-
Notifications
You must be signed in to change notification settings - Fork 1.3k
Expand file tree
/
Copy pathnet.cc
More file actions
1473 lines (1344 loc) · 66.2 KB
/
net.cc
File metadata and controls
1473 lines (1344 loc) · 66.2 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
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
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
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
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*************************************************************************
* Copyright (c) 2016-2022, NVIDIA CORPORATION. All rights reserved.
*
* See LICENSE.txt for license information
************************************************************************/
#include "comm.h"
#include "net.h"
#include "graph.h"
#include "proxy.h"
#include "collectives.h"
#include "gdrwrap.h"
#include "shmutils.h"
#include "p2p.h"
#include "profiler.h"
#include "transport.h"
#include "shm.h"
static_assert(sizeof(ncclNetHandle_t) <= CONNECT_SIZE, "NET Connect info is too large");
#define NCCL_NET_MAP_HOSTMEM 0
#define NCCL_NET_MAP_DEVMEM 1
#define NCCL_NET_MAP_SHARED_HOSTMEM 2
#define NCCL_NET_MAP_SHARED_DEVMEM 3
#define NCCL_NET_MAP_GDCMEM 4
#define NCCL_NET_MAP_MEMS 5
#define NCCL_NET_MAP_MASK_DEVMEM 0x40000000
#define NCCL_NET_MAP_MASK_SHARED 0x80000000
#define NCCL_NET_MAP_MASK_USED 0x20000000
#define NCCL_NET_MAP_MASK_OFFSET 0x1fffffff
#define NCCL_NET_MAP_OFFSET_BANK(mapStruct, offsetName) \
((mapStruct)->offsets.offsetName >> 30)
#define NCCL_NET_MAP_OFFSET_NULL(mapStruct, offsetName) \
(((mapStruct)->offsets.offsetName >> 29) == 0)
#define NCCL_NET_MAP_GET_POINTER(mapStruct, cpuOrGpu, offsetName) \
(NCCL_NET_MAP_OFFSET_NULL(mapStruct, offsetName) ? NULL : \
(mapStruct)->mems[NCCL_NET_MAP_OFFSET_BANK(mapStruct, offsetName)].cpuOrGpu##Ptr + ((mapStruct)->offsets.offsetName & NCCL_NET_MAP_MASK_OFFSET))
#define NCCL_NET_MAP_DEV_MEM(mapStruct, offsetName) \
(((mapStruct)->offsets.offsetName & NCCL_NET_MAP_MASK_DEVMEM) != 0)
#define NCCL_NET_MAP_ADD_POINTER(mapStruct, shared, dev, memSize, offsetName) do { \
int bank = NCCL_NET_MAP_MASK_USED + (dev)*NCCL_NET_MAP_MASK_DEVMEM + (shared)*NCCL_NET_MAP_MASK_SHARED; \
if ((shared) == 0) { \
if (dev) { \
(mapStruct)->offsets.offsetName = bank + (mapStruct)->mems[NCCL_NET_MAP_DEVMEM].size; \
(mapStruct)->mems[NCCL_NET_MAP_DEVMEM].size += memSize; \
} else { \
(mapStruct)->offsets.offsetName = bank + (mapStruct)->mems[NCCL_NET_MAP_HOSTMEM].size; \
(mapStruct)->mems[NCCL_NET_MAP_HOSTMEM].size += memSize; \
} \
} else { \
(mapStruct)->offsets.offsetName = bank; \
} \
} while (0);
struct connectMapMem{
char* gpuPtr;
char* cpuPtr;
int size;
ncclIpcDesc ipcDesc;
ncclShmIpcDesc_t attachDesc;
ncclShmIpcDesc_t createDesc;
};
struct connectMap {
int sameProcess;
int shared;
int cudaDev;
// First 3 bits of offsets determine the mem bank. 001 is host mem, 011 is dev mem, 101 is shared host mem and 111 is shared dev mem.
struct connectMapMem mems[NCCL_NET_MAP_MEMS];
// Offsets. 3 MSBs indicate mem bank, 111 indicates NULL.
struct {
uint32_t sendMem;
uint32_t recvMem;
uint32_t buffs[NCCL_NUM_PROTOCOLS];
} offsets;
};
struct sendNetResources {
struct connectMap map;
void* netSendComm;
struct ncclSendMem* sendMem;
struct ncclRecvMem* recvMem;
int tpRank;
int tpLocalRank;
int tpRemoteRank;
int netDev;
int useGdr;
int useDmaBuf;
int maxRecvs;
uint64_t* gdcSync;
void* gdrDesc;
int shared;
int channelId;
int connIndex;
char* buffers[NCCL_NUM_PROTOCOLS];
int buffSizes[NCCL_NUM_PROTOCOLS];
void* mhandles[NCCL_NUM_PROTOCOLS];
uint64_t step;
uint64_t llLastCleaning;
int netDeviceVersion;
ncclNetDeviceType netDeviceType;
ncclNetDeviceHandle_t* netDeviceHandle;
};
struct recvNetResources {
struct connectMap map;
void* netListenComm;
void* netRecvComm;
struct ncclSendMem* sendMem;
struct ncclRecvMem* recvMem;
int tpRank;
int tpLocalRank;
int tpRemoteRank;
int tpRemoteProxyRank;
int netDev;
int useGdr;
int useDmaBuf;
int needFlush;
int maxRecvs;
uint64_t* gdcSync;
uint64_t* gdcFlush;
void* gdrDesc;
int shared;
int channelId;
int connIndex;
char* buffers[NCCL_NUM_PROTOCOLS];
int buffSizes[NCCL_NUM_PROTOCOLS];
void* mhandles[NCCL_NUM_PROTOCOLS];
uint64_t step;
uint64_t llLastCleaning;
int netDeviceVersion;
ncclNetDeviceType netDeviceType;
ncclNetDeviceHandle_t* netDeviceHandle;
};
/* Determine if two peers can communicate with NET */
static ncclResult_t canConnect(int* ret, struct ncclComm* comm, struct ncclTopoGraph* graph, struct ncclPeerInfo* info1, struct ncclPeerInfo* info2) {
*ret = 1;
if (info1->hostHash == info2->hostHash) {
// If on the same host, check intra-node net is not disabled.
