nvidia-nccl-cu12 2.23.4__py3-none-manylinux2014_aarch64.whl

nvidia/nccl/include/nccl.h

@@ -0,0 +1,472 @@
+/*************************************************************************
+ * Copyright (c) 2015-2021, NVIDIA CORPORATION. All rights reserved.
+ *
+ * See LICENSE.txt for license information
+ ************************************************************************/
+
+#ifndef NCCL_H_
+#define NCCL_H_
+
+#include <cuda_runtime.h>
+#include <cuda_fp16.h>
+#if CUDART_VERSION >= 11000
+#include <cuda_bf16.h>
+#endif
+
+#define NCCL_MAJOR 2
+#define NCCL_MINOR 23
+#define NCCL_PATCH 4
+#define NCCL_SUFFIX ""
+
+#define NCCL_VERSION_CODE 22304
+#define NCCL_VERSION(X,Y,Z) (((X) <= 2 && (Y) <= 8) ? (X) * 1000 + (Y) * 100 + (Z) : (X) * 10000 + (Y) * 100 + (Z))
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <limits.h>
+/* Opaque handle to communicator */
+typedef struct ncclComm* ncclComm_t;
+#define NCCL_COMM_NULL NULL
+
+#define NCCL_UNIQUE_ID_BYTES 128
+typedef struct { char internal[NCCL_UNIQUE_ID_BYTES]; } ncclUniqueId;
+
+/* Error type */
+typedef enum { ncclSuccess                 =  0,
+               ncclUnhandledCudaError      =  1,
+               ncclSystemError             =  2,
+               ncclInternalError           =  3,
+               ncclInvalidArgument         =  4,
+               ncclInvalidUsage            =  5,
+               ncclRemoteError             =  6,
+               ncclInProgress              =  7,
+               ncclNumResults              =  8 } ncclResult_t;
+
+#define NCCL_CONFIG_UNDEF_INT INT_MIN
+#define NCCL_CONFIG_UNDEF_PTR NULL
+#define NCCL_SPLIT_NOCOLOR -1
+#define NCCL_UNDEF_FLOAT -1.0f
+
+/* Communicator configuration. Users can assign value to attributes to specify the
+ * behavior of a communicator. */
+typedef struct ncclConfig_v21700 {
+  /* attributes that users should never touch. */
+  size_t size;
+  unsigned int magic;
+  unsigned int version;
+  /* attributes that users are able to customize. */
+  int blocking;
+  int cgaClusterSize;
+  int minCTAs;
+  int maxCTAs;
+  const char *netName;
+  int splitShare;
+} ncclConfig_t;
+
+/* Config initializer must be assigned to initialize config structure when it is created.
+ * Not initialized config will result in NCCL error. */
+#define NCCL_CONFIG_INITIALIZER {                                       \
+  sizeof(ncclConfig_t), /* size */                                      \
+  0xcafebeef,           /* magic */                                     \
+  NCCL_VERSION(NCCL_MAJOR, NCCL_MINOR, NCCL_PATCH), /* version */       \
+  NCCL_CONFIG_UNDEF_INT,                    /* blocking */              \
+  NCCL_CONFIG_UNDEF_INT,                    /* cgaClusterSize */        \
+  NCCL_CONFIG_UNDEF_INT,                    /* minCTAs */               \
+  NCCL_CONFIG_UNDEF_INT,                    /* maxCTAs */               \
+  NCCL_CONFIG_UNDEF_PTR,                    /* netName */               \
+  NCCL_CONFIG_UNDEF_INT                     /* splitShare */            \
+}
+
+/* This struct will be used by ncclGroupSimulateEnd() API to query information about simulation. */
+typedef struct ncclSimInfo_v22200 {
+    size_t size;
+    unsigned int magic;
+    unsigned int version;
+    float estimatedTime;
+} ncclSimInfo_t;
+
+/* NCCL_SIM_INFO_INITIALIZER must be assigned to initialize simInfo structure when it is created.
+ * Not initialized simInfo will result in NCCL error. */
+#define NCCL_SIM_INFO_INITIALIZER {                                         \
+  sizeof(ncclSimInfo_t),                            /* size */              \
+  0x74685283,                                       /* magic */             \
+  NCCL_VERSION(NCCL_MAJOR, NCCL_MINOR, NCCL_PATCH), /* version */           \
+  NCCL_UNDEF_FLOAT                                  /* estimated time */    \
+}
+
+/* NCCL malloc and free function for all types of NCCL optimizations
+ * (e.g. user buffer registration). The actual allocated size might
+ * be larger than requested due to granularity requirement. */
+ncclResult_t  ncclMemAlloc(void** ptr, size_t size);
+ncclResult_t pncclMemAlloc(void** ptr, size_t size);
+
+ncclResult_t  ncclMemFree(void *ptr);
+ncclResult_t pncclMemFree(void *ptr);
+
+/* Return the NCCL_VERSION_CODE of the NCCL library in the supplied integer.
+ * This integer is coded with the MAJOR, MINOR and PATCH level of the
+ * NCCL library
+ */
+ncclResult_t  ncclGetVersion(int *version);
+ncclResult_t pncclGetVersion(int *version);
+
+/* Generates an Id to be used in ncclCommInitRank. ncclGetUniqueId should be
+ * called once and the Id should be distributed to all ranks in the
+ * communicator before calling ncclCommInitRank. */
+ncclResult_t  ncclGetUniqueId(ncclUniqueId* uniqueId);
+ncclResult_t pncclGetUniqueId(ncclUniqueId* uniqueId);
+
+/* Create a new communicator (multi thread/process version) with a configuration
+ * set by users. */
+ncclResult_t  ncclCommInitRankConfig(ncclComm_t* comm, int nranks, ncclUniqueId commId, int rank, ncclConfig_t* config);
+ncclResult_t pncclCommInitRankConfig(ncclComm_t* comm, int nranks, ncclUniqueId commId, int rank, ncclConfig_t* config);
+
+/* Creates a new communicator (multi thread/process version).
+ * rank must be between 0 and nranks-1 and unique within a communicator clique.
+ * Each rank is associated to a CUDA device, which has to be set before calling
+ * ncclCommInitRank.
+ * ncclCommInitRank implicitly syncronizes with other ranks, so it must be
+ * called by different threads/processes or use ncclGroupStart/ncclGroupEnd. */
+ncclResult_t  ncclCommInitRank(ncclComm_t* comm, int nranks, ncclUniqueId commId, int rank);
+ncclResult_t pncclCommInitRank(ncclComm_t* comm, int nranks, ncclUniqueId commId, int rank);
+
+/* Creates a clique of communicators (single process version).
+ * This is a convenience function to create a single-process communicator clique.
+ * Returns an array of ndev newly initialized communicators in comm.
+ * comm should be pre-allocated with size at least ndev*sizeof(ncclComm_t).
+ * If devlist is NULL, the first ndev CUDA devices are used.
