Distributed communication package - torch.distributed¶
Note
Please refer to PyTorch Distributed Overview for a brief introduction to all features related to distributed training.
Backends¶
torch.distributed
supports three built-in 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 |
|
|
|
|||
---|---|---|---|---|---|---|
Device |
CPU |
GPU |
CPU |
GPU |
CPU |
GPU |
send |
✓ |
✘ |
✓ |
? |
✘ |
✓ |
recv |
✓ |
✘ |
✓ |
? |
✘ |
✓ |
broadcast |
✓ |
✓ |
✓ |
? |
✘ |
✓ |
all_reduce |
✓ |
✓ |
✓ |
? |
✘ |
✓ |
reduce |
✓ |
✘ |
✓ |
? |
✘ |
✓ |
all_gather |
✓ |
✘ |
✓ |
? |
✘ |
✓ |
gather |
✓ |
✘ |
✓ |
? |
✘ |
✓ |
scatter |
✓ |
✘ |
✓ |
? |
✘ |
✓ |
reduce_scatter |
✘ |
✘ |
✘ |
✘ |
✘ |
✓ |
all_to_all |
✘ |
✘ |
✓ |
? |
✘ |
✓ |
barrier |
✓ |
✘ |
✓ |
? |
✘ |
✓ |
Backends that come with PyTorch¶
PyTorch distributed package supports Linux (stable), MacOS (stable), and Windows (prototype). By default for Linux, the Gloo and NCCL backends are built and included in PyTorch distributed (NCCL only when building with CUDA). MPI is an optional backend that can only be included if you build PyTorch from source. (e.g. building PyTorch on a host that has MPI installed.)
Note
As of PyTorch v1.8, Windows supports all collective communications backend but NCCL,
If the init_method argument of init_process_group()
points to a file it must adhere
to the following schema:
Local file system,
init_method="file:///d:/tmp/some_file"
Shared file system,
init_method="file://////{machine_name}/{share_folder_name}/some_file"
Same as on Linux platform, you can enable TcpStore by setting environment variables, MASTER_ADDR and MASTER_PORT.
Which backend to use?¶
In the past, we were often asked: “which backend should I use?”.
Rule of thumb
Use the NCCL backend for distributed GPU training
Use the Gloo backend for distributed CPU training.
GPU hosts with InfiniBand interconnect
Use NCCL, since it’s the only backend that currently supports InfiniBand and GPUDirect.
GPU hosts with Ethernet interconnect
Use NCCL, since it currently provides the best distributed GPU training performance, especially for multiprocess single-node or multi-node distributed training. If you encounter any problem with NCCL, use Gloo as the fallback option. (Note that Gloo currently runs slower than NCCL for GPUs.)
CPU hosts with InfiniBand interconnect
If your InfiniBand has enabled IP over IB, use Gloo, otherwise, use MPI instead. We are planning on adding InfiniBand support for Gloo in the upcoming releases.
CPU hosts with Ethernet interconnect
Use Gloo, unless you have specific reasons to use MPI.
Common environment variables¶
Choosing the network interface to use¶
By default, both the NCCL and Gloo backends will try to find the right network interface to use. If the automatically detected interface is not correct, you can override it using the following environment variables (applicable to the respective backend):
NCCL_SOCKET_IFNAME, for example
export NCCL_SOCKET_IFNAME=eth0
GLOO_SOCKET_IFNAME, for example
export GLOO_SOCKET_IFNAME=eth0
If you’re using the Gloo backend, you can specify multiple interfaces by separating
them by a comma, like this: export GLOO_SOCKET_IFNAME=eth0,eth1,eth2,eth3
.
The backend will dispatch operations in a round-robin fashion across these interfaces.
It is imperative that all processes specify the same number of interfaces in this variable.
Other NCCL environment variables¶
Debugging - in case of NCCL failure, you can set NCCL_DEBUG=INFO
to print an explicit
warning message as well as basic NCCL initialization information.
You may also use NCCL_DEBUG_SUBSYS
to get more details about a specific
aspect of NCCL. For example, NCCL_DEBUG_SUBSYS=COLL
would print logs of
collective calls, which may be helpful when debugging hangs, especially those
caused by collective type or message size mismatch. In case of topology
detection failure, it would be helpful to set NCCL_DEBUG_SUBSYS=GRAPH
to inspect the detailed detection result and save as reference if further help
from NCCL team is needed.