NCCLCHECK(ncclTopoCheckNet(comm->topo, info1->rank, info2->rank, ret));
}
return ncclSuccess;
}
NCCL_PARAM(NetSharedBuffers, "NET_SHARED_BUFFERS", -2);
NCCL_PARAM(NetSharedComms, "NET_SHARED_COMMS", 1);
struct setupReq {
int tpRank;
int tpLocalRank;
int tpRemoteRank;
int shared;
int netDev;
int useGdr;
int needFlush;
int channelId;
int connIndex;
};
// Forward declaration
static ncclResult_t sendProxyProgress(struct ncclProxyState* proxyState, struct ncclProxyArgs* args);
/* Determine if we will use this transport for this peer and return connect
* information for this peer */
static ncclResult_t sendSetup(struct ncclComm* comm, struct ncclTopoGraph* graph, struct ncclPeerInfo* myInfo, struct ncclPeerInfo* peerInfo, struct ncclConnect* connectInfo, struct ncclConnector* send, int channelId, int connIndex) {
struct setupReq req = { 0 };
send->conn.shared = req.shared = graph || connIndex == 0 ? 0 : ncclParamNetSharedBuffers() != -2 ? ncclParamNetSharedBuffers() : 1;
req.channelId = channelId;
req.connIndex = connIndex;
int proxyRank;
int64_t netId;
NCCLCHECK(ncclTopoGetNetDev(comm, myInfo->rank, graph, channelId, peerInfo->rank, &netId, &req.netDev, &proxyRank));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->busId, netId, 1, &req.useGdr));
send->conn.flags |= req.useGdr ? NCCL_DIRECT_NIC : 0;
NCCLCHECK(ncclProxyConnect(comm, TRANSPORT_NET, 1, proxyRank, &send->proxyConn));
req.tpLocalRank = comm->topParentLocalRanks[comm->localRank];
req.tpRank = comm->topParentRanks[myInfo->rank];
req.tpRemoteRank = comm->topParentRanks[peerInfo->rank];
NCCLCHECK(ncclProxyCallBlocking(comm, &send->proxyConn, ncclProxyMsgSetup, &req, sizeof(req), NULL, 0));
if (proxyRank == myInfo->rank) {
INFO(NCCL_INIT|NCCL_NET,"Channel %02d/%d : %d[%d] -> %d[%d] [send] via NET/%s/%d%s%s", channelId, connIndex, myInfo->rank, myInfo->nvmlDev, peerInfo->rank, peerInfo->nvmlDev, comm->ncclNet->name, req.netDev,
req.useGdr ? "/GDRDMA" : "", req.shared ? "/Shared" : "");
} else {
INFO(NCCL_INIT|NCCL_NET,"Channel %02d/%d : %d[%d] -> %d[%d] [send] via NET/%s/%d(%d)%s%s", channelId, connIndex, myInfo->rank, myInfo->nvmlDev, peerInfo->rank, peerInfo->nvmlDev, comm->ncclNet->name, req.netDev,
proxyRank, req.useGdr ? "/GDRDMA" : "", req.shared ? "/Shared" : "");
}
*((int*)connectInfo) = comm->topParentRanks[proxyRank];
return ncclSuccess;
}
// GDRCOPY support: TAIL_ENABLE When enabled locates the RX proxy tail in CUDA memory
NCCL_PARAM(GdrCopySyncEnable, "GDRCOPY_SYNC_ENABLE", 1);
// GDRCOPY support: FLUSH_ENABLE When enabled uses a PCI-E read to flush GDRDMA buffers
NCCL_PARAM(GdrCopyFlushEnable, "GDRCOPY_FLUSH_ENABLE", 0);
/* Setup recv connector */
static ncclResult_t recvSetup(struct ncclComm* comm, struct ncclTopoGraph* graph, struct ncclPeerInfo* myInfo, struct ncclPeerInfo* peerInfo, struct ncclConnect* connectInfo, struct ncclConnector* recv, int channelId, int connIndex) {
struct setupReq req = { 0 };
recv->conn.shared = req.shared = graph || connIndex == 0 ? 0 : ncclParamNetSharedBuffers() != -2 ? ncclParamNetSharedBuffers() : 1;
req.channelId = channelId;
req.connIndex = connIndex;
// Use myInfo->rank as the receiver uses its own NIC
int proxyRank;
int64_t netId;
NCCLCHECK(ncclTopoGetNetDev(comm, myInfo->rank, graph, channelId, myInfo->rank, &netId, &req.netDev, &proxyRank));
NCCLCHECK(ncclTopoCheckGdr(comm->topo, myInfo->busId, netId, 0, &req.useGdr));
// Determine whether we need to flush the GDR buffer on recv or not
if (req.useGdr) NCCLCHECK(ncclTopoNeedFlush(comm->topo, myInfo->busId, &req.needFlush));
// We don't support PXN on receive yet
NCCLCHECK(ncclProxyConnect(comm, TRANSPORT_NET, 0, myInfo->rank, &recv->proxyConn));
req.tpLocalRank = comm->topParentLocalRanks[comm->localRank];
req.tpRank = comm->topParentRanks[myInfo->rank];
req.tpRemoteRank = comm->topParentRanks[peerInfo->rank];
NCCLCHECK(ncclProxyCallBlocking(comm, &recv->proxyConn, ncclProxyMsgSetup, &req, sizeof(req), connectInfo, sizeof(ncclNetHandle_t)));
INFO(NCCL_INIT|NCCL_NET,"Channel %02d/%d : %d[%d] -> %d[%d] [receive] via NET/%s/%d%s%s", channelId, connIndex, peerInfo->rank, peerInfo->nvmlDev, myInfo->rank, myInfo->nvmlDev, comm->ncclNet->name, req.netDev,
req.useGdr ? "/GDRDMA" : "", req.shared ? "/Shared" : "");
return ncclSuccess;
}
static ncclResult_t netMapShm(struct ncclComm *comm, struct connectMapMem* mem) {
NCCLCHECK(ncclShmImportShareableBuffer(comm, &mem->createDesc, (void**)&mem->cpuPtr, (void**)&mem->gpuPtr, &mem->attachDesc));
return ncclSuccess;
}
static ncclResult_t netCreateShm(struct ncclProxyState* proxyState, struct connectMapMem* mem) {
NCCLCHECK(ncclShmAllocateShareableBuffer(proxyState->tpRank, mem->size, false, &mem->createDesc, (void**)&mem->cpuPtr, (void**)&mem->gpuPtr));
return ncclSuccess;
}
static ncclResult_t netDumpMap(struct connectMap* map) {
printf("Dump map same process %d shared %d\n", map->sameProcess, map->shared);
struct connectMapMem *mem = map->mems+NCCL_NET_MAP_HOSTMEM;
printf("Mem 0: Host mem (%x B) CPU %p GPU %p\n", mem->size, mem->cpuPtr, mem->gpuPtr);
mem = map->mems+NCCL_NET_MAP_DEVMEM;
printf("Mem 1: Vid mem (%x B) CPU %p GPU %p\n", mem->size, mem->cpuPtr, mem->gpuPtr);