+ * Order of devlist defines user-order of processors within the communicator. */
+ncclResult_t  ncclCommInitAll(ncclComm_t* comm, int ndev, const int* devlist);
+ncclResult_t pncclCommInitAll(ncclComm_t* comm, int ndev, const int* devlist);
+
+/* Finalize a communicator. ncclCommFinalize flushes all issued communications,
+ * and marks communicator state as ncclInProgress. The state will change to ncclSuccess
+ * when the communicator is globally quiescent and related resources are freed; then,
+ * calling ncclCommDestroy can locally free the rest of the resources (e.g. communicator
+ * itself) without blocking. */
+ncclResult_t  ncclCommFinalize(ncclComm_t comm);
+ncclResult_t pncclCommFinalize(ncclComm_t comm);
+
+/* Frees local resources associated with communicator object. */
+ncclResult_t  ncclCommDestroy(ncclComm_t comm);
+ncclResult_t pncclCommDestroy(ncclComm_t comm);
+
+/* Frees resources associated with communicator object and aborts any operations
+ * that might still be running on the device. */
+ncclResult_t  ncclCommAbort(ncclComm_t comm);
+ncclResult_t pncclCommAbort(ncclComm_t comm);
+
+/* Creates one or more communicators from an existing one.
+ * Ranks with the same color will end up in the same communicator.
+ * Within the new communicator, key will be used to order ranks.
+ * NCCL_SPLIT_NOCOLOR as color will indicate the rank will not be part of any group
+ * and will therefore return a NULL communicator.
+ * If config is NULL, the new communicator will inherit the original communicator's
+ * configuration*/
+ncclResult_t  ncclCommSplit(ncclComm_t comm, int color, int key, ncclComm_t *newcomm, ncclConfig_t* config);
+ncclResult_t pncclCommSplit(ncclComm_t comm, int color, int key, ncclComm_t *newcomm, ncclConfig_t* config);
+
+/* Creates a new communicator (multi thread/process version), similar to ncclCommInitRankConfig.
+ * Allows to use more than one ncclUniqueId (up to one per rank), indicated by nId, to accelerate the init operation.
+ * The number of ncclUniqueIds and their order must be the same for every rank.
+ */
+ncclResult_t ncclCommInitRankScalable(ncclComm_t* newcomm, int nranks, int myrank, int nId, ncclUniqueId* commIds, ncclConfig_t* config);
+ncclResult_t pncclCommInitRankScalable(ncclComm_t* newcomm, int nranks, int myrank, int nId, ncclUniqueId* commIds, ncclConfig_t* config);
+
+/* Returns a string for each error code. */
+const char*  ncclGetErrorString(ncclResult_t result);
+const char* pncclGetErrorString(ncclResult_t result);
+
+/* Returns a human-readable message of the last error that occurred. */
+const char*  ncclGetLastError(ncclComm_t comm);
+const char* pncclGetLastError(ncclComm_t comm);
+
+/* Checks whether the comm has encountered any asynchronous errors */
+ncclResult_t  ncclCommGetAsyncError(ncclComm_t comm, ncclResult_t *asyncError);
+ncclResult_t pncclCommGetAsyncError(ncclComm_t comm, ncclResult_t *asyncError);
+
+/* Gets the number of ranks in the communicator clique. */
+ncclResult_t  ncclCommCount(const ncclComm_t comm, int* count);
+ncclResult_t pncclCommCount(const ncclComm_t comm, int* count);
+
+/* Returns the cuda device number associated with the communicator. */
+ncclResult_t  ncclCommCuDevice(const ncclComm_t comm, int* device);
+ncclResult_t pncclCommCuDevice(const ncclComm_t comm, int* device);
+
+/* Returns the user-ordered "rank" associated with the communicator. */
+ncclResult_t  ncclCommUserRank(const ncclComm_t comm, int* rank);
+ncclResult_t pncclCommUserRank(const ncclComm_t comm, int* rank);
+
+/* Register CUDA buffer for zero-copy operation */
+ncclResult_t  ncclCommRegister(const ncclComm_t comm, void* buff, size_t size, void** handle);
+ncclResult_t pncclCommRegister(const ncclComm_t comm, void* buff, size_t size, void** handle);
+
+/* Deregister CUDA buffer */
+ncclResult_t  ncclCommDeregister(const ncclComm_t comm, void* handle);
+ncclResult_t pncclCommDeregister(const ncclComm_t comm, void* handle);
+
+/* Reduction operation selector */
+typedef enum { ncclNumOps_dummy = 5 } ncclRedOp_dummy_t;
+typedef enum { ncclSum        = 0,
+               ncclProd       = 1,
+               ncclMax        = 2,
+               ncclMin        = 3,
+               ncclAvg        = 4,
+               /* ncclNumOps: The number of built-in ncclRedOp_t values. Also
+                * serves as the least possible value for dynamic ncclRedOp_t's
+                * as constructed by ncclRedOpCreate*** functions. */
+               ncclNumOps     = 5,
+               /* ncclMaxRedOp: The largest valid value for ncclRedOp_t.
+                * It is defined to be the largest signed value (since compilers
+                * are permitted to use signed enums) that won't grow
+                * sizeof(ncclRedOp_t) when compared to previous NCCL versions to
+                * maintain ABI compatibility. */
+               ncclMaxRedOp   = 0x7fffffff>>(32-8*sizeof(ncclRedOp_dummy_t))
+             } ncclRedOp_t;
+
+/* Data types */
+typedef enum { ncclInt8       = 0, ncclChar       = 0,
+               ncclUint8      = 1,
+               ncclInt32      = 2, ncclInt        = 2,
+               ncclUint32     = 3,
+               ncclInt64      = 4,
+               ncclUint64     = 5,
+               ncclFloat16    = 6, ncclHalf       = 6,
+               ncclFloat32    = 7, ncclFloat      = 7,
+               ncclFloat64    = 8, ncclDouble     = 8,
+#if defined(__CUDA_BF16_TYPES_EXIST__)
+               ncclBfloat16   = 9,
+               ncclNumTypes   = 10
+#else
+               ncclNumTypes   = 9
+#endif
+} ncclDataType_t;
+
+/* ncclScalarResidence_t: Location and dereferencing logic for scalar arguments. */
+typedef enum {
+  /* ncclScalarDevice: The scalar is in device-visible memory and will be
+   * dereferenced while the collective is running. */
+  ncclScalarDevice = 0,
+
+  /* ncclScalarHostImmediate: The scalar is in host-visible memory and will be
+   * dereferenced before the ncclRedOpCreate***() function returns. */
+  ncclScalarHostImmediate = 1
+} ncclScalarResidence_t;
+
+/*
+ * ncclRedOpCreatePreMulSum
+ *
+ * Creates a new reduction operator which pre-multiplies input values by a given
+ * scalar locally before reducing them with peer values via summation. For use
+ * only with collectives launched against *comm* and *datatype*. The
+ * *residence* argument indicates how/when the memory pointed to by *scalar*
+ * will be dereferenced. Upon return, the newly created operator's handle
+ * is stored in *op*.