Performance tuning - NCCL performs automatic tuning based on its topology detection to save users’
tuning effort. On some socket-based systems, users may still try tuning
NCCL_SOCKET_NTHREADS
and NCCL_NSOCKS_PERTHREAD
to increase socket
network bandwidth. These two environment variables have been pre-tuned by NCCL
for some cloud providers, such as AWS or GCP.
For a full list of NCCL environment variables, please refer to NVIDIA NCCL’s official documentation
Basics¶
The torch.distributed package provides PyTorch support and communication primitives
for multiprocess parallelism across several computation nodes running on one or more
machines. The class torch.nn.parallel.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
Multiprocessing package - torch.multiprocessing and 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 or the
torch.nn.parallel.DistributedDataParallel()
wrapper may still have advantages over other
approaches to data-parallelism, including 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 torch.distributed.init_process_group()
or torch.distributed.device_mesh.init_device_mesh()
function before calling any other methods.
Both block until all processes have joined.
- torch.distributed.is_available()[source]¶
Return
True
if the distributed package is available.Otherwise,
torch.distributed
does not expose any other APIs. Currently,torch.distributed
is available on Linux, MacOS and Windows. SetUSE_DISTRIBUTED=1
to enable it when building PyTorch from source. Currently, the default value isUSE_DISTRIBUTED=1
for Linux and Windows,USE_DISTRIBUTED=0
for MacOS.- Return type
- torch.distributed.init_process_group(backend=None, init_method=None, timeout=None, world_size=-1, rank=-1, store=None, group_name='', pg_options=None, device_id=None)[source]¶
Initialize the default distributed process group.
This will also initialize the distributed package.
- There are 2 main ways to initialize a process group:
Specify
store
,rank
, andworld_size
explicitly.Specify
init_method
(a URL string) which indicates where/how to discover peers. Optionally specifyrank
andworld_size
, or encode all required parameters in the URL and omit them.
If neither is specified,
init_method
is assumed to be “env://”.- Parameters
backend (str or Backend, optional) – The backend to use. Depending on build-time configurations, valid values include
mpi
,gloo
,nccl
, anducc
. If the backend is not provided, then both agloo
andnccl
backend will be created, see notes below for how multiple backends are managed. This field can be given as a lowercase string (e.g.,"gloo"
), which can also be accessed viaBackend
attributes (e.g.,Backend.GLOO
). If using multiple processes per machine withnccl
backend, each process must have exclusive access to every GPU it uses, as sharing GPUs between processes can result in deadlocks.ucc
backend is experimental.init_method (str, optional) – URL specifying how to initialize the process group. Default is “env://” if no
init_method
orstore
is specified. Mutually exclusive withstore
.world_size (int, optional) – Number of processes participating in the job. Required if
store
is specified.rank (int, optional) – Rank of the current process (it should be a number between 0 and
world_size
-1). Required ifstore
is specified.store (Store, optional) – Key/value store accessible to all workers, used to exchange connection/address information. Mutually exclusive with
init_method
.timeout (timedelta, optional) – Timeout for operations executed against the process group. Default value is 10 minutes for NCCL and 30 minutes for other backends. This is the duration after which collectives will be aborted asynchronously and the process will crash. This is done since CUDA execution is async and it is no longer safe to continue executing user code since failed async NCCL operations might result in subsequent CUDA operations running on corrupted data. When TORCH_NCCL_BLOCKING_WAIT is set, the process will block and wait for this timeout.
group_name (str, optional, deprecated) – Group name. This argument is ignored
pg_options (ProcessGroupOptions, optional) – process group options specifying what additional options need to be passed in during the construction of specific process groups. As of now, the only options we support is
ProcessGroupNCCL.Options
for thenccl
backend,is_high_priority_stream
can be specified so that the nccl backend can pick up high priority cuda streams when there’re compute kernels waiting.device_id (torch.device, optional) – a single, specific device to “bind” this process to, allowing for backend-specific optimizations. Currently this has two effects, only under NCCL: the communicator is immediately formed (calling
ncclCommInit*
immediately rather than the normal lazy call) and sub-groups will usencclCommSplit
when possible to avoid unnecessary overhead of group creation. If you want to know NCCL initialization error early, you can also use this field.