mem = map->mems+NCCL_NET_MAP_SHARED_HOSTMEM;
printf("Mem 2: Shared Host mem (%x B) CPU %p GPU %p\n", mem->size, mem->cpuPtr, mem->gpuPtr);
mem = map->mems+NCCL_NET_MAP_SHARED_DEVMEM;
printf("Mem 3: Shared Vid mem (%x B) CPU %p GPU %p\n", mem->size, mem->cpuPtr, mem->gpuPtr);
printf("SendMem -> Used %d Bank %d Offset %x, cpu %p gpu %p\n",
map->offsets.sendMem & NCCL_NET_MAP_MASK_USED ? 1 : 0,
NCCL_NET_MAP_OFFSET_BANK(map, sendMem), map->offsets.sendMem & NCCL_NET_MAP_MASK_OFFSET,
NCCL_NET_MAP_GET_POINTER(map, cpu, sendMem), NCCL_NET_MAP_GET_POINTER(map, gpu, sendMem));
printf("RecvMem -> Used %d Bank %d Offset %x, cpu %p gpu %p\n",
map->offsets.recvMem & NCCL_NET_MAP_MASK_USED ? 1 : 0,
NCCL_NET_MAP_OFFSET_BANK(map, recvMem), map->offsets.recvMem & NCCL_NET_MAP_MASK_OFFSET,
NCCL_NET_MAP_GET_POINTER(map, cpu, recvMem), NCCL_NET_MAP_GET_POINTER(map, gpu, recvMem));
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++) {
printf("Proto %d -> Used %d Bank %d Offset %x, cpu %p, gpu %p\n", p,
map->offsets.buffs[p] & NCCL_NET_MAP_MASK_USED ? 1 : 0,
NCCL_NET_MAP_OFFSET_BANK(map, buffs[p]), map->offsets.buffs[p] & NCCL_NET_MAP_MASK_OFFSET,
NCCL_NET_MAP_GET_POINTER(map, cpu, buffs[p]), NCCL_NET_MAP_GET_POINTER(map, gpu, buffs[p]));
}
printf("End of dump\n");
return ncclSuccess;
}
struct netSendConnectArgs {
ncclNetHandle_t handle;
};
struct netRecvConnectArgs {
int proxyRank;
};
static ncclResult_t sendConnect(struct ncclComm* comm, struct ncclConnect* connectInfo, int nranks, int rank, struct ncclConnector* send) {
struct connectMap* map = (connectMap*) send->transportResources;
void* opId;
// map isn't allocated thus this op hasn't been submitted yet
if (!map) {
// Setup device pointers
NCCLCHECK(ncclCalloc(&map, 1));
send->transportResources = map;
opId = send;
INFO(NCCL_PROXY, "sendConnect ncclProxyCallAsync opId=%p", opId);
netSendConnectArgs args = {0};
memcpy(&args.handle, connectInfo, sizeof(ncclNetHandle_t));
NCCLCHECK(ncclProxyCallAsync(comm, &send->proxyConn, ncclProxyMsgConnect, &args, sizeof(netSendConnectArgs), sizeof(struct connectMap), opId));
} else {
opId = send;
}
ncclResult_t ret;
ret = ncclPollProxyResponse(comm, &send->proxyConn, map, opId);
if (ret != ncclSuccess) {
if (ret != ncclInProgress) {
free(map);
send->transportResources = NULL;
}
return ret;
}
INFO(NCCL_PROXY, "sendConnect ncclPollProxyResponse opId=%p", opId);
if (map->sameProcess && !ncclCuMemEnable()) {
if (map->cudaDev != comm->cudaDev) {
// Enable P2P access for Legacy IPC
cudaError_t err = cudaDeviceEnablePeerAccess(map->cudaDev, 0);
if (err == cudaErrorPeerAccessAlreadyEnabled) {
cudaGetLastError();
} else if (err != cudaSuccess) {
WARN("failed to peer with device %d: %d %s", map->cudaDev, err, cudaGetErrorString(err));
return ncclInternalError;
}
}
} else if (!(map->sameProcess && map->cudaDev == comm->cudaDev)) {
if (!map->sameProcess) NCCLCHECK(netMapShm(comm, map->mems + NCCL_NET_MAP_HOSTMEM));
if (map->mems[NCCL_NET_MAP_DEVMEM].size) {
map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr = NULL;
NCCLCHECK(ncclP2pImportShareableBuffer(comm, send->proxyConn.rank,
map->mems[NCCL_NET_MAP_DEVMEM].size,
&map->mems[NCCL_NET_MAP_DEVMEM].ipcDesc,
(void**)&map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr));
map->mems[NCCL_NET_MAP_DEVMEM].cpuPtr = NULL;
}
if (map->mems[NCCL_NET_MAP_SHARED_DEVMEM].size) {
void** sharedDevMemPtr = comm->proxyState->sharedDevMems + send->proxyConn.tpLocalRank;
if (*sharedDevMemPtr == NULL) {
map->mems[NCCL_NET_MAP_SHARED_DEVMEM].gpuPtr = NULL;
NCCLCHECK(ncclP2pImportShareableBuffer(comm, send->proxyConn.rank,
map->mems[NCCL_NET_MAP_SHARED_DEVMEM].size,
&map->mems[NCCL_NET_MAP_SHARED_DEVMEM].ipcDesc,
sharedDevMemPtr));
}
map->mems[NCCL_NET_MAP_SHARED_DEVMEM].gpuPtr = (char*)(*sharedDevMemPtr);
map->mems[NCCL_NET_MAP_SHARED_DEVMEM].cpuPtr = NULL;
}
}
//NCCLCHECK(netDumpMap(map));
struct ncclSendMem *sendMem = (struct ncclSendMem*) NCCL_NET_MAP_GET_POINTER(map, gpu, sendMem);
void* gdcMem = map->mems[NCCL_NET_MAP_GDCMEM].gpuPtr;
send->conn.head = gdcMem ? (uint64_t*)gdcMem : &sendMem->head;
struct ncclRecvMem *recvMem = (struct ncclRecvMem*) NCCL_NET_MAP_GET_POINTER(map, gpu, recvMem);
send->conn.tail = &recvMem->tail;
send->conn.stepSize = comm->buffSizes[NCCL_PROTO_SIMPLE]/NCCL_STEPS;
send->conn.connFifo = recvMem->connFifo;
// Only fuse P2P buffers, continue to allocate dedicated buffers for ring/tree
for (int i=0; i<NCCL_STEPS; i++) {
send->conn.connFifo[i].offset = -1;
recvMem->connFifo[i].mode = map->shared ? NCCL_MODE_OFFSET : NCCL_MODE_NORMAL;
}
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++)
send->conn.buffs[p] = NCCL_NET_MAP_GET_POINTER(map, gpu, buffs[p]);
if (send->proxyConn.sameProcess) {
if (send->proxyConn.connection->netDeviceHandle) {
send->conn.netDeviceHandle = *send->proxyConn.connection->netDeviceHandle;
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++)
send->conn.mhandles[p] = send->proxyConn.connection->mhandles[p];
}
if (send->proxyConn.connection->needsProxyProgress) {
send->proxyConn.proxyProgress = sendProxyProgress;
} else {
send->proxyConn.proxyProgress = NULL;
}
} else {
send->proxyConn.proxyProgress = sendProxyProgress;