+ */
+ncclResult_t  ncclRedOpCreatePreMulSum(ncclRedOp_t *op, void *scalar, ncclDataType_t datatype, ncclScalarResidence_t residence, ncclComm_t comm);
+ncclResult_t pncclRedOpCreatePreMulSum(ncclRedOp_t *op, void *scalar, ncclDataType_t datatype, ncclScalarResidence_t residence, ncclComm_t comm);
+
+/*
+ * ncclRedOpDestroy
+ *
+ * Destroys the reduction operator *op*. The operator must have been created by
+ * ncclRedOpCreatePreMul with the matching communicator *comm*. An operator may be
+ * destroyed as soon as the last NCCL function which is given that operator returns.
+ */
+ncclResult_t ncclRedOpDestroy(ncclRedOp_t op, ncclComm_t comm);
+ncclResult_t pncclRedOpDestroy(ncclRedOp_t op, ncclComm_t comm);
+
+/*
+ * Collective communication operations
+ *
+ * Collective communication operations must be called separately for each
+ * communicator in a communicator clique.
+ *
+ * They return when operations have been enqueued on the CUDA stream.
+ *
+ * Since they may perform inter-CPU synchronization, each call has to be done
+ * from a different thread or process, or need to use Group Semantics (see
+ * below).
+ */
+
+/*
+ * Reduce
+ *
+ * Reduces data arrays of length count in sendbuff into recvbuff using op
+ * operation.
+ * recvbuff may be NULL on all calls except for root device.
+ * root is the rank (not the CUDA device) where data will reside after the
+ * operation is complete.
+ *
+ * In-place operation will happen if sendbuff == recvbuff.
+ */
+ncclResult_t  ncclReduce(const void* sendbuff, void* recvbuff, size_t count, ncclDataType_t datatype,
+    ncclRedOp_t op, int root, ncclComm_t comm, cudaStream_t stream);
+ncclResult_t pncclReduce(const void* sendbuff, void* recvbuff, size_t count, ncclDataType_t datatype,
+    ncclRedOp_t op, int root, ncclComm_t comm, cudaStream_t stream);
+
+/*
+ * (deprecated) Broadcast (in-place)
+ *
+ * Copies count values from root to all other devices.
+ * root is the rank (not the CUDA device) where data resides before the
+ * operation is started.
+ *
+ * This operation is implicitely in place.
+ */
+ncclResult_t  ncclBcast(void* buff, size_t count, ncclDataType_t datatype, int root,
+    ncclComm_t comm, cudaStream_t stream);
+ncclResult_t pncclBcast(void* buff, size_t count, ncclDataType_t datatype, int root,
+    ncclComm_t comm, cudaStream_t stream);
+
+/*
+ * Broadcast
+ *
+ * Copies count values from root to all other devices.
+ * root is the rank (not the CUDA device) where data resides before the
+ * operation is started.
+ *
+ * In-place operation will happen if sendbuff == recvbuff.
+ */
+ncclResult_t  ncclBroadcast(const void* sendbuff, void* recvbuff, size_t count, ncclDataType_t datatype, int root,
+    ncclComm_t comm, cudaStream_t stream);
+ncclResult_t pncclBroadcast(const void* sendbuff, void* recvbuff, size_t count, ncclDataType_t datatype, int root,
+    ncclComm_t comm, cudaStream_t stream);
+
+/*
+ * All-Reduce
+ *
+ * Reduces data arrays of length count in sendbuff using op operation, and
+ * leaves identical copies of result on each recvbuff.
+ *
+ * In-place operation will happen if sendbuff == recvbuff.
+ */
+ncclResult_t  ncclAllReduce(const void* sendbuff, void* recvbuff, size_t count,
+    ncclDataType_t datatype, ncclRedOp_t op, ncclComm_t comm, cudaStream_t stream);
+ncclResult_t pncclAllReduce(const void* sendbuff, void* recvbuff, size_t count,
+    ncclDataType_t datatype, ncclRedOp_t op, ncclComm_t comm, cudaStream_t stream);
+
+/*
+ * Reduce-Scatter
+ *
+ * Reduces data in sendbuff using op operation and leaves reduced result
+ * scattered over the devices so that recvbuff on rank i will contain the i-th
+ * block of the result.
+ * Assumes sendcount is equal to nranks*recvcount, which means that sendbuff
+ * should have a size of at least nranks*recvcount elements.
+ *
+ * In-place operations will happen if recvbuff == sendbuff + rank * recvcount.
+ */
+ncclResult_t  ncclReduceScatter(const void* sendbuff, void* recvbuff,
+    size_t recvcount, ncclDataType_t datatype, ncclRedOp_t op, ncclComm_t comm,
+    cudaStream_t stream);
+ncclResult_t pncclReduceScatter(const void* sendbuff, void* recvbuff,
+    size_t recvcount, ncclDataType_t datatype, ncclRedOp_t op, ncclComm_t comm,
+    cudaStream_t stream);
+
+/*
+ * All-Gather
+ *
+ * Each device gathers sendcount values from other GPUs into recvbuff,
+ * receiving data from rank i at offset i*sendcount.
+ * Assumes recvcount is equal to nranks*sendcount, which means that recvbuff
+ * should have a size of at least nranks*sendcount elements.
+ *
+ * In-place operations will happen if sendbuff == recvbuff + rank * sendcount.
+ */
+ncclResult_t  ncclAllGather(const void* sendbuff, void* recvbuff, size_t sendcount,
+    ncclDataType_t datatype, ncclComm_t comm, cudaStream_t stream);
+ncclResult_t pncclAllGather(const void* sendbuff, void* recvbuff, size_t sendcount,
+    ncclDataType_t datatype, ncclComm_t comm, cudaStream_t stream);
+
+/*
+ * Send
+ *
+ * Send data from sendbuff to rank peer.
+ *
+ * Rank peer needs to call ncclRecv with the same datatype and the same count from this
+ * rank.
+ *
+ * This operation is blocking for the GPU. If multiple ncclSend and ncclRecv operations
+ * need to progress concurrently to complete, they must be fused within a ncclGroupStart/
+ * ncclGroupEnd section.
+ */
+ncclResult_t  ncclSend(const void* sendbuff, size_t count, ncclDataType_t datatype, int peer,
+    ncclComm_t comm, cudaStream_t stream);
+ncclResult_t pncclSend(const void* sendbuff, size_t count, ncclDataType_t datatype, int peer,
+    ncclComm_t comm, cudaStream_t stream);
+
+/*
+ * Receive
+ *
+ * Receive data from rank peer into recvbuff.
+ *
+ * Rank peer needs to call ncclSend with the same datatype and the same count to this
+ * rank.
+ *
+ * This operation is blocking for the GPU. If multiple ncclSend and ncclRecv operations
+ * need to progress concurrently to complete, they must be fused within a ncclGroupStart/
+ * ncclGroupEnd section.
