model-net-lp.c 31.9 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11
/*
 * Copyright (C) 2014 University of Chicago.
 * See COPYRIGHT notice in top-level directory.
 *
 */

#include <stddef.h>
#include <assert.h>
#include "codes/model-net.h"
#include "codes/model-net-method.h"
#include "codes/model-net-lp.h"
12
#include "codes/model-net-sched.h"
13 14 15 16 17 18 19 20
#include "codes/codes_mapping.h"
#include "codes/jenkins-hash.h"

#define MN_NAME "model_net_base"

/**** BEGIN SIMULATION DATA STRUCTURES ****/

int model_net_base_magic;
21
int mn_sample_enabled = 0;
22 23 24 25 26

// message-type specific offsets - don't want to get bitten later by alignment
// issues...
static int msg_offsets[MAX_NETS];

27 28
typedef struct model_net_base_params_s {
    model_net_sched_cfg_params sched_params;
29
    uint64_t packet_size;
Nikhil's avatar
Nikhil committed
30
    int num_queues;
31
    int use_recv_queue;
Nikhil's avatar
Nikhil committed
32 33
    tw_stime nic_seq_delay;
    int node_copy_queues;
34 35
} model_net_base_params;

36
/* annotation-specific parameters (unannotated entry occurs at the
37 38 39 40 41
 * last index) */
static int                       num_params = 0;
static const char              * annos[CONFIGURATION_MAX_ANNOS];
static model_net_base_params     all_params[CONFIGURATION_MAX_ANNOS];

42
static tw_stime mn_sample_interval = 0.0;
43
static tw_stime mn_sample_end = 0.0;
Nikhil's avatar
Nikhil committed
44 45
static int servers_per_node_queue = -1;
extern tw_stime codes_cn_delay;
46

47
typedef struct model_net_base_state {
Nikhil's avatar
Nikhil committed
48
    int net_id, nics_per_router;
49
    // whether scheduler loop is running
Nikhil's avatar
Nikhil committed
50
    int *in_sched_send_loop, in_sched_recv_loop;
51 52 53
    // unique message id counter. This doesn't get decremented on RC to prevent
    // optimistic orderings using "stale" ids
    uint64_t msg_id;
54
    // model-net schedulers
Nikhil's avatar
Nikhil committed
55
    model_net_sched **sched_send, *sched_recv;
56 57
    // parameters
    const model_net_base_params * params;
58 59 60
    // lp type and state of underlying model net method - cache here so we
    // don't have to constantly look up
    const tw_lptype *sub_type;
61
    const st_model_types *sub_model_type;
62
    void *sub_state;
Nikhil's avatar
Nikhil committed
63 64
    tw_stime next_available_time;
    tw_stime *node_copy_next_available_time;
65 66
} model_net_base_state;

67

68 69 70 71
/**** END SIMULATION DATA STRUCTURES ****/

/**** BEGIN LP, EVENT PROCESSING FUNCTION DECLS ****/

72
/* ROSS LP processing functions */
73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
static void model_net_base_lp_init(
        model_net_base_state * ns,
        tw_lp * lp);
static void model_net_base_event(
        model_net_base_state * ns,
        tw_bf * b,
        model_net_wrap_msg * m,
        tw_lp * lp);
static void model_net_base_event_rc(
        model_net_base_state * ns,
        tw_bf * b,
        model_net_wrap_msg * m,
        tw_lp * lp);
static void model_net_base_finalize(
        model_net_base_state * ns,
        tw_lp * lp);

/* event type handlers */
static void handle_new_msg(
        model_net_base_state * ns,
        tw_bf *b,
        model_net_wrap_msg * m,
        tw_lp * lp);
static void handle_sched_next(
        model_net_base_state * ns,
        tw_bf *b,
        model_net_wrap_msg * m,
        tw_lp * lp);
static void handle_new_msg_rc(
        model_net_base_state * ns,
        tw_bf *b,
        model_net_wrap_msg * m,
        tw_lp * lp);
static void handle_sched_next_rc(
        model_net_base_state * ns,
        tw_bf *b,
        model_net_wrap_msg * m,
        tw_lp * lp);

/* ROSS function pointer table for this LP */
tw_lptype model_net_base_lp = {
114 115 116 117
    (init_f) model_net_base_lp_init,
    (pre_run_f) NULL,
    (event_f) model_net_base_event,
    (revent_f) model_net_base_event_rc,
118 119
    (commit_f) NULL,
    (final_f)  model_net_base_finalize,
120 121
    (map_f) codes_mapping,
    sizeof(model_net_base_state),
122 123
};