}
return ncclSuccess;
}
// Forward declare
static ncclResult_t recvProxyProgress(struct ncclProxyState* proxyState, struct ncclProxyArgs* args);
/* Connect to this peer */
static ncclResult_t recvConnect(struct ncclComm* comm, struct ncclConnect* connectInfo, int nranks, int rank, struct ncclConnector* recv) {
struct connectMap* map = (connectMap*) recv->transportResources;
void* opId;
if (!map) {
NCCLCHECK(ncclCalloc(&map, 1));
recv->transportResources = map;
// Use recv connector as unique identifier
opId = recv;
INFO(NCCL_PROXY, "recvConnect ncclProxyCallAsync opId=%p &recv->proxyConn=%p connectInfo=%p",
opId, &recv->proxyConn, connectInfo);
netRecvConnectArgs args = {0};
args.proxyRank = *((int*)connectInfo);
NCCLCHECK(ncclProxyCallAsync(comm, &recv->proxyConn, ncclProxyMsgConnect, &args, sizeof(netRecvConnectArgs), sizeof(struct connectMap), opId));
} else {
opId = recv;
}
ncclResult_t ret;
NCCLCHECK(ret = ncclPollProxyResponse(comm, &recv->proxyConn, map, opId));
if (ret != ncclSuccess) {
if (ret != ncclInProgress) {
free(map);
recv->transportResources = NULL;
}
return ret;
}
INFO(NCCL_PROXY, "recvConnect ncclPollProxyResponse opId=%p", opId);
//NCCLCHECK(netDumpMap(map));
struct ncclSendMem *sendMem = (struct ncclSendMem*) NCCL_NET_MAP_GET_POINTER(map, gpu, sendMem);
recv->conn.head = &sendMem->head;
struct ncclRecvMem *recvMem = (struct ncclRecvMem*) NCCL_NET_MAP_GET_POINTER(map, gpu, recvMem);
void* gdcMem = map->mems[NCCL_NET_MAP_GDCMEM].gpuPtr;
recv->conn.tail = gdcMem ? (uint64_t*)gdcMem : &recvMem->tail;
recv->conn.stepSize = comm->buffSizes[NCCL_PROTO_SIMPLE]/NCCL_STEPS;
recv->conn.connFifo = recvMem->connFifo;
// Only fuse P2P buffers, continue to allocate dedicated buffers for ring/tree
for (int i=0; i<NCCL_STEPS; i++) {
recvMem->connFifo[i].mode = map->shared ? NCCL_MODE_OFFSET : NCCL_MODE_NORMAL;
}
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++)
recv->conn.buffs[p] = NCCL_NET_MAP_GET_POINTER(map, gpu, buffs[p]);
if (recv->proxyConn.sameProcess) {
if (recv->proxyConn.connection->netDeviceHandle) {
recv->conn.netDeviceHandle = *recv->proxyConn.connection->netDeviceHandle;
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++)
recv->conn.mhandles[p] = recv->proxyConn.connection->mhandles[p];
}
if (recv->proxyConn.connection->needsProxyProgress) {
recv->proxyConn.proxyProgress = recvProxyProgress;
} else {
recv->proxyConn.proxyProgress = NULL;
}
} else {
recv->proxyConn.proxyProgress = recvProxyProgress;
}
return ncclSuccess;
}
static ncclResult_t sendFree(struct ncclConnector* send) {
struct connectMap* map = (struct connectMap*)(send->transportResources);
if (map) {
int cudaDev;
CUDACHECK(cudaGetDevice(&cudaDev));
if (map->cudaDev != cudaDev && map->mems[NCCL_NET_MAP_DEVMEM].size) {
if (ncclCuMemEnable()) {
// cuMem API support
NCCLCHECK(ncclP2pFreeShareableBuffer(&map->mems[NCCL_NET_MAP_DEVMEM].ipcDesc));
NCCLCHECK(ncclCuMemFree(map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr));
} else {
// Legacy CUDA IPC support
CUDACHECK(cudaIpcCloseMemHandle(map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr));
}
}
if (!map->sameProcess) {
NCCLCHECK(ncclShmIpcClose(&map->mems[NCCL_NET_MAP_HOSTMEM].attachDesc));
}
free(map);
}
return ncclSuccess;
}
static ncclResult_t recvFree(struct ncclConnector* recv) {
if (recv->transportResources) free(recv->transportResources);
return ncclSuccess;
}
#define NCCL_SHARED_STEPS 16
static ncclResult_t sharedNetBuffersInit(struct ncclProxyState* proxyState, int cuda, int tpLocalRank, int type, int sameProcess,
int nChannels, char** gpuPtr, char** cpuPtr, int* size, ncclIpcDesc *ipcDesc) {
if (cuda == 0 && sameProcess == 0) {
WARN("PXN should not use host buffers for data");
return ncclInternalError;
}
struct ncclProxyProgressState* progressState = &proxyState->progressState;
if (progressState->localPeers == NULL) {
NCCLCHECK(ncclCalloc(&progressState->localPeers, proxyState->tpLocalnRanks));
}
struct ncclProxyPeer** localPeers = progressState->localPeers;
if (localPeers[tpLocalRank] == NULL) {
NCCLCHECK(ncclCalloc(localPeers + tpLocalRank, 1));
}
struct ncclProxyPeer* peer = localPeers[tpLocalRank];
struct ncclProxySharedP2p* state = type == 0 ? &peer->send : &peer->recv;
state->refcount++;
if (state->size == 0) {
state->size = nChannels * NCCL_SHARED_STEPS * proxyState->p2pChunkSize;
}
if (size) *size = state->size;
if (cuda && state->cudaBuff == NULL) {
if (sameProcess == 0 || ncclCuMemEnable()) {
NCCLCHECK(ncclP2pAllocateShareableBuffer(state->size, 0, &state->ipcDesc, (void**)&state->cudaBuff));
} else {
NCCLCHECK(ncclCudaCalloc(&state->cudaBuff, state->size));
}
}
if (!cuda && state->hostBuff == NULL) {
NCCLCHECK(ncclCudaHostCalloc(&state->hostBuff, state->size));
}
if (cpuPtr) *cpuPtr = cuda ? state->cudaBuff : state->hostBuff;
if (gpuPtr) *gpuPtr = (cpuPtr && sameProcess) ? *cpuPtr : NULL;
if (ipcDesc) memcpy(ipcDesc, &state->ipcDesc, sizeof(state->ipcDesc));
return ncclSuccess;
}
static ncclResult_t sharedBuffersGet(struct ncclProxyState* proxyState, int channel, int slot, int* offset, int* size) {
// Use different pools for different channels and also separate send/recv.
