.. role:: hidden :class: hidden-section Distributed communication package (deprecated) - torch.distributed.deprecated ============================================================================= .. warning:: torch.distributed.deprecated is the older version of torch.distributed and currently deprecated. It will be removed soon. Please use and refer the doc for torch.distributed, which is the latest distributed communication package for PyTorch .. automodule:: torch.distributed.deprecated .. currentmodule:: torch.distributed.deprecated Currently torch.distributed.deprecated supports four backends, each with different capabilities. The table below shows which functions are available for use with CPU / CUDA tensors. MPI supports cuda only if the implementation used to build PyTorch supports it. +------------+-----------+-----------+-----------+-----------+ | Backend | ``tcp`` | ``gloo`` | ``mpi`` | ``nccl`` | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | Device | CPU | GPU | CPU | GPU | CPU | GPU | CPU | GPU | +============+=====+=====+=====+=====+=====+=====+=====+=====+ | send | ✓ | ✘ | ✘ | ✘ | ✓ | ? | ✘ | ✘ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | recv | ✓ | ✘ | ✘ | ✘ | ✓ | ? | ✘ | ✘ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | broadcast | ✓ | ✘ | ✓ | ✓ | ✓ | ? | ✘ | ✓ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | all_reduce | ✓ | ✘ | ✓ | ✓ | ✓ | ? | ✘ | ✓ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | reduce | ✓ | ✘ | ✘ | ✘ | ✓ | ? | ✘ | ✓ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | all_gather | ✓ | ✘ | ✘ | ✘ | ✓ | ? | ✘ | ✓ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | gather | ✓ | ✘ | ✘ | ✘ | ✓ | ? | ✘ | ✘ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | scatter | ✓ | ✘ | ✘ | ✘ | ✓ | ? | ✘ | ✘ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ | barrier | ✓ | ✘ | ✓ | ✓ | ✓ | ? | ✘ | ✘ | +------------+-----+-----+-----+-----+-----+-----+-----+-----+ .. _distributed-deprecated-basics: Basics ------ The `torch.distributed.deprecated` package provides PyTorch support and communication primitives for multiprocess parallelism across several computation nodes running on one or more machines. The class :func:`torch.nn.parallel.deprecated.DistributedDataParallel` builds on this functionality to provide synchronous distributed training as a wrapper around any PyTorch model. This differs from the kinds of parallelism provided by :doc:`multiprocessing` and :func:`torch.nn.DataParallel` in that it supports multiple network-connected machines and in that the user must explicitly launch a separate copy of the main training script for each process. In the single-machine synchronous case, `torch.distributed.deprecated` or the :func:`torch.nn.parallel.deprecated.DistributedDataParallel` wrapper may still have advantages over other approaches to data-parallelism, including :func:`torch.nn.DataParallel`: * Each process maintains its own optimizer and performs a complete optimization step with each iteration. While this may appear redundant, since the gradients have already been gathered together and averaged across processes and are thus the same for every process, this means that no parameter broadcast step is needed, reducing time spent transferring tensors between nodes. * Each process contains an independent Python interpreter, eliminating the extra interpreter overhead and "GIL-thrashing" that comes from driving several execution threads, model replicas, or GPUs from a single Python process. This is especially important for models that make heavy use of the Python runtime, including models with recurrent layers or many small components. Initialization -------------- The package needs to be initialized using the :func:`torch.distributed.deprecated.init_process_group` function before calling any other methods. This blocks until all processes have joined. .. autofunction:: init_process_group .. autofunction:: get_rank .. autofunction:: get_world_size -------------------------------------------------------------------------------- Currently three initialization methods are supported: TCP initialization ^^^^^^^^^^^^^^^^^^ There are two ways to initialize using TCP, both requiring a network address reachable from all processes and a desired ``world_size``. The first way requires specifying an address that belongs to the rank 0 process. This initialization method requires that all processes have manually specified ranks. Alternatively, the address has to be a valid IP multicast address, in which case ranks can be assigned automatically. Multicast initialization also supports a ``group_name`` argument, which allows you to use the same address for multiple jobs, as long as they use different group names. :: import torch.distributed.deprecated as dist # Use address of one of the machines dist.init_process_group(backend, init_method='tcp://10.1.1.20:23456', rank=args.rank, world_size=4) # or a multicast address - rank will be assigned automatically if unspecified dist.init_process_group(backend, init_method='tcp://[ff15:1e18:5d4c:4cf0:d02d:b659:53ba:b0a7]:23456', world_size=4) Shared file-system initialization ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Another initialization method makes use of a file system that is shared and visible from all machines in a group, along with a desired ``world_size``. The URL should start