124 125 126 127
/* setup for the ROSS event tracing
 * can have a different function for  rbev_trace_f and ev_trace_f
 * but right now it is set to the same function for both
 */
128
void mn_event_collect(model_net_wrap_msg *m, tw_lp *lp, char *buffer, int *collect_flag)
129
{
130 131 132
    // assigning large numbers to message types to make it easier to
    // determine which messages are model net base LP msgs
    int type;
133
    void * sub_msg;
134 135 136
    switch (m->h.event_type){
        case MN_BASE_NEW_MSG:
            type = 9000;
137
            memcpy(buffer, &type, sizeof(type));
138 139 140
            break;
        case MN_BASE_SCHED_NEXT:
            type = 9001;
141
            memcpy(buffer, &type, sizeof(type));
142 143 144
            break;
        case MN_BASE_SAMPLE: 
            type = 9002;
145
            memcpy(buffer, &type, sizeof(type));
146 147
            break;
        case MN_BASE_PASS:
148
            sub_msg = ((char*)m)+msg_offsets[((model_net_base_state*)lp->cur_state)->net_id];
149 150
            if (g_st_ev_trace == RB_TRACE || g_st_ev_trace == COMMIT_TRACE)
                (((model_net_base_state*)lp->cur_state)->sub_model_type->rbev_trace)(sub_msg, lp, buffer, collect_flag);
151
            else if (g_st_ev_trace == FULL_TRACE)
152
                (((model_net_base_state*)lp->cur_state)->sub_model_type->ev_trace)(sub_msg, lp, buffer, collect_flag);
153
            break;
154
        default:  // this shouldn't happen, but can help detect an issue
155 156 157
            type = 9004;
            break;
    }
158 159
}

160 161 162 163 164 165 166 167 168 169
void mn_model_stat_collect(model_net_base_state *s, tw_lp *lp, char *buffer)
{
    // need to call the model level stats collection fn
    (*s->sub_model_type->model_stat_fn)(s->sub_state, lp, buffer);
    return;
}

st_model_types mn_model_types[MAX_NETS];

st_model_types mn_model_base_type = {
170
    (rbev_trace_f) mn_event_collect,
171
     sizeof(int),
172
     (ev_trace_f) mn_event_collect,
173
     sizeof(int),
174 175
     (model_stat_f) mn_model_stat_collect,
     0
176 177
};

178 179 180 181
/**** END LP, EVENT PROCESSING FUNCTION DECLS ****/

/**** BEGIN IMPLEMENTATIONS ****/

182
void model_net_enable_sampling(tw_stime interval, tw_stime end)
183 184
{
    mn_sample_interval = interval;
185
    mn_sample_end = end;
186 187 188 189 190 191 192 193 194 195 196
    mn_sample_enabled = 1;
}

int model_net_sampling_enabled(void)
{
    return mn_sample_enabled;
}

// schedule sample event - want to be precise, so no noise here
static void issue_sample_event(tw_lp *lp)
{
197 198 199 200 201 202
    if (tw_now(lp) + mn_sample_interval < mn_sample_end + 0.0001) {
        tw_event *e = tw_event_new(lp->gid, mn_sample_interval, lp);
        model_net_wrap_msg *m = tw_event_data(e);
        msg_set_header(model_net_base_magic, MN_BASE_SAMPLE, lp->gid, &m->h);
        tw_event_send(e);
    }
203 204
}

205 206 207 208
void model_net_base_register(int *do_config_nets){
    // here, we initialize ALL lp types to use the base type
    for (int i = 0; i < MAX_NETS; i++){
        if (do_config_nets[i]){
209 210 211 212 213 214 215
            // some model-net lps need custom registration hooks (dragonfly).
            // Those that don't NULL out the reg. function
            if (method_array[i]->mn_register == NULL)
                lp_type_register(model_net_lp_config_names[i],
                        &model_net_base_lp);
            else
                method_array[i]->mn_register(&model_net_base_lp);
216
            if (g_st_ev_trace || g_st_model_stats) // for ROSS event tracing
217
            {
218 219 220 221
                memcpy(&mn_model_types[i], &mn_model_base_type, sizeof(st_model_types));

                if (method_array[i]->mn_model_stat_register == NULL)
                    st_model_type_register(model_net_lp_config_names[i], &mn_model_types[i]);
222
                else
223
                    method_array[i]->mn_model_stat_register(&mn_model_types[i]);
224
            }
225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
        }
    }
}

static void base_read_config(const char * anno, model_net_base_params *p){
    char sched[MAX_NAME_LENGTH];
    long int packet_size_l = 0;
    uint64_t packet_size;
    int ret;

    ret = configuration_get_value(&config, "PARAMS", "modelnet_scheduler",
            anno, sched, MAX_NAME_LENGTH);
    configuration_get_value_longint(&config, "PARAMS", "packet_size", anno,
            &packet_size_l);
    packet_size = packet_size_l;

241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
    if (ret > 0){
        int i;
        for (i = 0; i < MAX_SCHEDS; i++){
            if (strcmp(sched_names[i], sched) == 0){
                p->sched_params.type = i;
                break;
            }
        }
        if (i == MAX_SCHEDS){
            tw_error(TW_LOC,"Unknown value for PARAMS:modelnet-scheduler : "
                    "%s", sched);
        }
    }
    else{
        // default: FCFS
        p->sched_params.type = MN_SCHED_FCFS;
    }

Nikhil's avatar
Nikhil committed
259 260 261 262 263 264 265 266
    p->num_queues = 1;
    ret = configuration_get_value_int(&config, "PARAMS", "num_injection_queues", anno,
            &p->num_queues);
    if(ret && !g_tw_mynode) {
        fprintf(stdout, "NIC num injection port not specified, "
                "setting to %d\n", p->num_queues);
    }

267
    p->nic_seq_delay = 10;
Nikhil's avatar
Nikhil committed
268 269 270 271 272 273 274
    ret = configuration_get_value_double(&config, "PARAMS", "nic_seq_delay", anno,
            &p->nic_seq_delay);
    if(ret && !g_tw_mynode) {
        fprintf(stdout, "NIC seq delay not specified, "
                "setting to %lf\n", p->nic_seq_delay);
    }