int globalSlot = (channel*NCCL_SHARED_STEPS)+slot;
*offset = proxyState->p2pChunkSize * globalSlot;
if (size) *size = proxyState->p2pChunkSize;
return ncclSuccess;
}
static ncclResult_t sharedNetBuffersDestroy(struct ncclProxyState* proxyState, int tpLocalRank, int type, struct ncclProxyConnection* connection) {
if (proxyState->progressState.localPeers == NULL) NCCLCHECK(ncclInternalError);
struct ncclProxyPeer* peer = proxyState->progressState.localPeers[tpLocalRank];
if (peer == NULL) NCCLCHECK(ncclInternalError);
struct ncclProxySharedP2p* state = type == 0 ? &peer->send : &peer->recv;
if (state->size == 0) NCCLCHECK(ncclInternalError);
if (ncclAtomicRefCountDecrement(&state->refcount) == 0) {
if (state->cudaBuff) {
if (!connection->sameProcess || ncclCuMemEnable()) {
NCCLCHECK(ncclP2pFreeShareableBuffer(&state->ipcDesc));
}
NCCLCHECK(ncclCudaFree(state->cudaBuff));
}
if (state->hostBuff) NCCLCHECK(ncclCudaHostFree(state->hostBuff));
}
if (peer->send.refcount || peer->recv.refcount) return ncclSuccess;
free(peer);
proxyState->progressState.localPeers[tpLocalRank] = NULL;
for (int r = 0; r < proxyState->tpLocalnRanks; r++) {
if (proxyState->progressState.localPeers[r]) return ncclSuccess;
}
// All peers are freed, free array
free(proxyState->progressState.localPeers);
proxyState->progressState.localPeers = NULL;
return ncclSuccess;
}
static ncclResult_t proxySharedInit(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, int nChannels) {
NCCLCHECK(sharedNetBuffersInit(proxyState, 1, connection->tpLocalRank, 0, connection->sameProcess, nChannels, NULL, NULL, NULL, NULL));
return ncclSuccess;
}
static ncclResult_t sendProxySetup(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, void* reqBuff, int reqSize, void* respBuff, int respSize, int* done) {
struct setupReq* req = (struct setupReq*) reqBuff;
if (reqSize != sizeof(struct setupReq)) return ncclInternalError;
struct sendNetResources* resources;
NCCLCHECK(ncclCalloc(&resources, 1));
connection->transportResources = resources;
resources->tpRank = req->tpRank;
resources->tpLocalRank = req->tpLocalRank;
resources->tpRemoteRank = req->tpRemoteRank;
resources->netDev = req->netDev;
resources->shared = connection->shared = req->shared;
resources->useGdr = req->useGdr;
resources->channelId = req->channelId;
resources->connIndex = req->connIndex;
ncclNetProperties_t props;
NCCLCHECK(proxyState->ncclNet->getProperties(req->netDev, &props));
/* DMA-BUF support */
resources->useDmaBuf = resources->useGdr && proxyState->dmaBufSupport && (props.ptrSupport & NCCL_PTR_DMABUF);
resources->maxRecvs = props.maxRecvs;
resources->netDeviceVersion = props.netDeviceVersion;
resources->netDeviceType = props.netDeviceType;
resources->netDeviceVersion = props.netDeviceVersion;
resources->netDeviceType = props.netDeviceType;
// We don't return any data
if (respSize != 0) return ncclInternalError;
*done = 1;
return ncclSuccess;
}
static ncclResult_t recvProxySetup(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, void* reqBuff, int reqSize, void* respBuff, int respSize, int* done) {
struct setupReq* req = (struct setupReq*) reqBuff;
if (reqSize != sizeof(struct setupReq)) return ncclInternalError;
struct recvNetResources* resources;
NCCLCHECK(ncclCalloc(&resources, 1));
connection->transportResources = resources;
resources->tpRank = req->tpRank;
resources->tpLocalRank = req->tpLocalRank;
resources->tpRemoteRank = req->tpRemoteRank;
resources->netDev = req->netDev;
resources->shared = connection->shared = req->shared;
resources->useGdr = req->useGdr;
resources->needFlush = req->needFlush;
resources->channelId = req->channelId;
resources->connIndex = req->connIndex;
ncclNetProperties_t props;
NCCLCHECK(proxyState->ncclNet->getProperties(req->netDev, &props));
/* DMA-BUF support */
resources->useDmaBuf = resources->useGdr && proxyState->dmaBufSupport && (props.ptrSupport & NCCL_PTR_DMABUF);
resources->maxRecvs = props.maxRecvs;
resources->netDeviceVersion = props.netDeviceVersion;
resources->netDeviceType = props.netDeviceType;
if (respSize != sizeof(ncclNetHandle_t)) return ncclInternalError;
NCCLCHECK(proxyState->ncclNet->listen(req->netDev, respBuff, &resources->netListenComm));
*done = 1;
return ncclSuccess;
}
// This function embeds plugin-specific rules given the current versions
static ncclResult_t ncclNetGetDeviceHandle(ncclNetDeviceType type, int version, bool isRecv, ncclNetDeviceHandle_t** handle) {
bool needsDeviceHandle = false;
if (type == NCCL_NET_DEVICE_UNPACK) {
if (version == NCCL_NET_DEVICE_UNPACK_VERSION && isRecv) {
needsDeviceHandle = true;
}
}
// Don't re-alloc netDeviceHandles
if (needsDeviceHandle && (*handle == NULL)) {
NCCLCHECK(ncclCalloc(handle, 1));
(*handle)->netDeviceType = type;
(*handle)->netDeviceVersion = version;
} else if (!needsDeviceHandle) {
*handle = NULL;
}
return ncclSuccess;
}
static ncclResult_t sendProxyConnect(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, void* reqBuff, int reqSize, void* respBuff, int respSize, int* done) {
struct sendNetResources* resources = (struct sendNetResources*)(connection->transportResources);
if (reqSize != sizeof(netSendConnectArgs)) return ncclInternalError;
ncclResult_t ret = ncclSuccess;
netSendConnectArgs* req = (netSendConnectArgs*) reqBuff;
NCCLCHECK(ncclNetGetDeviceHandle(resources->netDeviceType, resources->netDeviceVersion, false /*isRecv*/, &resources->netDeviceHandle));
if (resources->shared) {
// Shared buffers
struct ncclProxyProgressState* progressState = &proxyState->progressState;
if (progressState->localPeers == NULL) {
NCCLCHECK(ncclCalloc(&progressState->localPeers, proxyState->tpLocalnRanks));
}
struct ncclProxyPeer** localPeers = progressState->localPeers;
if (localPeers[resources->tpLocalRank] == NULL) {
NCCLCHECK(ncclCalloc(localPeers + resources->tpLocalRank, 1));
}
connection->proxyAppendPtr = localPeers[resources->tpLocalRank]->send.proxyAppend + resources->channelId;
if (resources->maxRecvs > 1 && ncclParamNetSharedComms()) {
// Connect or reuse connection for a netdev/remote rank.