Note
To enable
backend == Backend.MPI
, PyTorch needs to be built from source on a system that supports MPI.Note
Support for multiple backends is experimental. Currently when no backend is specified, both
gloo
andnccl
backends will be created. Thegloo
backend will be used for collectives with CPU tensors and thenccl
backend will be used for collectives with CUDA tensors. A custom backend can be specified by passing in a string with format “<device_type>:<backend_name>,<device_type>:<backend_name>”, e.g. “cpu:gloo,cuda:custom_backend”.
- torch.distributed.device_mesh.init_device_mesh(device_type, mesh_shape, *, mesh_dim_names=None)[source]¶
Initializes a DeviceMesh based on device_type, mesh_shape, and mesh_dim_names parameters.
This creates a DeviceMesh with an n-dimensional array layout, where n is the length of mesh_shape. If mesh_dim_names is provided, each dimension is labeled as mesh_dim_names[i].
Note
init_device_mesh follows SPMD programming model, meaning the same PyTorch Python program runs on all processes/ranks in the cluster. Ensure mesh_shape (the dimensions of the nD array describing device layout) is identical across all ranks. Inconsistent mesh_shape may lead to hanging.
Note
If no process group is found, init_device_mesh will initialize distributed process group/groups required for distributed communications behind the scene.
- Parameters
device_type (str) – The device type of the mesh. Currently supports: “cpu”, “cuda/cuda-like”.
mesh_shape (Tuple[int]) – A tuple defining the dimensions of the multi-dimensional array describing the layout of devices.
mesh_dim_names (Tuple[str], optional) – A tuple of mesh dimension names to assign to each dimension of the multi-dimensional array describing the layout of devices. Its length must match the length of mesh_shape. Each string in mesh_dim_names must be unique.
- Returns
A
DeviceMesh
object representing the device layout.- Return type
- Example::
>>> from torch.distributed.device_mesh import init_device_mesh >>> >>> mesh_1d = init_device_mesh("cuda", mesh_shape=(8,)) >>> mesh_2d = init_device_mesh("cuda", mesh_shape=(2, 8), mesh_dim_names=("dp", "tp"))
- torch.distributed.is_initialized()[source]¶
Check if the default process group has been initialized.
- Return type
- torch.distributed.is_torchelastic_launched()[source]¶
Check whether this process was launched with
torch.distributed.elastic
(aka torchelastic).The existence of
TORCHELASTIC_RUN_ID
environment variable is used as a proxy to determine whether the current process was launched with torchelastic. This is a reasonable proxy sinceTORCHELASTIC_RUN_ID
maps to the rendezvous id which is always a non-null value indicating the job id for peer discovery purposes..- Return type
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.
Note that multicast address is not supported anymore in the latest distributed
package. group_name
is deprecated as well.
import torch.distributed 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)
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 0MASTER_ADDR
- required (except for rank 0); address of rank 0 nodeWORLD_SIZE
- required; can be set either here, or in a call to init functionRANK
- 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://
).
Post-Initialization¶
Once torch.distributed.init_process_group()
was run, the following functions can be used. To
check whether the process group has already been initialized use torch.distributed.is_initialized()
.
- class torch.distributed.Backend(name)[source]¶
An enum-like class for backends.
Available backends: GLOO, NCCL, UCC, MPI, and other registered backends.
The values of this class are lowercase strings, e.g.,
"gloo"
. They can be accessed as attributes, e.g.,Backend.NCCL
.This class can be directly called to parse the string, e.g.,
Backend(backend_str)
will check ifbackend_str
is valid, and return the parsed lowercase string if so. It also accepts uppercase strings, e.g.,Backend("GLOO")
returns"gloo"
.Note
The entry
Backend.UNDEFINED
is present but only used as initial value of some fields. Users should neither use it directly nor assume its existence.- classmethod register_backend(name, func, extended_api=False, devices=None)[source]¶
Register a new backend with the given name and instantiating function.