Nikhil's avatar
Nikhil committed
275
    p->node_copy_queues = 1;
Nikhil's avatar
Nikhil committed
276 277 278 279 280 281 282
    ret = configuration_get_value_int(&config, "PARAMS", "node_copy_queues", anno,
            &p->node_copy_queues);
    if(ret && !g_tw_mynode) {
        fprintf(stdout, "NIC num copy queues not specified, "
                "setting to %d\n", p->node_copy_queues);
    }

283 284
    // get scheduler-specific parameters
    if (p->sched_params.type == MN_SCHED_PRIO){
285
        // prio scheduler uses default parameters
286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
        int             * num_prios = &p->sched_params.u.prio.num_prios;
        enum sched_type * sub_stype = &p->sched_params.u.prio.sub_stype;
        // number of priorities to allocate
        ret = configuration_get_value_int(&config, "PARAMS",
                "prio-sched-num-prios", anno, num_prios);
        if (ret != 0)
            *num_prios = 10;

        ret = configuration_get_value(&config, "PARAMS",
                "prio-sched-sub-sched", anno, sched, MAX_NAME_LENGTH);
        if (ret == 0)
            *sub_stype = MN_SCHED_FCFS;
        else{
            int i;
            for (i = 0; i < MAX_SCHEDS; i++){
                if (strcmp(sched_names[i], sched) == 0){
                    *sub_stype = i;
                    break;
                }
            }
            if (i == MAX_SCHEDS){
                tw_error(TW_LOC, "Unknown value for "
                        "PARAMS:prio-sched-sub-sched %s", sched);
            }
            else if (i == MN_SCHED_PRIO){
                tw_error(TW_LOC, "priority scheduler cannot be used as a "
                        "priority scheduler's sub sched "
                        "(PARAMS:prio-sched-sub-sched)");
            }
        }
316 317
    }

318 319 320 321 322 323 324 325 326 327 328 329 330
    if (p->sched_params.type == MN_SCHED_FCFS_FULL ||
            (p->sched_params.type == MN_SCHED_PRIO &&
             p->sched_params.u.prio.sub_stype == MN_SCHED_FCFS_FULL)){
        // override packet size to something huge (leave a bit in the unlikely
        // case that an op using packet size causes overflow)
        packet_size = 1ull << 62;
    }
    else if (!packet_size &&
            (p->sched_params.type != MN_SCHED_FCFS_FULL ||
             (p->sched_params.type == MN_SCHED_PRIO &&
              p->sched_params.u.prio.sub_stype != MN_SCHED_FCFS_FULL))){
        packet_size = 512;
        fprintf(stderr, "WARNING, no packet size specified, setting packet "
Jonathan Jenkins's avatar
Jonathan Jenkins committed
331
                "size to %llu\n", LLU(packet_size));
332 333 334
    }


335 336 337 338
    p->packet_size = packet_size;
}

void model_net_base_configure(){
339 340 341 342 343
    uint32_t h1=0, h2=0;

    bj_hashlittle2(MN_NAME, strlen(MN_NAME), &h1, &h2);
    model_net_base_magic = h1+h2;

344 345 346
    // set up offsets - doesn't matter if they are actually used or not
    msg_offsets[SIMPLENET] =
        offsetof(model_net_wrap_msg, msg.m_snet);
Jonathan Jenkins's avatar
Jonathan Jenkins committed
347 348
    msg_offsets[SIMPLEP2P] =
        offsetof(model_net_wrap_msg, msg.m_sp2p);
349 350 351
    msg_offsets[TORUS] =
        offsetof(model_net_wrap_msg, msg.m_torus);
    msg_offsets[DRAGONFLY] =
352
        offsetof(model_net_wrap_msg, msg.m_dfly);
353 354
    // note: dragonfly router uses the same event struct
    msg_offsets[DRAGONFLY_ROUTER] =
355
        offsetof(model_net_wrap_msg, msg.m_dfly);
356
    msg_offsets[DRAGONFLY_CUSTOM] =
357
        offsetof(model_net_wrap_msg, msg.m_custom_dfly);
358
    msg_offsets[DRAGONFLY_CUSTOM_ROUTER] =
359
        offsetof(model_net_wrap_msg, msg.m_custom_dfly);
360 361
    msg_offsets[SLIMFLY] =
        offsetof(model_net_wrap_msg, msg.m_slim);
362 363
    msg_offsets[FATTREE] =
	offsetof(model_net_wrap_msg, msg.m_fat);
364 365
    msg_offsets[LOGGP] =
        offsetof(model_net_wrap_msg, msg.m_loggp);
Nikhil's avatar
Nikhil committed
366 367 368 369
    msg_offsets[EXPRESS_MESH] =
        offsetof(model_net_wrap_msg, msg.m_em);
    msg_offsets[EXPRESS_MESH_ROUTER] =
        offsetof(model_net_wrap_msg, msg.m_em);
370

371 372 373 374 375 376 377 378 379 380
    // perform the configuration(s)
    // This part is tricky, as we basically have to look up all annotations that
    // have LP names of the form modelnet_*. For each of those, we need to read
    // the base parameters
    // - the init is a little easier as we can use the LP-id to look up the
    // annotation

