AI/rocm-systems
2
0
Форкнуть 0
Код Задачи Запросы на слияние Действия Пакеты Проекты Релизы Вики Активность
Files
3ea7eedf3b9b94f1d9f99f4e55536dfcbd23c1ca
rocm-systems/src/include/transport.h
T
Kamil Iskra 72d2432094 NCCL 2.27.3-1 
Symmetric memory API and symmetric kernels
 * Redesign from the ground up, enabling major latency and bandwidth
   improvements.
 * Add new API calls to register user-allocated memory among communicator
   ranks into a NCCL window: ncclCommWindowRegister() and
   ncclCommWindowDeregister(). The calls currently support symmetric
   registration for P2P and NVLS, and require VMM memory buffers (i.e.,
   CUMEM must be operational).
 * Implement specialized kernels taking advantage of symmetrically
   registered memory, with performance gains expected particularly for
   small to medium message sizes.
 * The kernels support 32 bit floating point types and smaller, and sum as
   the reduction operator, with no more than one collective operation per
   group.
 * Floating point summation is always done in fp32 accumulators (with the
   exception of fp8 on NVLS, where it uses fp16 inside the switch). Thus,
   the accuracy with fp8 and fp16 data types should be much improved.
 * This initial implementation supports non-network communicators only (P2P
   and NVLS transports).
 * To explore this functionality users need to use the new memory
   registration API calls with the NCCL_WIN_COLL_SYMMETRIC flag and all
   ranks of a communicator must pass buffers at the same offset in the same
   registration when invoking a collective NCCL operation.

Add support for DGX Spark.

Add support for DirectNIC (CX8) to the internal IB plugin.

Add a new ncclCommShrink() API call
 * It is a non-collective call similar to ncclCommSplit(), which makes it
   possible to exclude some (possibly unresponsive) ranks from the parent
   communicator.

Add support for loading multiple network plugins
 * This enables the creation of generic containers that can work across a
   range of providers.
 * Allow NCCL_NET_PLUGIN to accept a comma-separated list of plugins to
   load.

NVLink SHARP (NVLS) improvements
 * Implement NVLS+IB SHARP support for AllGather and ReduceScatter with
   user buffer registration. This improves performance and reduces the
   number of CTAs needed to achieve peak bandwidth.
 * Gracefully fall back by default to other transports if NVLS
   initialization fails (the old behavior of returning an error code from a
   NCCL call can be preserved by setting NCCL_NVLS_ENABLE=1).
 * Decrease the NVLS channel count to 24 on Blackwell systems with multiple
   NVLink domains per communicator.
 * Enable fine-tuning of NCCL behavior per communicator using new
   "ncclConfig_t" members "collnetEnable", "CTAPolicy", and "nvlsCTAs".

Profiler improvements
 * Extend the init function by adding communicator name, comm id (hash),
   rank, number of ranks, number of nodes, and the NCCL log function to the
   argument list. This makes the name and the comm id available to all
   events in the communicator without explicitly passing them to each
   individual event. Add the communicator id and rank to the profiler trace
   filename. Now, the communicator name can be set via a new "ncclConfig_t"
   member "commName".
 * Improve the accuracy of the GPU kernel events by providing GPU-generated
   timestamps for the start and stop of every NCCL operation.
 * Harmonize proxy events, removing overlaps between ProxyOp and ProxyStep
   states.
 * Add support for network-defined event updates (through
   "recordEventState").
 * Report the correct number of channels used by every collective/p2p
   operation (used to be set to nMaxChannels for collectives and absent for
   p2ps).
 * Fix the logic on proxyCtrl Idle/Active events (Issue #1162).
 * Fix an issue where the network proxy profiler could lose track of an
   event identifier (Issue #1682).
 * Improve the backward compatibility with plugins older than v4.
 * Ensure that the work counters are 0-initialized.
 * Fix a potential race condition in the network profiler that could result
   in an event being linked to a wrong parent.