with ``file://`` and contain a path to a non-existent file (in an existing directory) on a shared file system. This initialization method also supports a ``group_name`` argument, which allows you to use the same shared file path for multiple jobs, as long as they use different group names. .. warning:: This method assumes that the file system supports locking using ``fcntl`` - most local systems and NFS support it. :: import torch.distributed.deprecated as dist # Rank will be assigned automatically if unspecified dist.init_process_group(backend, init_method='file:///mnt/nfs/sharedfile', world_size=4, group_name=args.group) Environment variable initialization ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ This method will read the configuration from environment variables, allowing one to fully customize how the information is obtained. The variables to be set are: * ``MASTER_PORT`` - required; has to be a free port on machine with rank 0 * ``MASTER_ADDR`` - required (except for rank 0); address of rank 0 node * ``WORLD_SIZE`` - required; can be set either here, or in a call to init function * ``RANK`` - required; can be set either here, or in a call to init function The machine with rank 0 will be used to set up all connections. This is the default method, meaning that ``init_method`` does not have to be specified (or can be ``env://``). Groups ------ By default collectives operate on the default group (also called the world) and require all processes to enter the distributed function call. However, some workloads can benefit from more fine-grained communication. This is where distributed groups come into play. :func:`~torch.distributed.deprecated.new_group` function can be used to create new groups, with arbitrary subsets of all processes. It returns an opaque group handle that can be given as a ``group`` argument to all collectives (collectives are distributed functions to exchange information in certain well-known programming patterns). .. autofunction:: new_group Point-to-point communication ---------------------------- .. autofunction:: send .. autofunction:: recv :func:`~torch.distributed.deprecated.isend` and :func:`~torch.distributed.deprecated.irecv` return distributed request objects when used. In general, the type of this object is unspecified as they should never be created manually, but they are guaranteed to support two methods: * ``is_completed()`` - returns True if the operation has finished * ``wait()`` - will block the process until the operation is finished. ``is_completed()`` is guaranteed to return True once it returns. When using the MPI backend, :func:`~torch.distributed.deprecated.isend` and :func:`~torch.distributed.deprecated.irecv` support non-overtaking, which has some guarantees on supporting message order. For more detail, see http://mpi-forum.org/docs/mpi-2.2/mpi22-report/node54.htm#Node54 .. autofunction:: isend .. autofunction:: irecv Collective functions -------------------- .. autofunction:: broadcast .. autofunction:: all_reduce .. autofunction:: reduce .. autofunction:: all_gather .. autofunction:: gather .. autofunction:: scatter .. autofunction:: barrier Multi-GPU collective functions ------------------------------ If you have more than one GPU on each node, when using the NCCL backend, :func:`~torch.distributed.deprecated.broadcast_multigpu` :func:`~torch.distributed.deprecated.all_reduce_multigpu` :func:`~torch.distributed.deprecated.reduce_multigpu` and :func:`~torch.distributed.deprecated.all_gather_multigpu` support distributed collective operations among multiple GPUs within each node. These functions can potentially improve the overall distributed training performance and be easily used by passing a list of tensors. Each Tensor in the passed tensor list needs to be on a separate GPU device of the host where the function is called. Note that the length of the tensor list needs to be identical among all the distributed processes. Also note that currently the multi-GPU collective functions are only supported by the NCCL backend. For example, if the system we use for distributed training has 2 nodes, each of which has 8 GPUs. On each of the 16 GPUs, there is a tensor that we would like to all-reduce. The following code can serve as a reference: Code running on Node 0 :: import torch import torch.distributed.deprecated as dist dist.init_process_group(backend="nccl", init_method="file:///distributed_test", world_size=2, rank=0) tensor_list = [] for dev_idx in range(torch.cuda.device_count()): tensor_list.append(torch.FloatTensor([1]).cuda(dev_idx)) dist.all_reduce_multigpu(tensor_list) Code running on Node 1 :: import torch import torch.distributed.deprecated as dist dist.init_process_group(backend="nccl", init_method="file:///distributed_test", world_size=2, rank=1) tensor_list = [] for dev_idx in range(torch.cuda.device_count()): tensor_list.append(torch.FloatTensor([1]).cuda(dev_idx)) dist.all_reduce_multigpu(tensor_list) After the call, all 16 tensors on the two nodes will have the all-reduced value of 16 .. autofunction:: broadcast_multigpu .. autofunction:: all_reduce_multigpu .. autofunction:: reduce_multigpu .. autofunction:: all_gather_multigpu Launch utility -------------- The `torch.distributed.deprecated` package also provides a launch utility in `torch.distributed.deprecated.launch`. .. automodule:: torch.distributed.launch