    // first grab all of the annotations and store locally
    for (int c = 0; c < lpconf.lpannos_count; c++){
        const config_anno_map_t *amap = &lpconf.lpannos[c];
381
        if (strncmp("modelnet_", amap->lp_name.ptr, 9) == 0){
382 383 384
            for (int n = 0; n < amap->num_annos; n++){
                int a;
                for (a = 0; a < num_params; a++){
385 386
                    if (annos[a] != NULL && amap->annotations[n].ptr != NULL &&
                            strcmp(amap->annotations[n].ptr, annos[a]) == 0){
387 388 389 390 391
                        break;
                    }
                }
                if (a == num_params){
                    // found a new annotation
392
                    annos[num_params++] = amap->annotations[n].ptr;
393 394 395 396 397 398 399 400 401 402 403 404 405
                }
            }
            if (amap->has_unanno_lp){
                int a;
                for (a = 0; a < num_params; a++){
                    if (annos[a] == NULL)
                        break;
                }
                if (a == num_params){
                    // found a new (empty) annotation
                    annos[num_params++] = NULL;
                }
            }
406 407
        }
    }
408 409 410 411 412 413

    // now that we have all of the annos for all of the networks, loop through
    // and read the configs
    for (int i = 0; i < num_params; i++){
        base_read_config(annos[i], &all_params[i]);
    }
414 415 416 417 418 419
}

void model_net_base_lp_init(
        model_net_base_state * ns,
        tw_lp * lp){
    // obtain the underlying lp type through codes-mapping
Nikhil's avatar
Nikhil committed
420
    char lp_type_name[MAX_NAME_LENGTH], anno[MAX_NAME_LENGTH], group[MAX_NAME_LENGTH];
421 422
    int dummy;

Nikhil's avatar
Nikhil committed
423
    codes_mapping_get_lp_info(lp->gid, group, &dummy,
424 425 426 427 428 429 430 431 432 433
            lp_type_name, &dummy, anno, &dummy, &dummy);

    // get annotation-specific parameters
    for (int i = 0; i < num_params; i++){
        if ((anno[0]=='\0' && annos[i] == NULL) ||
                strcmp(anno, annos[i]) == 0){
            ns->params = &all_params[i];
            break;
        }
    }
434 435 436 437 438 439 440 441

    // find the corresponding method name / index
    for (int i = 0; i < MAX_NETS; i++){
        if (strcmp(model_net_lp_config_names[i], lp_type_name) == 0){
            ns->net_id = i;
            break;
        }
    }
Nikhil's avatar
Nikhil committed
442 443 444 445 446 447 448 449 450 451
    
    ns->nics_per_router = codes_mapping_get_lp_count(group, 1,
            lp_type_name, NULL, 1);
    
    ns->msg_id = 0;
    ns->next_available_time = 0;
    ns->node_copy_next_available_time = (tw_stime*)malloc(ns->params->node_copy_queues * sizeof(tw_stime));
    for(int i = 0; i < ns->params->node_copy_queues; i++) {
        ns->node_copy_next_available_time[i] = 0;
    }
452

Nikhil's avatar
Nikhil committed
453 454 455 456 457 458 459 460
    ns->in_sched_send_loop = (int *)malloc(ns->params->num_queues * sizeof(int));
    ns->sched_send = (model_net_sched**)malloc(ns->params->num_queues * sizeof(model_net_sched*));
    for(int i = 0; i < ns->params->num_queues; i++) {
        ns->sched_send[i] = (model_net_sched*)malloc(sizeof(model_net_sched));
        model_net_sched_init(&ns->params->sched_params, 0, method_array[ns->net_id],
                ns->sched_send[i]);
        ns->in_sched_send_loop[i] = 0;
    }
461 462 463
    ns->sched_recv = malloc(sizeof(model_net_sched));
    model_net_sched_init(&ns->params->sched_params, 1, method_array[ns->net_id],
            ns->sched_recv);
464

465
    ns->sub_type = model_net_get_lp_type(ns->net_id);
466

467 468 469 470 471 472
    /* some ROSS instrumentation setup */
    if (g_st_ev_trace || g_st_model_stats)
    {
        ns->sub_model_type = model_net_get_model_stat_type(ns->net_id);
        mn_model_types[ns->net_id].mstat_sz = ns->sub_model_type->mstat_sz;
    }
473

474 475 476 477 478 479
    // NOTE: some models actually expect LP state to be 0 initialized...
    // *cough anything that uses mn_stats_array cough*
    ns->sub_state = calloc(1, ns->sub_type->state_sz);

    // initialize the model-net method
    ns->sub_type->init(ns->sub_state, lp);
480 481 482 483 484 485