+ */
+ncclResult_t pncclRecv(void* recvbuff, size_t count, ncclDataType_t datatype, int peer,
+    ncclComm_t comm, cudaStream_t stream);
+ncclResult_t  ncclRecv(void* recvbuff, size_t count, ncclDataType_t datatype, int peer,
+    ncclComm_t comm, cudaStream_t stream);
+
+/*
+ * Group semantics
+ *
+ * When managing multiple GPUs from a single thread, and since NCCL collective
+ * calls may perform inter-CPU synchronization, we need to "group" calls for
+ * different ranks/devices into a single call.
+ *
+ * Grouping NCCL calls as being part of the same collective operation is done
+ * using ncclGroupStart and ncclGroupEnd. ncclGroupStart will enqueue all
+ * collective calls until the ncclGroupEnd call, which will wait for all calls
+ * to be complete. Note that for collective communication, ncclGroupEnd only
+ * guarantees that the operations are enqueued on the streams, not that
+ * the operation is effectively done.
+ *
+ * Both collective communication and ncclCommInitRank can be used in conjunction
+ * of ncclGroupStart/ncclGroupEnd, but not together.
+ *
+ * Group semantics also allow to fuse multiple operations on the same device
+ * to improve performance (for aggregated collective calls), or to permit
+ * concurrent progress of multiple send/receive operations.
+ */
+
+/*
+ * Group Start
+ *
+ * Start a group call. All calls to NCCL until ncclGroupEnd will be fused into
+ * a single NCCL operation. Nothing will be started on the CUDA stream until
+ * ncclGroupEnd.
+ */
+ncclResult_t  ncclGroupStart();
+ncclResult_t pncclGroupStart();
+
+/*
+ * Group End
+ *
+ * End a group call. Start a fused NCCL operation consisting of all calls since
+ * ncclGroupStart. Operations on the CUDA stream depending on the NCCL operations
+ * need to be called after ncclGroupEnd.
+ */
+ncclResult_t  ncclGroupEnd();
+ncclResult_t pncclGroupEnd();
+
+/*
+ * Group Simulate End
+ *
+ * Simulate a ncclGroupEnd() call and return NCCL's simulation info in a struct.
+ */
+ncclResult_t  ncclGroupSimulateEnd(ncclSimInfo_t* simInfo);
+ncclResult_t pncclGroupSimulateEnd(ncclSimInfo_t* simInfo);
+
+#ifdef __cplusplus
+} // end extern "C"
+#endif
+
+#endif // end include guard

nvidia/nccl/include/nccl_net.h

@@ -0,0 +1,456 @@
+/*************************************************************************
+ * Copyright (c) 2017-2022, NVIDIA CORPORATION. All rights reserved.
+ *
+ * See LICENSE.txt for license information
+ ************************************************************************/
+
+#ifndef NCCL_NET_H_
+#define NCCL_NET_H_
+
+#include "nccl.h"
+#include "nccl_common.h"
+#include "net_device.h"
+#include <stdint.h>
+
+#define NCCL_NET_HANDLE_MAXSIZE 128
+
+#define NCCL_PTR_HOST 0x1
+#define NCCL_PTR_CUDA 0x2
+#define NCCL_PTR_DMABUF 0x4
+
+// Maximum number of requests per comm object
+#define NCCL_NET_MAX_REQUESTS 32
+
+typedef struct {
+  char* name;                      // Used mostly for logging.
+  char* pciPath;                   // Path to the PCI device in /sys.
+  uint64_t guid;                   // Unique identifier for the NIC chip. Important for
+                                   // cards with multiple PCI functions (Physical or virtual).
+  int ptrSupport;                  // [NCCL_PTR_HOST|NCCL_PTR_CUDA|NCCL_PTR_DMABUF]
+  int regIsGlobal;                 // regMr is not tied to a particular comm
+  int speed;                       // Port speed in Mbps.
+  int port;                        // Port number.
+  float latency;                   // Network latency
+  int maxComms;                    // Maximum number of comms we can create
+  int maxRecvs;                    // Maximum number of grouped receives.
+  ncclNetDeviceType netDeviceType; // Network offload type
+  int netDeviceVersion;            // Version number for network offload
+} ncclNetProperties_v8_t;
+
+typedef ncclNetProperties_v8_t ncclNetProperties_t;
+
+typedef struct {
+  // Name of the network (mainly for logs)
+  const char* name;
+  // Initialize the network.
+  ncclResult_t (*init)(ncclDebugLogger_t logFunction);
+  // Return the number of adapters.
+  ncclResult_t (*devices)(int* ndev);
+  // Get various device properties.
+  ncclResult_t (*getProperties)(int dev, ncclNetProperties_v8_t* props);
+  // Create a receiving object and provide a handle to connect to it. The
+  // handle can be up to NCCL_NET_HANDLE_MAXSIZE bytes and will be exchanged
+  // between ranks to create a connection.
+  ncclResult_t (*listen)(int dev, void* handle, void** listenComm);
+  // Connect to a handle and return a sending comm object for that peer.
+  // This call must not block for the connection to be established, and instead
+  // should return successfully with sendComm == NULL with the expectation that
+  // it will be called again until sendComm != NULL.
+  // If *sendDevComm points to a valid object, then NCCL is requesting device offload for this connection
+  ncclResult_t (*connect)(int dev, void* handle, void** sendComm, ncclNetDeviceHandle_v8_t** sendDevComm);
+  // Finalize connection establishment after remote peer has called connect.
+  // This call must not block for the connection to be established, and instead
+  // should return successfully with recvComm == NULL with the expectation that
+  // it will be called again until recvComm != NULL.
+  // If *recvDevComm points to a valid object, then NCCL is requesting device offload for this connection
+  ncclResult_t (*accept)(void* listenComm, void** recvComm, ncclNetDeviceHandle_v8_t** recvDevComm);
+  // Register/Deregister memory. Comm can be either a sendComm or a recvComm.
+  // Type is either NCCL_PTR_HOST or NCCL_PTR_CUDA.
+  ncclResult_t (*regMr)(void* comm, void* data, size_t size, int type, void** mhandle);
+  /* DMA-BUF support */
+  ncclResult_t (*regMrDmaBuf)(void* comm, void* data, size_t size, int type, uint64_t offset, int fd, void** mhandle);
+  ncclResult_t (*deregMr)(void* comm, void* mhandle);
+  // Asynchronous send to a peer.
+  // May return request == NULL if the call cannot be performed (or would block)
+  ncclResult_t (*isend)(void* sendComm, void* data, int size, int tag, void* mhandle, void** request);
+  // Asynchronous recv from a peer.