MNNVL improvements
 * Increase to 16 the number of NICs used to communicate between MNNVL
   domains on GB200 systems, to optimize the performance of collective
   operations.
 * Add support for more complex MNNVL topologies with up to 32 NICs per
   node.
 * If the MNNVL fabric initialization was unsuccessful, NCCL will now fail
   by default, so as to avoid inadvertently falling back to a potentially
   much slower network transport. Such failures are typically due to a
   misconfigured IMEX support on the system. To continue without MNNVL,
   restart the job with NCCL_MNNVL_ENABLE=0.
 * Fix a potential hang in alltoall-like communication patterns at a scale
   of over 80 ranks.
 * Make NCCL_P2P_DISABLE=1 imply NCCL_MNNVL_ENABLE=0 (so the latter no
   longer needs to be specified on MNNVL systems).
 * Fix an initialization failure when NCCL_TOPO_FILE is used on MNNVL
   systems.
 * Fix the graph search to exclude non-local NICs.
 * Fix the SHM transport to use fabric handles on MNNVL systems.

NIC Fusion improvements
 * Disable the creation of fused NICs for physical devices that haven't
   been merged.
 * Flatten multiple ports to a single PCI device within the internal IB
   plugin and reparent dual-port NICs under the first PCI parent. If the
   parent is not a PCI switch, PCI devices for fused NICs won't be
   duplicated.
 * Route traffic on GB200-CX8 systems through DirectNIC, not the host
   interface.

Improve support for platforms with C2C connectivity (e.g., GB200)
 * Enable GPUDirect RDMA for the NICs by default.
 * Add support for P2C (PXN over C2C) and the LL128 protocol.

Extend NCCL fault tolerance in multithreaded scenarios
 * Support the creation of multiple nonblocking communicators within a
   single group and polling in parallel for the completion using multiple
   threads (one per communicator).

Enable ncclImplicitOrderLaunch for CUDA 12.9+
 * This can potentially speed up NCCL_IMPLICIT_LAUNCH_ORDER.

Improve the netSocket transport latency and control
 * Provide finer control over the size of the socket send/receive buffers,
   the task size, and the number of sockets that a single peer can open.
 * Add support for the inlining of small messages behind the header when
   using multiple sockets per connection.

Improve the readability of the CPU affinity in the debug output
 * Print it as a range string rather than a bitmask.

Fix a potential race condition in graph execution
 * A contention could arise when mixing graph and non-graph execution.

Improve PXN connection code
 * Avoid duplicate and unused connections.

RAS fixes
 * Fix a memory corruption at job termination time in case of a previously
   failed initialization of a RAS socket connection.
 * Fix a race condition leading to a crash when generating a RAS report
   during communicator initialization (Issues #1669, #1718).
 * Fix a potential race condition when gathering data for a RAS status
   report.

Fix a potential memory corruption in ncclCommSplit()
 * Memory could get corrupted when resource sharing was in use and the size
   of the NVLink domain in the new communicator was smaller than in the old
   one.

Fix asynchronous graph upload
 * Fix a small memory leak.
 * Fix oversychronization.

Add a check for out-of-memory conditions in ncclMemAlloc()

Clean up the NCCL socket code
 * accept() will retry also if just reading the magic failed (Issue #1613).
 * connect() will retry also if poll() did not return a POLLOUT event
   (Issue #1618).
 * Add error checking in a few instances (Issue #1539).
 * Fix the loop condition in ncclFindInterfaceMatchSubnet() (Issue #1574).
 * Clean up the debug output, downgrading WARN messages to INFO in
   non-critical cases, and printing the peer's address where relevant.

Switch NCCL_DEBUG_FILE to line buffering
 * This should help avoid mixed-up partial output lines in multithreaded
   cases.