    // check validity of sampling function
    event_f  sample  = method_array[ns->net_id]->mn_sample_fn;
    revent_f rsample = method_array[ns->net_id]->mn_sample_rc_fn;
    if (model_net_sampling_enabled()) {
        if (sample == NULL) {
486 487 488
            /* MM: Commented out temporarily--- */
            //tw_error(TW_LOC,
            //        "Sampling requested for a model that doesn't provide it\n");
489 490 491 492
        }
        else if (rsample == NULL &&
                (g_tw_synchronization_protocol == OPTIMISTIC ||
                 g_tw_synchronization_protocol == OPTIMISTIC_DEBUG)) {
493 494 495
            /* MM: Commented out temporarily--- */
            //tw_error(TW_LOC,
            //        "Sampling requested for a model that doesn't provide it\n");
496 497 498 499 500 501 502 503
        }
        else {
            init_f sinit = method_array[ns->net_id]->mn_sample_init_fn;
            if (sinit != NULL)
                sinit(ns->sub_state, lp);
            issue_sample_event(lp);
        }
    }
504 505 506 507 508 509 510
}

void model_net_base_event(
        model_net_base_state * ns,
        tw_bf * b,
        model_net_wrap_msg * m,
        tw_lp * lp){
511 512

    if(m->h.magic != model_net_base_magic)
513
        printf("\n LP ID mismatched %llu ", lp->gid);
514

515
    assert(m->h.magic == model_net_base_magic);
516 517

    void * sub_msg;
518
    switch (m->h.event_type){
519 520 521 522 523 524
        case MN_BASE_NEW_MSG:
            handle_new_msg(ns, b, m, lp);
            break;
        case MN_BASE_SCHED_NEXT:
            handle_sched_next(ns, b, m, lp);
            break;
525 526 527 528 529 530 531
        case MN_BASE_SAMPLE: ;
            event_f sample = method_array[ns->net_id]->mn_sample_fn;
            assert(model_net_sampling_enabled() && sample != NULL);
            sub_msg = ((char*)m)+msg_offsets[ns->net_id];
            sample(ns->sub_state, b, sub_msg, lp);
            issue_sample_event(lp);
            break;
532
        case MN_BASE_PASS: ;
533
            sub_msg = ((char*)m)+msg_offsets[ns->net_id];
534 535 536 537 538 539 540 541 542 543 544 545 546 547
            ns->sub_type->event(ns->sub_state, b, sub_msg, lp);
            break;
        /* ... */
        default:
            assert(!"model_net_base event type not known");
            break;
    }
}

void model_net_base_event_rc(
        model_net_base_state * ns,
        tw_bf * b,
        model_net_wrap_msg * m,
        tw_lp * lp){
548
    assert(m->h.magic == model_net_base_magic);
549 550

    void * sub_msg;
551
    switch (m->h.event_type){
552 553 554 555 556 557
        case MN_BASE_NEW_MSG:
            handle_new_msg_rc(ns, b, m, lp);
            break;
        case MN_BASE_SCHED_NEXT:
            handle_sched_next_rc(ns, b, m, lp);
            break;
558 559 560 561 562 563
        case MN_BASE_SAMPLE: ;
            revent_f sample_rc = method_array[ns->net_id]->mn_sample_rc_fn;
            assert(model_net_sampling_enabled() && sample_rc != NULL);
            sub_msg = ((char*)m)+msg_offsets[ns->net_id];
            sample_rc(ns->sub_state, b, sub_msg, lp);
            break;
564
        case MN_BASE_PASS: ;
565
            sub_msg = ((char*)m)+msg_offsets[ns->net_id];
566 567 568 569 570 571 572 573 574 575 576 577
            ns->sub_type->revent(ns->sub_state, b, sub_msg, lp);
            break;
        /* ... */
        default:
            assert(!"model_net_base event type not known");
            break;
    }
}

void model_net_base_finalize(
        model_net_base_state * ns,
        tw_lp * lp){
578 579 580
    final_f sfini = method_array[ns->net_id]->mn_sample_fini_fn;
    if (sfini != NULL)
        sfini(ns->sub_state, lp);
581 582 583 584
    ns->sub_type->final(ns->sub_state, lp);
    free(ns->sub_state);
}

585
/// bitfields used:
586
/// c31 - we initiated a sched_next event
587 588 589 590 591
void handle_new_msg(
        model_net_base_state * ns,
        tw_bf *b,
        model_net_wrap_msg * m,
        tw_lp * lp){
Nikhil's avatar
Nikhil committed
592 593 594 595 596 597 598 599 600 601 602 603
    static int num_servers = -1;
    static int servers_per_node = -1;
    if(num_servers == -1) {
        char const *sender_group;
        char const *sender_lpname;
        int rep_id, offset;
        model_net_request *r = &m->msg.m_base.req;
        codes_mapping_get_lp_info2(r->src_lp, &sender_group, &sender_lpname, 
                NULL, &rep_id, &offset);
        num_servers = codes_mapping_get_lp_count(sender_group, 1,
                sender_lpname, NULL, 1);
        servers_per_node = num_servers/ns->params->num_queues; //this is for entire switch
604
        if(servers_per_node == 0) servers_per_node = 1;
Nikhil's avatar
Nikhil committed
605
        servers_per_node_queue = num_servers/ns->nics_per_router/ns->params->node_copy_queues;
606
        if(servers_per_node_queue == 0) servers_per_node_queue = 1;
Nikhil's avatar
Nikhil committed
607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623
        if(!g_tw_mynode) {
            fprintf(stdout, "Set num_servers per router %d, servers per "
                "injection queue per router %d, servers per node copy queue "
                "per node %d\n", num_servers, servers_per_node, 
                servers_per_node_queue);
        }
    } 

