         count++;
 415                 } else {
 416                         count = 1;
 417                 }
 418                 if (count > best_count) {
 419                         best_count = count;
 420                         best_reply = reply;
 421                 }
 422         }
 423
 424         /* Pick the first move of each line as group. */
 425         char *dead = gtp_replies[best_reply];
 426         dead = strchr(dead, ' '); // skip "id "
 427         while (dead && *++dead != '\n') {
 428                 mq_add(mq, str2scoord(dead, board_size(b)), 0);
 429                 dead = strchr(dead, '\n');
 430         }
 431         protocol_unlock();
 432 }
 433
 434 static struct distributed *
 435 distributed_state_init(char *arg, struct board *b)
 436 {
 437         struct distributed *dist = calloc2(1, sizeof(struct distributed));
 438
 439         dist->stats_hbits = DEFAULT_STATS_HBITS;
 440         dist->max_slaves = DEFAULT_MAX_SLAVES;
 441         dist->shared_nodes = DEFAULT_SHARED_NODES;
 442         if (arg) {
 443                 char *optspec, *next = arg;
 444                 while (*next) {
 445                         optspec = next;
 446                         next += strcspn(next, ",");
 447                         if (*next) { *next++ = 0; } else { *next = 0; }
 448
 449                         char *optname = optspec;
 450                         char *optval = strchr(optspec, '=');
 451                         if (optval) *optval++ = 0;
 452
 453                         if (!strcasecmp(optname, "slave_port") && optval) {
 454                                 dist->slave_port = strdup(optval);
 455                         } else if (!strcasecmp(optname, "proxy_port") && optval) {
 456                                 dist->proxy_port = strdup(optval);
 457                         } else if (!strcasecmp(optname, "max_slaves") && optval) {
 458                                 dist->max_slaves = atoi(optval);
 459                         } else if (!strcasecmp(optname, "shared_nodes") && optval) {
 460                                 /* Share at most shared_nodes between master and slave at each genmoves.
 461                                  * Must use the same value in master and slaves. */
 462                                 dist->shared_nodes = atoi(optval);
 463                         } else if (!strcasecmp(optname, "stats_hbits") && optval) {
 464                                 /* Set hash table size to 2^stats_hbits for the shared stats. */
 465                                 dist->stats_hbits = atoi(optval);
 466                         } else if (!strcasecmp(optname, "slaves_quit")) {
 467                                 dist->slaves_quit = !optval || atoi(optval);
 468                         } else {
 469                                 fprintf(stderr, "distributed: Invalid engine argument %s or missing value\n", optname);
 470                         }
 471                 }
 472         }
 473
 474         gtp_replies = calloc2(dist->max_slaves, sizeof(char *));
 475
 476         if (!dist->slave_port) {
 477                 fprintf(stderr, "distributed: missing slave_port\n");
 478                 exit(1);
 479         }
 480
 481         merge_init(&default_sstate, dist->shared_nodes, dist->stats_hbits, dist->max_slaves);
 482         protocol_init(dist->slave_port, dist->proxy_port, dist->max_slaves);
 483
 484         return dist;
 485 }
 486
 487 struct engine *
 488 engine_distributed_init(char *arg, struct board *b)
 489 {
 490         struct distributed *dist = distributed_state_init(arg, b);
 491         struct engine *e = calloc2(1, sizeof(struct engine));
 492         e->name = "Distributed Engine";
 493         e->comment = "I'm playing the distributed engine. When I'm losing, I will resign, "
 494                 "if I think I win, I play until you pass. "
 495                 "Anyone can send me 'winrate' in private chat to get my assessment of the position.";
 496         e->notify = distributed_notify;
 497         e->genmove = distributed_genmove;
 498         e->dead_group_list = distributed_dead_group_list;
 499         e->chat = distributed_chat;
 500         e->data = dist;
 501         // Keep the threads and the open socket connections:
 502         e->keep_on_clear = true;
 503
 504         return e;
 505 }