if (progressState->netComms[resources->netDev] == NULL) {
NCCLCHECK(ncclCalloc(progressState->netComms + resources->netDev, proxyState->tpnRanks));
}
struct ncclSharedNetComms* comms = progressState->netComms[resources->netDev] + resources->tpRemoteRank;
if (comms->sendComm[resources->channelId] == NULL) ret = proxyState->ncclNet->connect(resources->netDev, req->handle, comms->sendComm + resources->channelId, &resources->netDeviceHandle);
resources->netSendComm = comms->sendComm[resources->channelId];
if (comms->sendComm[resources->channelId]) comms->sendRefCount[resources->channelId]++;
} else {
ret = proxyState->ncclNet->connect(resources->netDev, req->handle, &resources->netSendComm, &resources->netDeviceHandle);
}
} else {
// Connect to remote peer
ret = proxyState->ncclNet->connect(resources->netDev, req->handle, &resources->netSendComm, &resources->netDeviceHandle);
connection->proxyAppendPtr = &connection->proxyAppend;
}
NCCLCHECK(ret);
if (resources->netSendComm == NULL) {
*done = 0;
return ncclInProgress;
}
*done = 1;
if (resources->netDeviceHandle) {
connection->netDeviceHandle = resources->netDeviceHandle;
connection->needsProxyProgress = connection->netDeviceHandle->needsProxyProgress;
} else {
connection->needsProxyProgress = 1;
}
// Create structures
struct connectMap* map = &resources->map;
map->sameProcess = connection->sameProcess;
map->shared = resources->shared;
CUDACHECK(cudaGetDevice(&map->cudaDev));
if (resources->shared == 0) { // Only allocate dedicated buffers for ring/tree, not for p2p
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++) {
NCCL_NET_MAP_ADD_POINTER(map, 0, p!= NCCL_PROTO_LL && resources->useGdr, proxyState->buffSizes[p], buffs[p]);
resources->buffSizes[p] = proxyState->buffSizes[p];
}
} else {
// Get shared buffers
int bank = resources->useGdr ? NCCL_NET_MAP_SHARED_DEVMEM : NCCL_NET_MAP_SHARED_HOSTMEM;
struct connectMapMem* mapMem = map->mems+bank;
NCCLCHECK(sharedNetBuffersInit(
proxyState, resources->useGdr, resources->tpLocalRank, 0, map->sameProcess, proxyState->p2pnChannels,
&mapMem->gpuPtr, &mapMem->cpuPtr, &mapMem->size, &mapMem->ipcDesc));
resources->buffSizes[NCCL_PROTO_SIMPLE] = mapMem->size;
if (proxyState->allocP2pNetLLBuffers) {
NCCL_NET_MAP_ADD_POINTER(map, 0, 0 /*p == NCCL_PROTO_LL*/, proxyState->buffSizes[NCCL_PROTO_LL], buffs[NCCL_PROTO_LL]);
resources->buffSizes[NCCL_PROTO_LL] = proxyState->buffSizes[NCCL_PROTO_LL];
}
NCCL_NET_MAP_ADD_POINTER(map, 1, resources->useGdr, mapMem->size, buffs[NCCL_PROTO_SIMPLE]);
}
NCCL_NET_MAP_ADD_POINTER(map, 0, 0, sizeof(struct ncclSendMem), sendMem);
NCCL_NET_MAP_ADD_POINTER(map, 0, 0, sizeof(struct ncclRecvMem), recvMem);
if (map->mems[NCCL_NET_MAP_DEVMEM].size) {
if (resources->shared == 0) {
if (!map->sameProcess || ncclCuMemEnable()) {
ALIGN_SIZE(map->mems[NCCL_NET_MAP_DEVMEM].size, CUDA_IPC_MIN);
NCCLCHECK(ncclP2pAllocateShareableBuffer(map->mems[NCCL_NET_MAP_DEVMEM].size, 0, &map->mems[NCCL_NET_MAP_DEVMEM].ipcDesc,
(void**)&map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr));
} else {
NCCLCHECK(ncclCudaCalloc(&map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr, map->mems[NCCL_NET_MAP_DEVMEM].size));
}
map->mems[NCCL_NET_MAP_DEVMEM].cpuPtr = map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr;
}
}
if (map->sameProcess) {
NCCLCHECK(ncclCudaHostCalloc(&map->mems[NCCL_NET_MAP_HOSTMEM].cpuPtr, map->mems[NCCL_NET_MAP_HOSTMEM].size));
map->mems[NCCL_NET_MAP_HOSTMEM].gpuPtr = map->mems[NCCL_NET_MAP_HOSTMEM].cpuPtr;
} else {
NCCLCHECK(netCreateShm(proxyState, map->mems+NCCL_NET_MAP_HOSTMEM));
void* sendMem = (void*)NCCL_NET_MAP_GET_POINTER(map, cpu, sendMem);
void* recvMem = (void*)NCCL_NET_MAP_GET_POINTER(map, cpu, recvMem);
memset(sendMem, 0, sizeof(struct ncclSendMem));
memset(recvMem, 0, sizeof(struct ncclRecvMem));
}
if (ncclGdrCopy && map->sameProcess && ncclParamGdrCopySyncEnable()) {
uint64_t *cpuPtr, *gpuPtr;
NCCLCHECK(ncclGdrCudaCalloc(&cpuPtr, &gpuPtr, 1, &resources->gdrDesc));
resources->gdcSync = cpuPtr;
struct connectMapMem* gdcMem = map->mems+NCCL_NET_MAP_GDCMEM;
gdcMem->cpuPtr = (char*)cpuPtr;
gdcMem->gpuPtr = (char*)gpuPtr;
gdcMem->size = sizeof(uint64_t); // sendMem->head
}
resources->sendMem = (struct ncclSendMem*) NCCL_NET_MAP_GET_POINTER(map, cpu, sendMem);
resources->recvMem = (struct ncclRecvMem*) NCCL_NET_MAP_GET_POINTER(map, cpu, recvMem);
// Don't give credits yet in shared mode.