    if(lp->gid == m->msg.m_base.req.dest_mn_lp) {
        model_net_request *r = &m->msg.m_base.req;
        int rep_id, offset;
        codes_mapping_get_lp_info2(r->src_lp, NULL, NULL, NULL, &rep_id, &offset);
        int queue = offset/ns->nics_per_router/servers_per_node_queue;
        m->msg.m_base.save_ts = ns->node_copy_next_available_time[queue];
        tw_stime exp_time = ((ns->node_copy_next_available_time[queue] 
                            > tw_now(lp)) ? ns->node_copy_next_available_time[queue] : tw_now(lp));
        exp_time += r->msg_size * codes_cn_delay;
Nikhil's avatar
Nikhil committed
624 625
        exp_time -= tw_now(lp);
        tw_stime delay = codes_local_latency(lp);
Nikhil's avatar
Nikhil committed
626 627 628 629 630
        ns->node_copy_next_available_time[queue] = exp_time;
        int remote_event_size = r->remote_event_size;
        int self_event_size = r->self_event_size;
        void *e_msg = (m+1);
        if (remote_event_size > 0) {
Nikhil's avatar
Nikhil committed
631 632
            exp_time += delay;
            tw_event *e = tw_event_new(r->final_dest_lp, exp_time, lp);
Nikhil's avatar
Nikhil committed
633 634 635 636 637
            memcpy(tw_event_data(e), e_msg, remote_event_size);
            tw_event_send(e);
            e_msg = (char*)e_msg + remote_event_size; 
        }
        if (self_event_size > 0) {
Nikhil's avatar
Nikhil committed
638 639
            exp_time += delay;
            tw_event *e = tw_event_new(r->src_lp, exp_time, lp);
Nikhil's avatar
Nikhil committed
640 641 642 643 644 645 646 647 648
            memcpy(tw_event_data(e), e_msg, self_event_size);
            tw_event_send(e);
        }
        return;
    }

    if(m->msg.m_base.isQueueReq) {
        m->msg.m_base.save_ts = ns->next_available_time;
        tw_stime exp_time = ((ns->next_available_time > tw_now(lp)) ? ns->next_available_time : tw_now(lp));
Nikhil's avatar
Nikhil committed
649
        exp_time += ns->params->nic_seq_delay + codes_local_latency(lp);
Nikhil's avatar
Nikhil committed
650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670
        ns->next_available_time = exp_time;
        tw_event *e = tw_event_new(lp->gid, exp_time - tw_now(lp), lp);
        model_net_wrap_msg *m_new = tw_event_data(e);
        memcpy(m_new, m, sizeof(model_net_wrap_msg));
        void *e_msg = (m+1);
        void *e_new_msg = (m_new+1);
        model_net_request *r = &m->msg.m_base.req;
        int remote_event_size = r->remote_event_size;
        int self_event_size = r->self_event_size;
        if (remote_event_size > 0){
            memcpy(e_new_msg, e_msg, remote_event_size);
            e_msg = (char*)e_msg + remote_event_size; 
            e_new_msg = (char*)e_new_msg + remote_event_size; 
        }
        if (self_event_size > 0){
            memcpy(e_new_msg, e_msg, self_event_size);
        }
        m_new->msg.m_base.isQueueReq = 0;
        tw_event_send(e);
        return;
    } 
671
    // simply pass down to the scheduler
672
    model_net_request *r = &m->msg.m_base.req;
673 674
    // don't forget to set packet size, now that we're responsible for it!
    r->packet_size = ns->params->packet_size;
675
    r->msg_id = ns->msg_id++;
676 677
    void * m_data = m+1;
    void *remote = NULL, *local = NULL;
678
    if (r->remote_event_size > 0){
679 680
        remote = m_data;
        m_data = (char*)m_data + r->remote_event_size;
681 682
    }
    if (r->self_event_size > 0){
683
        local = m_data;
684
    }
685

Nikhil's avatar
Nikhil committed
686 687 688 689 690 691 692 693 694 695 696
    int queue_offset = 0;
    if(!m->msg.m_base.is_from_remote && ns->params->num_queues != 1) {
        int rep_id, offset;
        if(num_servers == -1) {
            char const *sender_group;
            char const *sender_lpname;
            codes_mapping_get_lp_info2(r->src_lp, &sender_group, &sender_lpname, 
                    NULL, &rep_id, &offset);
            num_servers = codes_mapping_get_lp_count(sender_group, 1,
                    sender_lpname, NULL, 1);
            servers_per_node = num_servers/ns->params->num_queues;
697
            if(servers_per_node == 0) servers_per_node = 1;
Nikhil's avatar
Nikhil committed
698 699 700 701 702 703 704
        } else {
            codes_mapping_get_lp_info2(r->src_lp, NULL, NULL, NULL, &rep_id, &offset);
        }
        queue_offset = offset/servers_per_node;
    }
    r->queue_offset = queue_offset;

705
    // set message-specific params
706
    int is_from_remote = m->msg.m_base.is_from_remote;
Nikhil's avatar
Nikhil committed
707
    model_net_sched *ss = is_from_remote ? ns->sched_recv : ns->sched_send[queue_offset];
708
    int *in_sched_loop = is_from_remote  ?
Nikhil's avatar
Nikhil committed
709
        &ns->in_sched_recv_loop : &ns->in_sched_send_loop[queue_offset];
710 711
    model_net_sched_add(r, &m->msg.m_base.sched_params, r->remote_event_size,
            remote, r->self_event_size, local, ss, &m->msg.m_base.rc, lp);
712