+  // May return request == NULL if the call cannot be performed (or would block)
+  ncclResult_t (*irecv)(void* recvComm, int n, void** data, int* sizes, int* tags, void** mhandles, void** request);
+  // Perform a flush/fence to make sure all data received with NCCL_PTR_CUDA is
+  // visible to the GPU
+  ncclResult_t (*iflush)(void* recvComm, int n, void** data, int* sizes, void** mhandles, void** request);
+  // Test whether a request is complete. If size is not NULL, it returns the
+  // number of bytes sent/received.
+  ncclResult_t (*test)(void* request, int* done, int* sizes);
+  // Close and free send/recv comm objects
+  ncclResult_t (*closeSend)(void* sendComm);
+  ncclResult_t (*closeRecv)(void* recvComm);
+  ncclResult_t (*closeListen)(void* listenComm);
+
+  // Copy the given mhandle to a dptr in a format usable by this plugin's device code
+  ncclResult_t (*getDeviceMr)(void* comm, void* mhandle, void** dptr_mhandle);
+
+  // Notify the plugin that a recv has completed by the device
+  ncclResult_t (*irecvConsumed)(void* recvComm, int n, void* request);
+} ncclNet_v8_t;
+
+typedef ncclNet_v8_t ncclNet_t;
+
+#define NCCL_NET_PLUGIN_SYMBOL ncclNetPlugin_v8
+
+typedef struct {
+  void* mhandle;
+  void* address;
+  uint32_t size;
+} ncclNetSGE_v8_t;
+
+typedef struct {
+  // Name of the collective network (mainly for logs)
+  const char* name;
+  // Initialize the collective network.
+  ncclResult_t (*init)(ncclDebugLogger_t logFunction);
+  // Return the number of adapters capable of doing collective operations.
+  // If ndev returns 0, all other functions might be set to NULL.
+  ncclResult_t (*devices)(int* ndev);
+  // Get various device properties.
+  ncclResult_t (*getProperties)(int dev, ncclNetProperties_v8_t* props);
+  // Create a receiving object and provide a handle to connect to it. The
+  // handle can be up to NCCL_NET_HANDLE_MAXSIZE bytes and will be exchanged
+  // between ranks to create connections.
+  ncclResult_t (*listen)(int dev, void* handle, void** listenComm);
+  // Create a group for collective operations. handles have been created
+  // using listen() above. rank indicates caller's rank in the collective network.
+  ncclResult_t (*connect)(void* handles[], int nranks, int rank, void* listenComm, void** collComm);
+  // Returns whether a reduction operation on a data type is supported.
+  // 1 for supported, 0 otherwise.
+  ncclResult_t (*reduceSupport)(ncclDataType_t dataType, ncclRedOp_t redOp, int* supported);
+  // Register/Deregister memory. Type is either NCCL_PTR_HOST or NCCL_PTR_CUDA.
+  ncclResult_t (*regMr)(void* collComm, void* data, size_t size, int type, void** mhandle);
+  /* DMA-BUF support */
+  ncclResult_t (*regMrDmaBuf)(void* collComm, void* data, size_t size, int type, uint64_t offset, int fd, void** mhandle);
+  ncclResult_t (*deregMr)(void* collComm, void* mhandle);
+  // Performs an asynchronous allreduce operation on the collective group.
+  // May return request == NULL if the call cannot be performed (or would block).
+  ncclResult_t (*iallreduce)(void* collComm, void* sendData, void* recvData, int count,
+      ncclDataType_t dataType, ncclRedOp_t redOp, void* sendMhandle, void* recvMhandle, void** request);
+  ncclResult_t (*iallgather)(void* collComm, void* sendData, int nRecvParts, ncclNetSGE_v8_t* recvParts,
+                             size_t bytesPerRank, size_t windowOffset, size_t windowBytes,
+                             void* sendMhandle, void** request);
+  ncclResult_t (*ireducescatter)(void* collComm, int nSendParts, ncclNetSGE_v8_t* sendParts, void* recvData,
+                                 size_t bytesPerRank, size_t windowOffset, size_t windowBytes,
+                                 ncclDataType_t dataType, ncclRedOp_t redOp,
+                                 void* recvMhandle, void** request);
+  // Perform a flush/fence to make sure all data received with NCCL_PTR_CUDA is
+  // visible to the GPU
+  ncclResult_t (*iflush)(void* collComm, void* data, int size, void* mhandle, void** request);
+  // Test whether a request is complete. If size is not NULL, it returns the
+  // number of bytes sent/received.
+  ncclResult_t (*test)(void* request, int* done, int* size);
+  // Close and free collective comm objects
+  ncclResult_t (*closeColl)(void* collComm);
+  ncclResult_t (*closeListen)(void* listenComm);
+} ncclCollNet_v8_t;
+
+typedef ncclCollNet_v8_t ncclCollNet_t;
+
+#define NCCL_COLLNET_PLUGIN_SYMBOL ncclCollNetPlugin_v8
+
+typedef struct {
+  char* name;                      // Used mostly for logging.
+  char* pciPath;                   // Path to the PCI device in /sys.
+  uint64_t guid;                   // Unique identifier for the NIC chip. Important for
+                                   // cards with multiple PCI functions (Physical or virtual).
+  int ptrSupport;                  // [NCCL_PTR_HOST|NCCL_PTR_CUDA|NCCL_PTR_DMABUF]
+  int speed;                       // Port speed in Mbps.
+  int port;                        // Port number.
+  float latency;                   // Network latency
+  int maxComms;                    // Maximum number of comms we can create
+  int maxRecvs;                    // Maximum number of grouped receives.
+  ncclNetDeviceType netDeviceType; // Network offload type
+  int netDeviceVersion;            // Version number for network offload
+} ncclNetProperties_v7_t;
+
+typedef struct {
+  // Name of the network (mainly for logs)
+  const char* name;
+  // Initialize the network.
+  ncclResult_t (*init)(ncclDebugLogger_t logFunction);
+  // Return the number of adapters.
+  ncclResult_t (*devices)(int* ndev);
+  // Get various device properties.
+  ncclResult_t (*getProperties)(int dev, ncclNetProperties_v7_t* props);
+  // Create a receiving object and provide a handle to connect to it. The
+  // handle can be up to NCCL_NET_HANDLE_MAXSIZE bytes and will be exchanged
+  // between ranks to create a connection.
+  ncclResult_t (*listen)(int dev, void* handle, void** listenComm);
+  // Connect to a handle and return a sending comm object for that peer.
+  // This call must not block for the connection to be established, and instead
+  // should return successfully with sendComm == NULL with the expectation that
+  // it will be called again until sendComm != NULL.
+  // If *sendDevComm points to a valid object, then NCCL is requesting device offload for this connection
+  ncclResult_t (*connect)(int dev, void* handle, void** sendComm, ncclNetDeviceHandle_v7_t** sendDevComm);
+  // Finalize connection establishment after remote peer has called connect.
+  // This call must not block for the connection to be established, and instead
+  // should return successfully with recvComm == NULL with the expectation that
+  // it will be called again until recvComm != NULL.