This class method is used by 3rd party
ProcessGroup
extension to register new backends.- Parameters
name (str) – Backend name of the
ProcessGroup
extension. It should match the one ininit_process_group()
.func (function) – Function handler that instantiates the backend. The function should be implemented in the backend extension and takes four arguments, including
store
,rank
,world_size
, andtimeout
.extended_api (bool, optional) – Whether the backend supports extended argument structure. Default:
False
. If set toTrue
, the backend will get an instance ofc10d::DistributedBackendOptions
, and a process group options object as defined by the backend implementation.device (str or list of str, optional) – device type this backend supports, e.g. “cpu”, “cuda”, etc. If None, assuming both “cpu” and “cuda”
Note
This support of 3rd party backend is experimental and subject to change.
- torch.distributed.get_backend(group=None)[source]¶
Return the backend of the given process group.
- Parameters
group (ProcessGroup, optional) – The process group to work on. The default is the general main process group. If another specific group is specified, the calling process must be part of
group
.- Returns
The backend of the given process group as a lower case string.
- Return type
- torch.distributed.get_rank(group=None)[source]¶
Return the rank of the current process in the provided
group
, default otherwise.Rank is a unique identifier assigned to each process within a distributed process group. They are always consecutive integers ranging from 0 to
world_size
.- Parameters
group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
- Returns
The rank of the process group -1, if not part of the group
- Return type
- torch.distributed.get_world_size(group=None)[source]¶
Return the number of processes in the current process group.
- Parameters
group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
- Returns
The world size of the process group -1, if not part of the group
- Return type
Distributed Key-Value Store¶
The distributed package comes with a distributed key-value store, which can be
used to share information between processes in the group as well as to
initialize the distributed package in
torch.distributed.init_process_group()
(by explicitly creating the store
as an alternative to specifying init_method
.) There are 3 choices for
Key-Value Stores: TCPStore
,
FileStore
, and HashStore
.
- class torch.distributed.Store¶
Base class for all store implementations, such as the 3 provided by PyTorch distributed: (
TCPStore
,FileStore
, andHashStore
).
- class torch.distributed.TCPStore¶
A TCP-based distributed key-value store implementation. The server store holds the data, while the client stores can connect to the server store over TCP and perform actions such as
set()
to insert a key-value pair,get()
to retrieve a key-value pair, etc. There should always be one server store initialized because the client store(s) will wait for the server to establish a connection.- Parameters
host_name (str) – The hostname or IP Address the server store should run on.
port (int) – The port on which the server store should listen for incoming requests.
world_size (int, optional) – The total number of store users (number of clients + 1 for the server). Default is None (None indicates a non-fixed number of store users).
is_master (bool, optional) – True when initializing the server store and False for client stores. Default is False.
timeout (timedelta, optional) – Timeout used by the store during initialization and for methods such as
get()
andwait()
. Default is timedelta(seconds=300)wait_for_workers (bool, optional) – Whether to wait for all the workers to connect with the server store. This is only applicable when world_size is a fixed value. Default is True.
multi_tenant (bool, optional) – If True, all
TCPStore
instances in the current process with the same host/port will use the same underlyingTCPServer
. Default is False.master_listen_fd (int, optional) – If specified, the underlying
TCPServer
will listen on this file descriptor, which must be a socket already bound toport
. Useful to avoid port assignment races in some scenarios. Default is None (meaning the server creates a new socket and attempts to bind it toport
).
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> # Run on process 1 (server) >>> server_store = dist.TCPStore("127.0.0.1", 1234, 2, True, timedelta(seconds=30)) >>> # Run on process 2 (client) >>> client_store = dist.TCPStore("127.0.0.1", 1234, 2, False) >>> # Use any of the store methods from either the client or server after initialization >>> server_store.set("first_key", "first_value") >>> client_store.get("first_key")
- class torch.distributed.HashStore¶
A thread-safe store implementation based on an underlying hashmap. This store can be used within the same process (for example, by other threads), but cannot be used across processes.
- Example::
>>> import torch.distributed as dist >>> store = dist.HashStore() >>> # store can be used from other threads >>> # Use any of the store methods after initialization >>> store.set("first_key", "first_value")
- class torch.distributed.FileStore¶
A store implementation that uses a file to store the underlying key-value pairs.