713
    if (*in_sched_loop == 0){
714 715
        b->c31 = 1;
        /* No need to issue an extra sched-next event if we're currently idle */
716
        *in_sched_loop = 1;
717 718 719 720 721
        /* NOTE: we can do this because the sched rc struct in the event is
         * *very* lightly used (there's harmless overlap in usage for the
         * priority scheduler) */
        handle_sched_next(ns, b, m, lp);
        assert(*in_sched_loop); // we shouldn't have fallen out of the loop
722 723
    }
}
724 725 726

void handle_new_msg_rc(
        model_net_base_state *ns,
727
        tw_bf *b,
728 729
        model_net_wrap_msg *m,
        tw_lp *lp){
Nikhil's avatar
Nikhil committed
730
    if(lp->gid == m->msg.m_base.req.dest_mn_lp) {
Nikhil's avatar
Nikhil committed
731
        codes_local_latency_reverse(lp);
Nikhil's avatar
Nikhil committed
732 733 734 735 736 737 738 739
        model_net_request *r = &m->msg.m_base.req;
        int rep_id, offset;
        codes_mapping_get_lp_info2(r->src_lp, NULL, NULL, NULL, &rep_id, &offset);
        int queue = offset/ns->nics_per_router/servers_per_node_queue;
        ns->node_copy_next_available_time[queue] = m->msg.m_base.save_ts;
        return;
    }
    if(m->msg.m_base.isQueueReq) {
Nikhil's avatar
Nikhil committed
740
        codes_local_latency_reverse(lp);
Nikhil's avatar
Nikhil committed
741 742 743 744
        ns->next_available_time = m->msg.m_base.save_ts;
        return;
    }
    model_net_request *r = &m->msg.m_base.req;
745
    int is_from_remote = m->msg.m_base.is_from_remote;
Nikhil's avatar
Nikhil committed
746
    model_net_sched *ss = is_from_remote ? ns->sched_recv : ns->sched_send[r->queue_offset];
747
    int *in_sched_loop = is_from_remote  ?
Nikhil's avatar
Nikhil committed
748
        &ns->in_sched_recv_loop : &ns->in_sched_send_loop[r->queue_offset];
749

750 751
    if (b->c31) {
        handle_sched_next_rc(ns, b, m, lp);
752
        *in_sched_loop = 0;
753
    }
754
    model_net_sched_add_rc(ss, &m->msg.m_base.rc, lp);
755
}
756 757 758 759

/// bitfields used
/// c0 - scheduler loop is finished
void handle_sched_next(
760 761 762 763
        model_net_base_state * ns,
        tw_bf *b,
        model_net_wrap_msg * m,
        tw_lp * lp){
764
    tw_stime poffset;
Nikhil's avatar
Nikhil committed
765
    model_net_request *r = &m->msg.m_base.req;
766
    int is_from_remote = m->msg.m_base.is_from_remote;
Nikhil's avatar
Nikhil committed
767
    model_net_sched * ss = is_from_remote ? ns->sched_recv : ns->sched_send[r->queue_offset];
768
    int *in_sched_loop = is_from_remote ?
Nikhil's avatar
Nikhil committed
769
        &ns->in_sched_recv_loop : &ns->in_sched_send_loop[r->queue_offset];
770
    int ret = model_net_sched_next(&poffset, ss, m+1, &m->msg.m_base.rc, lp);
771 772 773 774
    // we only need to know whether scheduling is finished or not - if not,
    // go to the 'next iteration' of the loop
    if (ret == -1){
        b->c0 = 1;
775
        *in_sched_loop = 0;
776
    }
777 778 779
    // Currently, only a subset of the network implementations use the
    // callback-based scheduling loop (model_net_method_idle_event).
    // For all others, we need to schedule the next packet
780
    // immediately
Jonathan Jenkins's avatar
Jonathan Jenkins committed
781
    else if (ns->net_id == SIMPLEP2P || ns->net_id == TORUS){
782
        tw_event *e = tw_event_new(lp->gid,
783
                poffset+codes_local_latency(lp), lp);
784
        model_net_wrap_msg *m_wrap = tw_event_data(e);
Nikhil's avatar
Nikhil committed
785
        model_net_request *r_wrap = &m_wrap->msg.m_base.req;
786
        msg_set_header(model_net_base_magic, MN_BASE_SCHED_NEXT, lp->gid,
787 788
                &m_wrap->h);
        m_wrap->msg.m_base.is_from_remote = is_from_remote;
Nikhil's avatar
Nikhil committed
789
        r_wrap->queue_offset = r->queue_offset;
790 791
        // no need to set m_base here
        tw_event_send(e);
792 793
    }
}
794

795 796 797 798 799
void handle_sched_next_rc(
        model_net_base_state * ns,
        tw_bf *b,
        model_net_wrap_msg * m,
        tw_lp * lp){
Nikhil's avatar
Nikhil committed
800
    model_net_request *r = &m->msg.m_base.req;
801
    int is_from_remote = m->msg.m_base.is_from_remote;
Nikhil's avatar
Nikhil committed
802
    model_net_sched * ss = is_from_remote ? ns->sched_recv : ns->sched_send[r->queue_offset];
803
    int *in_sched_loop = is_from_remote ?
Nikhil's avatar
Nikhil committed
804
        &ns->in_sched_recv_loop : &ns->in_sched_send_loop[r->queue_offset];
805