Other minor fixes
 * Improve the checks for buffer overflows in the graph code (Issue #1585).
 * Extend logging and state clearing to all four events in the internal IB
   plugin (Issue #1650).
 * Fix the error path in case IB communication is not ready (Issue #1489).
 * Add ECE logging for IB fabric.
 * Fix various minor issues in the graph module (Issue #1635).
 * Clean up the debug output in the graph code, downgrading WARN messages
   to INFO in non-critical cases.
 * Add a missing argument to a directSend() call (Issue #1628).
 * Remove duplicate code in sendProxySetup() (Issue #1420).
 * Fix the order of arguments of cudaDeviceCanAccessPeer() (Issue #1507).
 * Fix compiler warnings with GCC 14.
 * Fix a typo in a comment (Issue #1236).
2025-05-29 20:56:40 -07:00

175 строки
9.2 KiB
C
Исходник Ответственный История

/*************************************************************************
* Copyright (c) 2016-2022, NVIDIA CORPORATION. All rights reserved.
*
* See LICENSE.txt for license information
************************************************************************/
#ifndef NCCL_TRANSPORT_H_
#define NCCL_TRANSPORT_H_
#include "device.h"
#include "graph.h"
#include "nvmlwrap.h"
#include "core.h"
#define NTRANSPORTS 4
#define TRANSPORT_UNDEFINED -1
#define TRANSPORT_P2P 0
#define TRANSPORT_SHM 1
#define TRANSPORT_NET 2
#define TRANSPORT_COLLNET 3
#define TRANSPORT_PROFILER 4
#include "proxy.h"
#include "comm.h"
#include "bootstrap.h"
extern struct ncclTransport p2pTransport;
extern struct ncclTransport shmTransport;
extern struct ncclTransport netTransport;
extern struct ncclTransport collNetTransport;
extern struct ncclTransport profilerTransport;
extern struct ncclTransport* ncclTransports[];
// Forward declarations
struct ncclRing;
struct ncclConnector;
struct ncclComm;
struct ncclPeerInfo {
int rank;
int cudaDev;
int nvmlDev;
int gdrSupport;
uint64_t hostHash;
uint64_t pidHash;
dev_t shmDev;
int64_t busId;
struct ncclComm* comm;
int cudaCompCap;
size_t totalGlobalMem;
// MNNVL support
nvmlGpuFabricInfoV_t fabricInfo;
int cuMemSupport;
int version;
};
#define CONNECT_SIZE 256
#define NCCL_MAX_PAGE_SIZE (512L * 1024L * 1024L)
#define NCCL_REC_PAGE_SIZE (2L * 1024L * 1024L)
struct ncclConnect {
char data[CONNECT_SIZE];
};
#if CUDART_VERSION >= 12010
#define NVLS_HANDLE_SIZE 64
struct ncclNvlsSharedRes {
int refCount;
bool inited;
CUmulticastObjectProp bufProp;
CUmulticastObjectProp signalProp;
CUmemAccessDesc accessDesc;
int dev;
size_t creditUCSize;
size_t creditMCSize;
size_t buffUCSize;
size_t buffMCSize;
CUmemGenericAllocationHandle mcBuffHandle; // Multicast handle for NVLS buffer
CUmemGenericAllocationHandle mcCreditHandle; // Multicast handle for NVLS credit buffer
char* mcBuff; // Multicast NVLS buffer address
char* mcCredit; // Multicast NVLS credit address
CUmemGenericAllocationHandle ucBuffHandle; // Unicast Handle for NVLS buffer
CUmemGenericAllocationHandle ucCreditHandle; // Unicast Handle for NVLS credit buffer
char* ucBuff; // Unicast NVLS buffer address
char* ucCredit; // Unicast NVLS credit address
int nChannels;
int nHeads;
struct ncclShmemCollBuff nvlsShmem;
void *nvlsShmemHandle;
};
#endif /* CUDART_VERSION >= 12010 */
struct ncclCollNetSharedRes {
int refCount;
int size;
char* cudaBuff;
char* hostBuff;
struct ncclProxyArgs* proxyAppend[2*NCCL_MAX_NETDEVS];
void* resources;