+  // If *recvDevComm points to a valid object, then NCCL is requesting device offload for this connection
+  ncclResult_t (*accept)(void* listenComm, void** recvComm, ncclNetDeviceHandle_v7_t** recvDevComm);
+  // Register/Deregister memory. Comm can be either a sendComm or a recvComm.
+  // Type is either NCCL_PTR_HOST or NCCL_PTR_CUDA.
+  ncclResult_t (*regMr)(void* comm, void* data, int size, int type, void** mhandle);
+  /* DMA-BUF support */
+  ncclResult_t (*regMrDmaBuf)(void* comm, void* data, size_t size, int type, uint64_t offset, int fd, void** mhandle);
+  ncclResult_t (*deregMr)(void* comm, void* mhandle);
+  // Asynchronous send to a peer.
+  // May return request == NULL if the call cannot be performed (or would block)
+  ncclResult_t (*isend)(void* sendComm, void* data, int size, int tag, void* mhandle, void** request);
+  // Asynchronous recv from a peer.
+  // May return request == NULL if the call cannot be performed (or would block)
+  ncclResult_t (*irecv)(void* recvComm, int n, void** data, int* sizes, int* tags, void** mhandles, void** request);
+  // Perform a flush/fence to make sure all data received with NCCL_PTR_CUDA is
+  // visible to the GPU
+  ncclResult_t (*iflush)(void* recvComm, int n, void** data, int* sizes, void** mhandles, void** request);
+  // Test whether a request is complete. If size is not NULL, it returns the
+  // number of bytes sent/received.
+  ncclResult_t (*test)(void* request, int* done, int* sizes);
+  // Close and free send/recv comm objects
+  ncclResult_t (*closeSend)(void* sendComm);
+  ncclResult_t (*closeRecv)(void* recvComm);
+  ncclResult_t (*closeListen)(void* listenComm);
+
+  // Copy the given mhandle to a dptr in a format usable by this plugin's device code
+  ncclResult_t (*getDeviceMr)(void* comm, void* mhandle, void** dptr_mhandle);
+
+  // Notify the plugin that a recv has completed by the device
+  ncclResult_t (*irecvConsumed)(void* recvComm, int n, void* request);
+} ncclNet_v7_t;
+
+typedef struct {
+  // Name of the collective network (mainly for logs)
+  const char* name;
+  // Initialize the collective network.
+  ncclResult_t (*init)(ncclDebugLogger_t logFunction);
+  // Return the number of adapters capable of doing collective operations.
+  // If ndev returns 0, all other functions might be set to NULL.
+  ncclResult_t (*devices)(int* ndev);
+  // Get various device properties.
+  ncclResult_t (*getProperties)(int dev, ncclNetProperties_v7_t* props);
+  // Create a receiving object and provide a handle to connect to it. The
+  // handle can be up to NCCL_NET_HANDLE_MAXSIZE bytes and will be exchanged
+  // between ranks to create connections.
+  ncclResult_t (*listen)(int dev, void* handle, void** listenComm);
+  // Create a group for collective operations. handles have been created
+  // using listen() above. rank indicates caller's rank in the collective network.
+  ncclResult_t (*connect)(void* handles[], int nranks, int rank, void* listenComm, void** collComm);
+  // Returns whether a reduction operation on a data type is supported.
+  // 1 for supported, 0 otherwise.
+  ncclResult_t (*reduceSupport)(ncclDataType_t dataType, ncclRedOp_t redOp, int* supported);
+  // Register/Deregister memory. Type is either NCCL_PTR_HOST or NCCL_PTR_CUDA.
+  ncclResult_t (*regMr)(void* collComm, void* data, int size, int type, void** mhandle);
+  /* DMA-BUF support */
+  ncclResult_t (*regMrDmaBuf)(void* collComm, void* data, size_t size, int type, uint64_t offset, int fd, void** mhandle);
+  ncclResult_t (*deregMr)(void* collComm, void* mhandle);
+  // Performs an asynchronous allreduce operation on the collective group.
+  // May return request == NULL if the call cannot be performed (or would block).
+  ncclResult_t (*iallreduce)(void* collComm, void* sendData, void* recvData, int count,
+      ncclDataType_t dataType, ncclRedOp_t redOp, void* sendMhandle, void* recvMhandle, void** request);
+  // Perform a flush/fence to make sure all data received with NCCL_PTR_CUDA is
+  // visible to the GPU
+  ncclResult_t (*iflush)(void* collComm, void* data, int size, void* mhandle, void** request);
+  // Test whether a request is complete. If size is not NULL, it returns the
+  // number of bytes sent/received.
+  ncclResult_t (*test)(void* request, int* done, int* size);
+  // Close and free collective comm objects
+  ncclResult_t (*closeColl)(void* collComm);
+  ncclResult_t (*closeListen)(void* listenComm);
+} ncclCollNet_v7_t;
+
+#define NCCL_NET_MAX_REQUESTS_V6 8
+
+// v6 struct for backwards compatibility
+typedef struct {
+  char* name;     // Used mostly for logging.
+  char* pciPath;  // Path to the PCI device in /sys.
+  uint64_t guid;  // Unique identifier for the NIC chip. Important for
+                  // cards with multiple PCI functions (Physical or virtual).
+  int ptrSupport; // [NCCL_PTR_HOST|NCCL_PTR_CUDA|NCCL_PTR_DMABUF]
+  int speed;      // Port speed in Mbps.
+  int port;       // Port number.
+  float latency;  // Network latency
+  int maxComms;   // Maximum number of comms we can create
+  int maxRecvs;   // Maximum number of grouped receives.
+} ncclNetProperties_v6_t;
+
+typedef struct {
+  // Name of the network (mainly for logs)
+  const char* name;
+  // Initialize the network.
+  ncclResult_t (*init)(ncclDebugLogger_t logFunction);
+  // Return the number of adapters.
+  ncclResult_t (*devices)(int* ndev);
+  // Get various device properties.
+  ncclResult_t (*getProperties)(int dev, ncclNetProperties_v6_t* props);
+  // Create a receiving object and provide a handle to connect to it. The
+  // handle can be up to NCCL_NET_HANDLE_MAXSIZE bytes and will be exchanged
+  // between ranks to create a connection.
+  ncclResult_t (*listen)(int dev, void* handle, void** listenComm);
+  // Connect to a handle and return a sending comm object for that peer.
+  // This call must not block for the connection to be established, and instead
+  // should return successfully with sendComm == NULL with the expectation that
+  // it will be called again until sendComm != NULL.
+  ncclResult_t (*connect)(int dev, void* handle, void** sendComm);
+  // Finalize connection establishment after remote peer has called connect.
+  // This call must not block for the connection to be established, and instead
+  // should return successfully with recvComm == NULL with the expectation that
+  // it will be called again until recvComm != NULL.
+  ncclResult_t (*accept)(void* listenComm, void** recvComm);
+  // Register/Deregister memory. Comm can be either a sendComm or a recvComm.
+  // Type is either NCCL_PTR_HOST or NCCL_PTR_CUDA.