806
    model_net_sched_next_rc(ss, m+1, &m->msg.m_base.rc, lp);
807
    if (b->c0){
808
        *in_sched_loop = 1;
809
    }
Jonathan Jenkins's avatar
Jonathan Jenkins committed
810
    else if (ns->net_id == SIMPLEP2P || ns->net_id == TORUS){
811 812
        codes_local_latency_reverse(lp);
    }
813 814 815 816 817 818 819 820 821 822 823 824 825
}

/**** END IMPLEMENTATIONS ****/

tw_event * model_net_method_event_new(
        tw_lpid dest_gid,
        tw_stime offset_ts,
        tw_lp *sender,
        int net_id,
        void **msg_data,
        void **extra_data){
    tw_event *e = tw_event_new(dest_gid, offset_ts, sender);
    model_net_wrap_msg *m_wrap = tw_event_data(e);
826 827
    msg_set_header(model_net_base_magic, MN_BASE_PASS, sender->gid,
            &m_wrap->h);
828 829 830 831 832 833 834 835
    *msg_data = ((char*)m_wrap)+msg_offsets[net_id];
    // extra_data is optional
    if (extra_data != NULL){
        *extra_data = m_wrap + 1;
    }
    return e;
}

836 837 838 839 840 841 842 843 844 845 846 847
void model_net_method_send_msg_recv_event(
        tw_lpid final_dest_lp,
        tw_lpid dest_mn_lp,
        tw_lpid src_lp, // the "actual" source (as opposed to the model net lp)
        uint64_t msg_size,
        int is_pull,
        uint64_t pull_size,
        int remote_event_size,
        const mn_sched_params *sched_params,
        const char * category,
        int net_id,
        void * msg,
848
        tw_stime offset,
849
        tw_lp *sender){
850
    tw_event *e =
851
        tw_event_new(dest_mn_lp, offset+codes_local_latency(sender), sender);
852 853 854 855 856 857 858 859
    model_net_wrap_msg *m = tw_event_data(e);
    msg_set_header(model_net_base_magic, MN_BASE_NEW_MSG, sender->gid, &m->h);

    if (sched_params != NULL)
        m->msg.m_base.sched_params = *sched_params;
    else
        model_net_sched_set_default_params(&m->msg.m_base.sched_params);

860 861 862
    model_net_request *r = &m->msg.m_base.req;
    r->final_dest_lp = final_dest_lp;
    r->src_lp = src_lp;
863
    // for "recv" events, set the "dest" to this LP in the case of a pull event
864 865 866 867 868 869 870 871 872
    r->dest_mn_lp = sender->gid;
    r->pull_size = pull_size;
    r->msg_size = msg_size;
    // TODO: document why we're setting packet_size this way
    r->packet_size = msg_size;
    r->net_id = net_id;
    r->is_pull = is_pull;
    r->remote_event_size = remote_event_size;
    r->self_event_size = 0;
873 874
    m->msg.m_base.is_from_remote = 1;

875 876
    strncpy(r->category, category, CATEGORY_NAME_MAX-1);
    r->category[CATEGORY_NAME_MAX-1] = '\0';
877 878 879 880 881 882 883 884 885

    if (remote_event_size > 0){
        void * m_dat = model_net_method_get_edata(net_id, msg);
        memcpy(m+1, m_dat, remote_event_size);
    }

    tw_event_send(e);
}

Jonathan Jenkins's avatar
Jonathan Jenkins committed
886 887 888 889
void model_net_method_send_msg_recv_event_rc(tw_lp *sender){
    codes_local_latency_reverse(sender);
}

890 891 892

void model_net_method_idle_event(tw_stime offset_ts, int is_recv_queue,
        tw_lp * lp){
Nikhil's avatar
Nikhil committed
893 894 895 896 897
    model_net_method_idle_event2(offset_ts, is_recv_queue, 0, lp);
}

void model_net_method_idle_event2(tw_stime offset_ts, int is_recv_queue, 
        int queue_offset, tw_lp * lp){
898 899
    tw_event *e = tw_event_new(lp->gid, offset_ts, lp);
    model_net_wrap_msg *m_wrap = tw_event_data(e);
Nikhil's avatar
Nikhil committed
900
    model_net_request *r_wrap = &m_wrap->msg.m_base.req;
901 902
    msg_set_header(model_net_base_magic, MN_BASE_SCHED_NEXT, lp->gid,
            &m_wrap->h);
903
    m_wrap->msg.m_base.is_from_remote = is_recv_queue;
Nikhil's avatar
Nikhil committed
904
    r_wrap->queue_offset = queue_offset;
905 906 907
    tw_event_send(e);
}

908 909 910 911 912 913 914 915 916 917 918 919
void * model_net_method_get_edata(int net_id, void *msg){
    return (char*)msg + sizeof(model_net_wrap_msg) - msg_offsets[net_id];
}

/*
 * Local variables:
 *  c-indent-level: 4
 *  c-basic-offset: 4
 * End:
 *
 * vim: ft=c ts=8 sts=4 sw=4 expandtab
 */