int nChannels;
size_t buffSize;
};
struct ncclTransportComm {
ncclResult_t (*setup)(struct ncclComm* comm, struct ncclTopoGraph* graph, struct ncclPeerInfo*, struct ncclPeerInfo*, struct ncclConnect*, struct ncclConnector*, int channelId, int connIndex);
ncclResult_t (*connect)(struct ncclComm* comm, struct ncclConnect*, int nranks, int rank, struct ncclConnector*);
ncclResult_t (*free)(struct ncclConnector*);
ncclResult_t (*proxySharedInit)(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, int nChannels);
ncclResult_t (*proxySetup)(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, void* reqBuff, int reqSize, void* respBuff, int respSize, int* done);
ncclResult_t (*proxyConnect)(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, void* reqBuff, int reqSize, void* respBuff, int respSize, int* done);
ncclResult_t (*proxyFree)(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState);
ncclResult_t (*proxyProgress)(struct ncclProxyState* proxyState, struct ncclProxyArgs*);
ncclResult_t (*proxyRegister)(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, void* reqBuff, int reqSize, void* respBuff, int respSize, int* done);
ncclResult_t (*proxyDeregister)(struct ncclProxyConnection* connection, struct ncclProxyState* proxyState, void* reqBuff, int reqSize, int* done);
};
struct ncclTransport {
const char name[8];
ncclResult_t (*canConnect)(int*, struct ncclComm* comm, struct ncclTopoGraph* graph, struct ncclPeerInfo*, struct ncclPeerInfo*);
struct ncclTransportComm send;
struct ncclTransportComm recv;
};
ncclResult_t ncclTransportP2pConnect(struct ncclComm* comm, int channelId, int nrecv, int* peerRecv, int nsend, int* peerSend, int connIndex);
ncclResult_t ncclTransportP2pSetup(struct ncclComm* comm, struct ncclTopoGraph* graph, int connIndex);
ncclResult_t ncclTransportCheckP2pType(struct ncclComm* comm, bool* isAllDirectP2p, bool* directMode);
ncclResult_t ncclTransportIsAllDirectP2p(struct ncclComm* comm, int* isAllDirectP2p);
ncclResult_t ncclNvlsInit(struct ncclComm* comm);
ncclResult_t ncclNvlsSetup(struct ncclComm* comm, struct ncclComm* parent);
ncclResult_t ncclNvlsBufferSetup(struct ncclComm* comm);
ncclResult_t ncclNvlsTreeConnect(struct ncclComm* comm);
ncclResult_t ncclNvlsGraphRegisterBuffer(struct ncclComm *comm, const void *sendbuff, void *recvbuff, size_t sendbuffSize, size_t recvbuffSize, int *outRegBufUsed, void **outRegBufSend, void **outRegBufRecv, struct ncclIntruQueue<struct ncclCommCallback, &ncclCommCallback::next>* cleanupQueue, int* nCleanupQueueElts);
ncclResult_t ncclNvlsLocalRegisterBuffer(struct ncclComm *comm, const void *sendbuff, void *recvbuff, size_t sendbuffSize, size_t recvbuffSize, int *outRegBufUsed, void **outRegBufSend, void **outRegBufRecv);
ncclResult_t ncclNvlsDeregBuffer(struct ncclComm* comm, CUmemGenericAllocationHandle *mcHandler, CUdeviceptr ptr, int dev, size_t ucsize, size_t mcsize);
ncclResult_t ncclNvlsFree(struct ncclComm* comm);
enum { collNetRecv=0, collNetSend=1 };
bool ncclTransportCollNetSetup(struct ncclComm* comm, struct ncclTopoGraph* collNetGraph, struct ncclChannel* channel, int masterRank, int masterPeer, int collNetGraphChannelId, int type, ncclConnect* connect);
ncclResult_t ncclTransportCollNetCheck(struct ncclComm* comm, int collNetSetupFail);
ncclResult_t ncclTransportCollNetFree(struct ncclComm* comm);