- Parameters
- Example::
>>> import torch.distributed as dist >>> store1 = dist.FileStore("/tmp/filestore", 2) >>> store2 = dist.FileStore("/tmp/filestore", 2) >>> # Use any of the store methods from either the client or server after initialization >>> store1.set("first_key", "first_value") >>> store2.get("first_key")
- class torch.distributed.PrefixStore¶
A wrapper around any of the 3 key-value stores (
TCPStore
,FileStore
, andHashStore
) that adds a prefix to each key inserted to the store.- Parameters
prefix (str) – The prefix string that is prepended to each key before being inserted into the store.
store (torch.distributed.store) – A store object that forms the underlying key-value store.
- torch.distributed.Store.set(self: torch._C._distributed_c10d.Store, arg0: str, arg1: str) None ¶
Inserts the key-value pair into the store based on the supplied
key
andvalue
. Ifkey
already exists in the store, it will overwrite the old value with the new suppliedvalue
.- Parameters
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> store.set("first_key", "first_value") >>> # Should return "first_value" >>> store.get("first_key")
- torch.distributed.Store.get(self: torch._C._distributed_c10d.Store, arg0: str) bytes ¶
Retrieves the value associated with the given
key
in the store. Ifkey
is not present in the store, the function will wait fortimeout
, which is defined when initializing the store, before throwing an exception.- Parameters
key (str) – The function will return the value associated with this key.
- Returns
Value associated with
key
ifkey
is in the store.
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> store.set("first_key", "first_value") >>> # Should return "first_value" >>> store.get("first_key")
- torch.distributed.Store.add(self: torch._C._distributed_c10d.Store, arg0: str, arg1: int) int ¶
The first call to add for a given
key
creates a counter associated withkey
in the store, initialized toamount
. Subsequent calls to add with the samekey
increment the counter by the specifiedamount
. Callingadd()
with a key that has already been set in the store byset()
will result in an exception.- Parameters
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> # Using TCPStore as an example, other store types can also be used >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> store.add("first_key", 1) >>> store.add("first_key", 6) >>> # Should return 7 >>> store.get("first_key")
- torch.distributed.Store.compare_set(self: torch._C._distributed_c10d.Store, arg0: str, arg1: str, arg2: str) bytes ¶
Inserts the key-value pair into the store based on the supplied
key
and performs comparison betweenexpected_value
anddesired_value
before inserting.desired_value
will only be set ifexpected_value
for thekey
already exists in the store or ifexpected_value
is an empty string.- Parameters
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> store.set("key", "first_value") >>> store.compare_set("key", "first_value", "second_value") >>> # Should return "second_value" >>> store.get("key")
- torch.distributed.Store.wait(*args, **kwargs)¶
Overloaded function.
wait(self: torch._C._distributed_c10d.Store, arg0: list[str]) -> None
Waits for each key in
keys
to be added to the store. If not all keys are set before thetimeout
(set during store initialization), thenwait
will throw an exception.- Parameters
keys (list) – List of keys on which to wait until they are set in the store.
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> # Using TCPStore as an example, other store types can also be used >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> # This will throw an exception after 30 seconds >>> store.wait(["bad_key"])
wait(self: torch._C._distributed_c10d.Store, arg0: list[str], arg1: datetime.timedelta) -> None
Waits for each key in
keys
to be added to the store, and throws an exception if the keys have not been set by the suppliedtimeout
.- Parameters
keys (list) – List of keys on which to wait until they are set in the store.
timeout (timedelta) – Time to wait for the keys to be added before throwing an exception.
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> # Using TCPStore as an example, other store types can also be used >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> # This will throw an exception after 10 seconds >>> store.wait(["bad_key"], timedelta(seconds=10))
- torch.distributed.Store.num_keys(self: torch._C._distributed_c10d.Store) int ¶
Returns the number of keys set in the store. Note that this number will typically be one greater than the number of keys added by
set()
andadd()
since one key is used to coordinate all the workers using the store.Warning
When used with the
TCPStore
,num_keys
returns the number of keys written to the underlying file. If the store is destructed and another store is created with the same file, the original keys will be retained.- Returns
The number of keys present in the store.
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> # Using TCPStore as an example, other store types can also be used >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> store.set("first_key", "first_value") >>> # This should return 2 >>> store.num_keys()
- torch.distributed.Store.delete_key(self: torch._C._distributed_c10d.Store, arg0: str) bool ¶
Deletes the key-value pair associated with
key
from the store. Returns true if the key was successfully deleted, and false if it was not.Warning
The
delete_key
API is only supported by theTCPStore
andHashStore
. Using this API with theFileStore
will result in an exception.- Parameters
key (str) – The key to be deleted from the store
- Returns
True if
key
was deleted, otherwise False.