+  ncclResult_t (*regMr)(void* comm, void* data, int size, int type, void** mhandle);
+  /* DMA-BUF support */
+  ncclResult_t (*regMrDmaBuf)(void* comm, void* data, size_t size, int type, uint64_t offset, int fd, void** mhandle);
+  ncclResult_t (*deregMr)(void* comm, void* mhandle);
+  // Asynchronous send to a peer.
+  // May return request == NULL if the call cannot be performed (or would block)
+  ncclResult_t (*isend)(void* sendComm, void* data, int size, int tag, void* mhandle, void** request);
+  // Asynchronous recv from a peer.
+  // May return request == NULL if the call cannot be performed (or would block)
+  ncclResult_t (*irecv)(void* recvComm, int n, void** data, int* sizes, int* tags, void** mhandles, void** request);
+  // Perform a flush/fence to make sure all data received with NCCL_PTR_CUDA is
+  // visible to the GPU
+  ncclResult_t (*iflush)(void* recvComm, int n, void** data, int* sizes, void** mhandles, void** request);
+  // Test whether a request is complete. If size is not NULL, it returns the
+  // number of bytes sent/received.
+  ncclResult_t (*test)(void* request, int* done, int* sizes);
+  // Close and free send/recv comm objects
+  ncclResult_t (*closeSend)(void* sendComm);
+  ncclResult_t (*closeRecv)(void* recvComm);
+  ncclResult_t (*closeListen)(void* listenComm);
+} ncclNet_v6_t;
+
+typedef struct {
+  // Name of the collective network (mainly for logs)
+  const char* name;
+  // Initialize the collective network.
+  ncclResult_t (*init)(ncclDebugLogger_t logFunction);
+  // Return the number of adapters capable of doing collective operations.
+  // If ndev returns 0, all other functions might be set to NULL.
+  ncclResult_t (*devices)(int* ndev);
+  // Get various device properties.
+  ncclResult_t (*getProperties)(int dev, ncclNetProperties_v6_t* props);
+  // Create a receiving object and provide a handle to connect to it. The
+  // handle can be up to NCCL_NET_HANDLE_MAXSIZE bytes and will be exchanged
+  // between ranks to create connections.
+  ncclResult_t (*listen)(int dev, void* handle, void** listenComm);
+  // Create a group for collective operations. handles have been created
+  // using listen() above. rank indicates caller's rank in the collective network.
+  ncclResult_t (*connect)(void* handles[], int nranks, int rank, void* listenComm, void** collComm);
+  // Returns whether a reduction operation on a data type is supported.
+  // 1 for supported, 0 otherwise.
+  ncclResult_t (*reduceSupport)(ncclDataType_t dataType, ncclRedOp_t redOp, int* supported);
+  // Register/Deregister memory. Type is either NCCL_PTR_HOST or NCCL_PTR_CUDA.
+  ncclResult_t (*regMr)(void* collComm, void* data, int size, int type, void** mhandle);
+  /* DMA-BUF support */
+  ncclResult_t (*regMrDmaBuf)(void* collComm, void* data, size_t size, int type, uint64_t offset, int fd, void** mhandle);
+  ncclResult_t (*deregMr)(void* collComm, void* mhandle);
+  // Performs an asynchronous allreduce operation on the collective group.
+  // May return request == NULL if the call cannot be performed (or would block).
+  ncclResult_t (*iallreduce)(void* collComm, void* sendData, void* recvData, int count,
+      ncclDataType_t dataType, ncclRedOp_t redOp, void* sendMhandle, void* recvMhandle, void** request);
+  // Perform a flush/fence to make sure all data received with NCCL_PTR_CUDA is
+  // visible to the GPU
+  ncclResult_t (*iflush)(void* collComm, void* data, int size, void* mhandle, void** request);
+  // Test whether a request is complete. If size is not NULL, it returns the
+  // number of bytes sent/received.
+  ncclResult_t (*test)(void* request, int* done, int* size);
+  // Close and free collective comm objects
+  ncclResult_t (*closeColl)(void* collComm);
+  ncclResult_t (*closeListen)(void* listenComm);
+} ncclCollNet_v6_t;
+
+// v5 struct for backwards compatibility
+typedef struct {
+  // Name of the network (mainly for logs)
+  const char* name;
+  // Initialize the network.
+  ncclResult_t (*init)(ncclDebugLogger_t logFunction);
+  // Return the number of adapters.
+  ncclResult_t (*devices)(int* ndev);
+  // Get various device properties.
+  ncclResult_t (*getProperties)(int dev, ncclNetProperties_v6_t* props);
+  // Create a receiving object and provide a handle to connect to it. The
+  // handle can be up to NCCL_NET_HANDLE_MAXSIZE bytes and will be exchanged
+  // between ranks to create a connection.
+  ncclResult_t (*listen)(int dev, void* handle, void** listenComm);
+  // Connect to a handle and return a sending comm object for that peer.
+  // This call must not block for the connection to be established, and instead
+  // should return successfully with sendComm == NULL with the expectation that
+  // it will be called again until sendComm != NULL.
+  ncclResult_t (*connect)(int dev, void* handle, void** sendComm);
+  // Finalize connection establishment after remote peer has called connect.
+  // This call must not block for the connection to be established, and instead
+  // should return successfully with recvComm == NULL with the expectation that
+  // it will be called again until recvComm != NULL.
+  ncclResult_t (*accept)(void* listenComm, void** recvComm);
+  // Register/Deregister memory. Comm can be either a sendComm or a recvComm.
+  // Type is either NCCL_PTR_HOST or NCCL_PTR_CUDA.
+  ncclResult_t (*regMr)(void* comm, void* data, int size, int type, void** mhandle);
+  ncclResult_t (*deregMr)(void* comm, void* mhandle);
+  // Asynchronous send to a peer.
+  // May return request == NULL if the call cannot be performed (or would block)
+  ncclResult_t (*isend)(void* sendComm, void* data, int size, int tag, void* mhandle, void** request);
+  // Asynchronous recv from a peer.
+  // May return request == NULL if the call cannot be performed (or would block)
+  ncclResult_t (*irecv)(void* recvComm, int n, void** data, int* sizes, int* tags, void** mhandles, void** request);
+  // Perform a flush/fence to make sure all data received with NCCL_PTR_CUDA is
+  // visible to the GPU
+  ncclResult_t (*iflush)(void* recvComm, int n, void** data, int* sizes, void** mhandles, void** request);
+  // Test whether a request is complete. If size is not NULL, it returns the
+  // number of bytes sent/received.
+  ncclResult_t (*test)(void* request, int* done, int* sizes);
+  // Close and free send/recv comm objects
+  ncclResult_t (*closeSend)(void* sendComm);
+  ncclResult_t (*closeRecv)(void* recvComm);
+  ncclResult_t (*closeListen)(void* listenComm);
+} ncclNet_v5_t;
+
+// v5 struct for backwards compatibility
+typedef struct {
+  // Name of the collective network (mainly for logs)
+  const char* name;
+  // Initialize the collective network.