(resources->gdcSync ? *resources->gdcSync : resources->sendMem->head) =
(map->shared ? -NCCL_STEPS : 0);
for (int i=0; i<NCCL_STEPS; i++) resources->recvMem->connFifo[i].size = -1;
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++) {
resources->buffers[p] = NCCL_NET_MAP_GET_POINTER(map, cpu, buffs[p]);
if (resources->buffers[p]) {
#if CUDA_VERSION >= 11070
/* DMA-BUF support */
int type = NCCL_NET_MAP_DEV_MEM(map, buffs[p]) ? NCCL_PTR_CUDA : NCCL_PTR_HOST;
if (type == NCCL_PTR_CUDA && resources->useDmaBuf) {
int dmabuf_fd;
CUCHECK(cuMemGetHandleForAddressRange((void *)&dmabuf_fd, (CUdeviceptr)resources->buffers[p], resources->buffSizes[p], CU_MEM_RANGE_HANDLE_TYPE_DMA_BUF_FD, 0));
NCCLCHECK(proxyState->ncclNet->regMrDmaBuf(resources->netSendComm, resources->buffers[p], resources->buffSizes[p], type, 0ULL, dmabuf_fd, &resources->mhandles[p]));
(void)close(dmabuf_fd);
} else // FALL-THROUGH to nv_peermem GDR path
#endif
{
NCCLCHECK(proxyState->ncclNet->regMr(resources->netSendComm, resources->buffers[p], resources->buffSizes[p], NCCL_NET_MAP_DEV_MEM(map, buffs[p]) ? NCCL_PTR_CUDA : NCCL_PTR_HOST, &resources->mhandles[p]));
}
// Copy the mhandle dptr, if implemented
if (resources->netDeviceHandle && proxyState->ncclNet->getDeviceMr)
NCCLCHECK(proxyState->ncclNet->getDeviceMr(resources->netSendComm, resources->mhandles[p], &connection->mhandles[p]));
}
}
//NCCLCHECK(netDumpMap(map));
if (respSize != sizeof(struct connectMap)) return ncclInternalError;
memcpy(respBuff, map, sizeof(struct connectMap));
return ncclSuccess;
}
static ncclResult_t recvProxyConnect(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, void* reqBuff, int reqSize, void* respBuff, int respSize, int* done) {
if (reqSize != sizeof(netRecvConnectArgs)) return ncclInternalError;
struct recvNetResources* resources = (struct recvNetResources*)(connection->transportResources);
netRecvConnectArgs* req = (netRecvConnectArgs*) reqBuff;
resources->tpRemoteProxyRank = req->proxyRank;
ncclResult_t ret = ncclSuccess;
NCCLCHECK(ncclNetGetDeviceHandle(resources->netDeviceType, resources->netDeviceVersion, true /*isRecv*/, &resources->netDeviceHandle));
// Finish connection establishment from remote peer
if (resources->shared) {
// Shared buffers
struct ncclProxyProgressState* progressState = &proxyState->progressState;
if (progressState->localPeers == NULL) {
NCCLCHECK(ncclCalloc(&progressState->localPeers, proxyState->tpLocalnRanks));
}
struct ncclProxyPeer** localPeers = progressState->localPeers;
if (localPeers[resources->tpLocalRank] == NULL) {
NCCLCHECK(ncclCalloc(localPeers + resources->tpLocalRank, 1));
}
connection->proxyAppendPtr = localPeers[resources->tpLocalRank]->recv.proxyAppend + resources->channelId;
if (resources->maxRecvs > 1 && ncclParamNetSharedComms()) {
// Connect or reuse connection for a netdev/remote rank.
if (progressState->netComms[resources->netDev] == NULL) {
NCCLCHECK(ncclCalloc(progressState->netComms + resources->netDev, proxyState->tpnRanks));
}
struct ncclSharedNetComms* comms = progressState->netComms[resources->netDev] + resources->tpRemoteProxyRank;
if (comms->recvComm[resources->channelId] == NULL) ret = proxyState->ncclNet->accept(resources->netListenComm, comms->recvComm+resources->channelId, &resources->netDeviceHandle);
resources->netRecvComm = comms->recvComm[resources->channelId];
if (comms->recvComm[resources->channelId]) comms->recvRefCount[resources->channelId]++;
} else {
ret = proxyState->ncclNet->accept(resources->netListenComm, &resources->netRecvComm, &resources->netDeviceHandle);
}
} else {
// Connect to remote peer
ret = proxyState->ncclNet->accept(resources->netListenComm, &resources->netRecvComm, &resources->netDeviceHandle);
connection->proxyAppendPtr = &connection->proxyAppend;
}
NCCLCHECK(ret);
if (resources->netRecvComm == NULL) {
*done = 0;
return ncclInProgress;
}
*done = 1;
if (resources->netDeviceHandle) {
connection->netDeviceHandle = resources->netDeviceHandle;
connection->needsProxyProgress = connection->netDeviceHandle->needsProxyProgress;
} else {
connection->needsProxyProgress = 1;
}
NCCLCHECK(proxyState->ncclNet->closeListen(resources->netListenComm));
// Create structures
struct connectMap* map = &resources->map;
map->sameProcess = connection->sameProcess;
if (map->sameProcess == 0) return ncclInternalError; // We don't support remote proxy for recv
map->shared = resources->shared;
if (resources->shared == 0) { // Only allocate dedicated buffers for ring/tree, not for p2p
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++) {
NCCL_NET_MAP_ADD_POINTER(map, 0, resources->useGdr, proxyState->buffSizes[p], buffs[p]);
resources->buffSizes[p] = proxyState->buffSizes[p];
}
} else {
// Get shared buffers
int bank = resources->useGdr ? NCCL_NET_MAP_SHARED_DEVMEM : NCCL_NET_MAP_SHARED_HOSTMEM;
struct connectMapMem* mapMem = map->mems+bank;
NCCLCHECK(sharedNetBuffersInit(
proxyState, resources->useGdr, resources->tpLocalRank, 1, 1, proxyState->p2pnChannels,
&mapMem->gpuPtr, &mapMem->cpuPtr, &mapMem->size, NULL));
resources->buffSizes[NCCL_PROTO_SIMPLE] = mapMem->size;
NCCL_NET_MAP_ADD_POINTER(map, 1, resources->useGdr, mapMem->size, buffs[NCCL_PROTO_SIMPLE]);
}
NCCL_NET_MAP_ADD_POINTER(map, 0, 0, sizeof(struct ncclSendMem), sendMem);