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> # Using TCPStore as an example, HashStore can also be used >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> store.set("first_key") >>> # This should return true >>> store.delete_key("first_key") >>> # This should return false >>> store.delete_key("bad_key")
- torch.distributed.Store.set_timeout(self: torch._C._distributed_c10d.Store, arg0: datetime.timedelta) None ¶
Sets the store’s default timeout. This timeout is used during initialization and in
wait()
andget()
.- Parameters
timeout (timedelta) – timeout to be set in the store.
- Example::
>>> import torch.distributed as dist >>> from datetime import timedelta >>> # Using TCPStore as an example, other store types can also be used >>> store = dist.TCPStore("127.0.0.1", 0, 1, True, timedelta(seconds=30)) >>> store.set_timeout(timedelta(seconds=10)) >>> # This will throw an exception after 10 seconds >>> store.wait(["bad_key"])
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. 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).
- torch.distributed.new_group(ranks=None, timeout=None, backend=None, pg_options=None, use_local_synchronization=False)[source]¶
Create a new distributed group.
This function requires that all processes in the main group (i.e. all processes that are part of the distributed job) enter this function, even if they are not going to be members of the group. Additionally, groups should be created in the same order in all processes.
Warning
Using multiple process groups with the
NCCL
backend concurrently is not safe and the user should perform explicit synchronization in their application to ensure only one process group is used at a time. This means collectives from one process group should have completed execution on the device (not just enqueued since CUDA execution is async) before collectives from another process group are enqueued. See Using multiple NCCL communicators concurrently for more details.- Parameters
ranks (list[int]) – List of ranks of group members. If
None
, will be set to all ranks. Default isNone
.timeout (timedelta, optional) – see init_process_group for details and default value.
backend (str or Backend, optional) – The backend to use. Depending on build-time configurations, valid values are
gloo
andnccl
. By default uses the same backend as the global group. This field should be given as a lowercase string (e.g.,"gloo"
), which can also be accessed viaBackend
attributes (e.g.,Backend.GLOO
). IfNone
is passed in, the backend corresponding to the default process group will be used. Default isNone
.pg_options (ProcessGroupOptions, optional) – process group options specifying what additional options need to be passed in during the construction of specific process groups. i.e. for the
nccl
backend,is_high_priority_stream
can be specified so that process group can pick up high priority cuda streams.use_local_synchronization (bool, optional) – perform a group-local barrier at the end of the process group creation. This is different in that non-member ranks don’t need to call into API and don’t join the barrier.
- Returns
A handle of distributed group that can be given to collective calls or None if the rank is not part of
ranks
.
N.B. use_local_synchronization doesn’t work with MPI.
N.B. While use_local_synchronization=True can be significantly faster with larger clusters and small process groups, care must be taken since it changes cluster behavior as non-member ranks don’t join the group barrier().
N.B. use_local_synchronization=True can lead to deadlocks when each rank creates multiple overlaping process groups. To avoid that, make sure all ranks follow the same global creation order.
- torch.distributed.get_group_rank(group, global_rank)[source]¶
Translate a global rank into a group rank.
global_rank
must be part ofgroup
otherwise this raises RuntimeError.- Parameters
group (ProcessGroup) – ProcessGroup to find the relative rank.
global_rank (int) – Global rank to query.
- Returns
Group rank of
global_rank
relative togroup
- Return type
N.B. calling this function on the default process group returns identity
- torch.distributed.get_global_rank(group, group_rank)[source]¶
Translate a group rank into a global rank.
group_rank
must be part of group otherwise this raises RuntimeError.- Parameters
group (ProcessGroup) – ProcessGroup to find the global rank from.
group_rank (int) – Group rank to query.
- Returns
Global rank of
group_rank
relative togroup
- Return type
N.B. calling this function on the default process group returns identity
DeviceMesh¶
DeviceMesh is a higher level abstraction that manages process groups (or NCCL communicators).
It allows user to easily create inter node and intra node process groups without worrying about
how to set up the ranks correctly for different sub process groups, and it helps manage those
distributed process group easily. init_device_mesh()
function can be
used to create new DeviceMesh, with a mesh shape describing the device topology.