+  ncclResult_t (*init)(ncclDebugLogger_t logFunction);
+  // Return the number of adapters capable of doing collective operations.
+  // If ndev returns 0, all other functions might be set to NULL.
+  ncclResult_t (*devices)(int* ndev);
+  // Get various device properties.
+  ncclResult_t (*getProperties)(int dev, ncclNetProperties_v6_t* props);
+  // Create a receiving object and provide a handle to connect to it. The
+  // handle can be up to NCCL_NET_HANDLE_MAXSIZE bytes and will be exchanged
+  // between ranks to create connections.
+  ncclResult_t (*listen)(int dev, void* handle, void** listenComm);
+  // Create a group for collective operations. handles have been created
+  // using listen() above. rank indicates caller's rank in the collective network.
+  ncclResult_t (*connect)(void* handles[], int nranks, int rank, void* listenComm, void** collComm);
+  // Returns whether a reduction operation on a data type is supported.
+  // 1 for supported, 0 otherwise.
+  ncclResult_t (*reduceSupport)(ncclDataType_t dataType, ncclRedOp_t redOp, int* supported);
+  // Register/Deregister memory. Type is either NCCL_PTR_HOST or NCCL_PTR_CUDA.
+  ncclResult_t (*regMr)(void* collComm, void* data, int size, int type, void** mhandle);
+  ncclResult_t (*deregMr)(void* collComm, void* mhandle);
+  // Performs an asynchronous allreduce operation on the collective group.
+  // May return request == NULL if the call cannot be performed (or would block).
+  ncclResult_t (*iallreduce)(void* collComm, void* sendData, void* recvData, int count,
+      ncclDataType_t dataType, ncclRedOp_t redOp, void* sendMhandle, void* recvMhandle, void** request);
+  // Perform a flush/fence to make sure all data received with NCCL_PTR_CUDA is
+  // visible to the GPU
+  ncclResult_t (*iflush)(void* collComm, void* data, int size, void* mhandle, void** request);
+  // Test whether a request is complete. If size is not NULL, it returns the
+  // number of bytes sent/received.
+  ncclResult_t (*test)(void* request, int* done, int* size);
+  // Close and free collective comm objects
+  ncclResult_t (*closeColl)(void* collComm);
+  ncclResult_t (*closeListen)(void* listenComm);
+} ncclCollNet_v5_t;
+
+#endif // end include guard

nvidia_nccl_cu12-2.23.4.dist-info/License.txt

@@ -0,0 +1,39 @@
+
+ Copyright (c) 2015-2020, NVIDIA CORPORATION. All rights reserved.
+
+ Redistribution and use in source and binary forms, with or without
+ modification, are permitted provided that the following conditions
+ are met:
+  * Redistributions of source code must retain the above copyright
+    notice, this list of conditions and the following disclaimer.
+  * Redistributions in binary form must reproduce the above copyright
+    notice, this list of conditions and the following disclaimer in the
+    documentation and/or other materials provided with the distribution.
+  * Neither the name of NVIDIA CORPORATION, Lawrence Berkeley National
+    Laboratory, the U.S. Department of Energy, nor the names of their
+    contributors may be used to endorse or promote products derived
+    from this software without specific prior written permission.
+
+ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY
+ EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+ PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+ CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+ EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+ PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+ PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
+ OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+ The U.S. Department of Energy funded the development of this software
+ under subcontract 7078610 with Lawrence Berkeley National Laboratory.
+
+
+This code also includes files from the NVIDIA Tools Extension SDK project.
+
+See:
+
+   https://github.com/NVIDIA/NVTX
+
+for more information and license details.

nvidia_nccl_cu12-2.23.4.dist-info/METADATA

@@ -0,0 +1,35 @@
+Metadata-Version: 2.1
+Name: nvidia-nccl-cu12
+Version: 2.23.4
+Summary: NVIDIA Collective Communication Library (NCCL) Runtime
+Home-page: https://developer.nvidia.com/cuda-zone
+Author: Nvidia CUDA Installer Team
+Author-email: compute_installer@nvidia.com
+License: BSD-3-Clause
+Keywords: cuda,nvidia,runtime,machine learning,deep learning
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Education
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: Other/Proprietary License
+Classifier: Natural Language :: English
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.5
+Classifier: Programming Language :: Python :: 3.6
+Classifier: Programming Language :: Python :: 3.7
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Scientific/Engineering
+Classifier: Topic :: Scientific/Engineering :: Mathematics
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Topic :: Software Development
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Operating System :: POSIX :: Linux
+Requires-Python: >=3
+License-File: License.txt
+
+NCCL (pronounced "Nickel") is a stand-alone library of standard collective communication routines for GPUs, implementing all-reduce, all-gather, reduce, broadcast, and reduce-scatter. It has been optimized to achieve high bandwidth on any platform using PCIe, NVLink, NVswitch, as well as networking using InfiniBand Verbs or TCP/IP sockets.

nvidia_nccl_cu12-2.23.4.dist-info/RECORD

@@ -0,0 +1,12 @@
+nvidia/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+nvidia/nccl/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+nvidia/nccl/include/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+nvidia/nccl/include/nccl.h,sha256=-xnqDygmn2YmOkYW3bakiRfIsnvwUAptTJeZoMWsXdM,20225
+nvidia/nccl/include/nccl_net.h,sha256=Q2yMEZBE6uKkX_nduRb3TaU64hC1jjOMstCypTKGSdk,25896
+nvidia/nccl/lib/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+nvidia/nccl/lib/libnccl.so.2,sha256=cpNUzYaGHk_iB-IyLjS-_cVxLENE_6MKeZS-bGDm3zU,267885488
+nvidia_nccl_cu12-2.23.4.dist-info/License.txt,sha256=DwF0prTgszrCY3W_cpUzB1sy9MUaW2gCo9dC19zcmnY,1895
+nvidia_nccl_cu12-2.23.4.dist-info/METADATA,sha256=gInA2V7kgbe5Q2WGriCuxDlUeVD70aQvto0S5vPqd8o,1822
+nvidia_nccl_cu12-2.23.4.dist-info/WHEEL,sha256=SFSBTBkI2Eso5pk64VndQikelT62YOGkTuu55QuBsyk,109
+nvidia_nccl_cu12-2.23.4.dist-info/top_level.txt,sha256=fTkAtiFuL16nUrB9ytDDtpytz2t0B4NvYTnRzwAhO14,7
+nvidia_nccl_cu12-2.23.4.dist-info/RECORD,,

nvidia_nccl_cu12-2.23.4.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (74.1.2)
+Root-Is-Purelib: true
+Tag: py3-none-manylinux2014_aarch64
+

nvidia_nccl_cu12-2.23.4.dist-info/top_level.txt

@@ -0,0 +1 @@
+nvidia