NCCL_NET_MAP_ADD_POINTER(map, 0, 0, sizeof(struct ncclRecvMem), recvMem);
if (proxyState->allocP2pNetLLBuffers) {
NCCL_NET_MAP_ADD_POINTER(map, 0, 0 /*resources->useGdr*/, proxyState->buffSizes[NCCL_PROTO_LL], buffs[NCCL_PROTO_LL]);
resources->buffSizes[NCCL_PROTO_LL] = proxyState->buffSizes[NCCL_PROTO_LL];
}
if (map->mems[NCCL_NET_MAP_DEVMEM].size) {
if (resources->shared == 0) {
if (ncclCuMemEnable()) {
NCCLCHECK(ncclP2pAllocateShareableBuffer(map->mems[NCCL_NET_MAP_DEVMEM].size, 0, &map->mems[NCCL_NET_MAP_DEVMEM].ipcDesc,
(void**)&map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr));
} else {
NCCLCHECK(ncclCudaCalloc(&map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr, map->mems[NCCL_NET_MAP_DEVMEM].size));
}
map->mems[NCCL_NET_MAP_DEVMEM].cpuPtr = map->mems[NCCL_NET_MAP_DEVMEM].gpuPtr;
}
}
NCCLCHECK(ncclCudaHostCalloc(&map->mems[NCCL_NET_MAP_HOSTMEM].cpuPtr, map->mems[NCCL_NET_MAP_HOSTMEM].size));
map->mems[NCCL_NET_MAP_HOSTMEM].gpuPtr = map->mems[NCCL_NET_MAP_HOSTMEM].cpuPtr;
if (ncclGdrCopy && map->sameProcess) {
uint64_t *cpuPtr, *gpuPtr;
NCCLCHECK(ncclGdrCudaCalloc(&cpuPtr, &gpuPtr, 2, &resources->gdrDesc));
if (ncclParamGdrCopySyncEnable()) {
resources->gdcSync = cpuPtr;
struct connectMapMem* gdcMem = map->mems+NCCL_NET_MAP_GDCMEM;
gdcMem->cpuPtr = (char*)cpuPtr;
gdcMem->gpuPtr = (char*)gpuPtr;
gdcMem->size = sizeof(uint64_t);
}
if (ncclParamGdrCopyFlushEnable()) resources->gdcFlush = cpuPtr + 1;
}
resources->sendMem = (struct ncclSendMem*) NCCL_NET_MAP_GET_POINTER(map, cpu, sendMem);
resources->recvMem = (struct ncclRecvMem*) NCCL_NET_MAP_GET_POINTER(map, cpu, recvMem);
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++) {
resources->buffers[p] = NCCL_NET_MAP_GET_POINTER(map, cpu, buffs[p]);
if (resources->buffers[p]) {
#if CUDA_VERSION >= 11070
/* DMA-BUF support */
int type = NCCL_NET_MAP_DEV_MEM(map, buffs[p]) ? NCCL_PTR_CUDA : NCCL_PTR_HOST;
if (type == NCCL_PTR_CUDA && resources->useDmaBuf) {
int dmabuf_fd;
CUCHECK(cuMemGetHandleForAddressRange((void *)&dmabuf_fd, (CUdeviceptr)resources->buffers[p], resources->buffSizes[p], CU_MEM_RANGE_HANDLE_TYPE_DMA_BUF_FD, 0));
NCCLCHECK(proxyState->ncclNet->regMrDmaBuf(resources->netRecvComm, resources->buffers[p], resources->buffSizes[p], type, 0ULL, dmabuf_fd, &resources->mhandles[p]));
(void)close(dmabuf_fd);
} else // FALL-THROUGH to nv_peermem GDR path
#endif
{
NCCLCHECK(proxyState->ncclNet->regMr(resources->netRecvComm, resources->buffers[p], resources->buffSizes[p], NCCL_NET_MAP_DEV_MEM(map, buffs[p]) ? NCCL_PTR_CUDA : NCCL_PTR_HOST, &resources->mhandles[p]));
}
// Copy the mhandle dptr
if (resources->netDeviceType != NCCL_NET_DEVICE_HOST && proxyState->ncclNet->getDeviceMr)
NCCLCHECK(proxyState->ncclNet->getDeviceMr(resources->netRecvComm, resources->mhandles[p], &connection->mhandles[p]));
}
}
//NCCLCHECK(netDumpMap(map));
if (respSize != sizeof(struct connectMap)) return ncclInternalError;
memcpy(respBuff, map, sizeof(struct connectMap));
return ncclSuccess;
}
static ncclResult_t sendProxyFree(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState) {
struct sendNetResources* resources = (struct sendNetResources*)(connection->transportResources);
if (connection->state == connSharedInitialized) { // NVB Preconnect
NCCLCHECK(sharedNetBuffersDestroy(proxyState, connection->tpLocalRank, 0, connection));
return ncclSuccess;
}
if (connection->state == connConnected) {
for (int p=0; p<NCCL_NUM_PROTOCOLS; p++) {
if (resources->buffers[p]) {
NCCLCHECK(proxyState->ncclNet->deregMr(resources->netSendComm, resources->mhandles[p]));
}
}
struct connectMapMem* mems = resources->map.mems;
if (resources->map.sameProcess) {
NCCLCHECK(ncclCudaHostFree(mems[NCCL_NET_MAP_HOSTMEM].cpuPtr));
} else {
NCCLCHECK(ncclShmIpcClose(&mems[NCCL_NET_MAP_HOSTMEM].createDesc));
}
NCCLCHECK(ncclCudaFree(mems[NCCL_NET_MAP_DEVMEM].cpuPtr));
if (!resources->map.sameProcess || ncclCuMemEnable()) {
// cuMem API support
if (mems[NCCL_NET_MAP_DEVMEM].size) {
NCCLCHECK(ncclP2pFreeShareableBuffer(&mems[NCCL_NET_MAP_DEVMEM].ipcDesc));
}
}
if (mems[NCCL_NET_MAP_GDCMEM].cpuPtr) NCCLCHECK(ncclGdrCudaFree(resources->gdrDesc));
if (resources->shared) {
NCCLCHECK(sharedNetBuffersDestroy(proxyState, resources->tpLocalRank, 0, connection));
if (resources->maxRecvs > 1 && ncclParamNetSharedComms()) {
struct ncclSharedNetComms* comms = proxyState->progressState.netComms[resources->netDev]+resources->tpRemoteRank;
comms->sendRefCount[resources->channelId]--;
if (comms->sendRefCount[resources->channelId] == 0) NCCLCHECK(proxyState->ncclNet->closeSend(comms->sendComm[resources->channelId]));
} else {
NCCLCHECK(proxyState->ncclNet->closeSend(resources->netSendComm));
}
} else {
NCCLCHECK(proxyState->ncclNet->closeSend(resources->netSendComm));
}
}
if (resources) free(resources);
return ncclSuccess;
}
static ncclResult_t recvProxyFree(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState) {
struct recvNetResources* resources = (struct recvNetResources*)(connection->transportResources);
if (connection->state == connSharedInitialized) { // NVB Preconnect
NCCLCHECK(sharedNetBuffersDestroy(proxyState, connection->tpLocalRank, 1, connection));
return ncclSuccess;
}
if (connection->state == connConnected) {