- class torch.distributed.device_mesh.DeviceMesh(device_type, mesh, *, mesh_dim_names=None, _init_backend=True)[source]¶
DeviceMesh represents a mesh of devices, where layout of devices could be represented as a n-d dimension array, and each value of the n-d dimensional array is the global id of the default process group ranks.
DeviceMesh could be used to describe the layout of devices across the cluster, and serves as a proxy for communication among the device lists within the cluster.
DeviceMesh can be used as a context manager.
Note
DeviceMesh follows SPMD programming model, which means the same PyTorch Python program is running on all processes/ranks in the cluster. Therefore, users need to make sure the mesh array (which describes the layout of devices) should be identical across all ranks. Inconsistent mesh will lead to silent hang.
- Parameters
device_type (str) – The device type of the mesh. Currently supports: “cpu”, “cuda/cuda-like”.
mesh (ndarray) – A multi-dimensional array or an integer tensor describing the layout of devices, where the IDs are global IDs of the default process group.
- Returns
A
DeviceMesh
object representing the device layout.- Return type
The following program runs on each process/rank in an SPMD manner. In this example, we have 2 hosts with 4 GPUs each. A reduction over the first dimension of mesh will reduce across columns (0, 4), .. and (3, 7), a reduction over the second dimension of mesh reduces across rows (0, 1, 2, 3) and (4, 5, 6, 7).
- Example::
>>> from torch.distributed.device_mesh import DeviceMesh >>> >>> # Initialize device mesh as (2, 4) to represent the topology >>> # of cross-host(dim 0), and within-host (dim 1). >>> mesh = DeviceMesh(device_type="cuda", mesh=[[0, 1, 2, 3],[4, 5, 6, 7]])
Point-to-point communication¶
- torch.distributed.send(tensor, dst, group=None, tag=0)[source]¶
Send a tensor synchronously.
- Parameters
tensor (Tensor) – Tensor to send.
dst (int) – Destination rank on global process group (regardless of
group
argument). Destination rank should not be the same as the rank of the current process.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
tag (int, optional) – Tag to match send with remote recv
- torch.distributed.recv(tensor, src=None, group=None, tag=0)[source]¶
Receives a tensor synchronously.
- Parameters
tensor (Tensor) – Tensor to fill with received data.
src (int, optional) – Source rank on global process group (regardless of
group
argument). Will receive from any process if unspecified.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
tag (int, optional) – Tag to match recv with remote send
- Returns
Sender rank -1, if not part of the group
- Return type
isend()
and 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 finishedwait()
- will block the process until the operation is finished.is_completed()
is guaranteed to return True once it returns.
- torch.distributed.isend(tensor, dst, group=None, tag=0)[source]¶
Send a tensor asynchronously.
Warning
Modifying
tensor
before the request completes causes undefined behavior.Warning
tag
is not supported with the NCCL backend.- Parameters
- Returns
A distributed request object. None, if not part of the group
- Return type
- torch.distributed.irecv(tensor, src=None, group=None, tag=0)[source]¶
Receives a tensor asynchronously.
Warning
tag
is not supported with the NCCL backend.- Parameters
tensor (Tensor) – Tensor to fill with received data.
src (int, optional) – Source rank on global process group (regardless of
group
argument). Will receive from any process if unspecified.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
tag (int, optional) – Tag to match recv with remote send
- Returns
A distributed request object. None, if not part of the group
- Return type
- torch.distributed.batch_isend_irecv(p2p_op_list)[source]¶
Send or Receive a batch of tensors asynchronously and return a list of requests.
Process each of the operations in
p2p_op_list
and return the corresponding requests. NCCL, Gloo, and UCC backend are currently supported.- Parameters
p2p_op_list – A list of point-to-point operations(type of each operator is
torch.distributed.P2POp
). The order of the isend/irecv in the list matters and it needs to match with corresponding isend/irecv on the remote end.- Returns
A list of distributed request objects returned by calling the corresponding op in the op_list.