ncclResult_t ncclCollnetLocalRegisterBuffer(struct ncclComm* comm, const void* userbuff, size_t buffSize, int type, int* outRegBufUsed, void** outHandle);
ncclResult_t ncclCollnetGraphRegisterBuffer(struct ncclComm* comm, const void* userbuff, size_t buffSize, int type, int* outRegBufFlag, void** outHandle, struct ncclIntruQueue<struct ncclCommCallback, &ncclCommCallback::next>* cleanupQueue, int* nCleanupQueueElts);
ncclResult_t ncclCollnetDeregBuffer(struct ncclComm* comm, struct ncclProxyConnector* proxyconn, void* handle);
ncclResult_t ncclTransportRingConnect(struct ncclComm* comm);
ncclResult_t ncclTransportTreeConnect(struct ncclComm* comm);
ncclResult_t ncclTransportPatConnect(struct ncclComm* comm);
ncclResult_t ncclCollNetSetup(ncclComm_t comm, ncclComm_t parent, struct ncclTopoGraph* graphs[]);
ncclResult_t ncclCollNetChainBufferSetup(ncclComm_t comm);
ncclResult_t ncclCollNetDirectBufferSetup(ncclComm_t comm);
ncclResult_t ncclNetDeregBuffer(struct ncclComm* comm, struct ncclProxyConnector* proxyConn, void* handle);
ncclResult_t ncclNetLocalRegisterBuffer(ncclComm* comm, const void* userbuff, size_t buffSize, struct ncclConnector** peerConns, int nPeers, int* outRegBufFlag, void** outHandle);
ncclResult_t ncclNetGraphRegisterBuffer(ncclComm* comm, const void* userbuff, size_t buffSize, struct ncclConnector** peerConns, int nPeers, int* outRegBufFlag, void** outHandle, struct ncclIntruQueue<struct ncclCommCallback, &ncclCommCallback::next>* cleanupQueue, int* nCleanupQueueElts);
ncclResult_t ncclRegisterP2pIpcBuffer(struct ncclComm* comm, void* userbuff, size_t size, int peerRank, int* regFlag, void** regAddr, struct ncclIntruQueue<struct ncclCommCallback, &ncclCommCallback::next>* cleanupQueue);
ncclResult_t ncclRegisterP2pNetBuffer(struct ncclComm* comm, void* userbuff, size_t size, struct ncclConnector* conn, int* regFlag, void** handle, struct ncclIntruQueue<struct ncclCommCallback, &ncclCommCallback::next>* cleanupQueue);
ncclResult_t ncclRegisterCollBuffers(struct ncclComm* comm, struct ncclTaskColl* info, void* outRegBufSend[NCCL_MAX_LOCAL_RANKS], void* outRegBufRecv[NCCL_MAX_LOCAL_RANKS], struct ncclIntruQueue<struct ncclCommCallback, &ncclCommCallback::next>* cleanupQueue, bool* regNeedConnect);
ncclResult_t ncclRegisterCollNvlsBuffers(struct ncclComm* comm, struct ncclTaskColl* info, void* outRegBufSend[NCCL_MAX_LOCAL_RANKS], void* outRegBufRecv[NCCL_MAX_LOCAL_RANKS], struct ncclIntruQueue<struct ncclCommCallback, &ncclCommCallback::next>* cleanupQueue, bool* regNeedConnect);
ncclResult_t ncclNvlsRegResourcesQuery(struct ncclComm* comm, struct ncclTaskColl* info, int* recChannels);
ncclResult_t ncclIpcSymmetricInit(struct ncclComm* comm);
ncclResult_t ncclIpcSymmetricMap(struct ncclComm* comm, size_t offset, size_t size, CUmemGenericAllocationHandle memHandle, void** symPtr);
ncclResult_t ncclIpcSymmetricFree(struct ncclComm* comm, size_t size, void* symPtr);
ncclResult_t ncclIpcSymmetricFinalize(struct ncclComm* comm);
ncclResult_t ncclNvlsSymmetricInit(struct ncclComm* comm);
ncclResult_t ncclNvlsSymmetricMap(struct ncclComm* comm, size_t offset, size_t ucsize, void* ucaddr);
ncclResult_t ncclNvlsSymmetricFree(struct ncclComm* comm, size_t ucsize, void* ucaddr);
ncclResult_t ncclNvlsSymmetricFinalize(struct ncclComm* comm);
#endif
Ссылка в новой задаче Показать git blame Копировать постоянную ссылку