Examples
>>> send_tensor = torch.arange(2, dtype=torch.float32) + 2 * rank >>> recv_tensor = torch.randn(2, dtype=torch.float32) >>> send_op = dist.P2POp(dist.isend, send_tensor, (rank + 1)%world_size) >>> recv_op = dist.P2POp(dist.irecv, recv_tensor, (rank - 1 + world_size)%world_size) >>> reqs = batch_isend_irecv([send_op, recv_op]) >>> for req in reqs: >>> req.wait() >>> recv_tensor tensor([2, 3]) # Rank 0 tensor([0, 1]) # Rank 1
Note
Note that when this API is used with the NCCL PG backend, users must set the current GPU device with torch.cuda.set_device, otherwise it will lead to unexpected hang issues.
In addition, if this API is the first collective call in the
group
passed todist.P2POp
, all ranks of thegroup
must participate in this API call; otherwise, the behavior is undefined. If this API call is not the first collective call in thegroup
, batched P2P operations involving only a subset of ranks of thegroup
are allowed.
- class torch.distributed.P2POp(op, tensor, peer, group=None, tag=0)[source]¶
A class to build point-to-point operations for
batch_isend_irecv
.This class builds the type of P2P operation, communication buffer, peer rank, Process Group, and tag. Instances of this class will be passed to
batch_isend_irecv
for point-to-point communications.- Parameters
op (Callable) – A function to send data to or receive data from a peer process. The type of
op
is eithertorch.distributed.isend
ortorch.distributed.irecv
.tensor (Tensor) – Tensor to send or receive.
peer (int) – Destination or source rank.
group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
tag (int, optional) – Tag to match send with recv.
Synchronous and asynchronous collective operations¶
Every collective operation function supports the following two kinds of operations,
depending on the setting of the async_op
flag passed into the collective:
Synchronous operation - the default mode, when async_op
is set to False
.
When the function returns, it is guaranteed that
the collective operation is performed. In the case of CUDA operations, it is not guaranteed
that the CUDA operation is completed, since CUDA operations are asynchronous. For CPU collectives, any
further function calls utilizing the output of the collective call will behave as expected. For CUDA collectives,
function calls utilizing the output on the same CUDA stream will behave as expected. Users must take care of
synchronization under the scenario of running under different streams. For details on CUDA semantics such as stream
synchronization, see CUDA Semantics.
See the below script to see examples of differences in these semantics for CPU and CUDA operations.
Asynchronous operation - when async_op
is set to True. The collective operation function
returns a distributed request object. In general, you don’t need to create it manually and it
is guaranteed to support two methods:
is_completed()
- in the case of CPU collectives, returnsTrue
if completed. In the case of CUDA operations, returnsTrue
if the operation has been successfully enqueued onto a CUDA stream and the output can be utilized on the default stream without further synchronization.wait()
- in the case of CPU collectives, will block the process until the operation is completed. In the case of CUDA collectives, will block until the operation has been successfully enqueued onto a CUDA stream and the output can be utilized on the default stream without further synchronization.get_future()
- returnstorch._C.Future
object. Supported for NCCL, also supported for most operations on GLOO and MPI, except for peer to peer operations. Note: as we continue adopting Futures and merging APIs,get_future()
call might become redundant.
Example
The following code can serve as a reference regarding semantics for CUDA operations when using distributed collectives. It shows the explicit need to synchronize when using collective outputs on different CUDA streams:
# Code runs on each rank.
dist.init_process_group("nccl", rank=rank, world_size=2)
output = torch.tensor([rank]).cuda(rank)
s = torch.cuda.Stream()
handle = dist.all_reduce(output, async_op=True)
# Wait ensures the operation is enqueued, but not necessarily complete.
handle.wait()
# Using result on non-default stream.
with torch.cuda.stream(s):
s.wait_stream(torch.cuda.default_stream())
output.add_(100)
if rank == 0:
# if the explicit call to wait_stream was omitted, the output below will be
# non-deterministically 1 or 101, depending on whether the allreduce overwrote
# the value after the add completed.
print(output)
Collective functions¶
- torch.distributed.broadcast(tensor, src, group=None, async_op=False)[source]¶
Broadcasts the tensor to the whole group.
tensor
must have the same number of elements in all processes participating in the collective.- Parameters
tensor (Tensor) – Data to be sent if
src
is the rank of current process, and tensor to be used to save received data otherwise.src (int) – Source rank on global process group (regardless of
group
argument).group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group
- torch.distributed.broadcast_object_list(object_list, src=0, group=None, device=None)[source]¶
Broadcasts picklable objects in
object_list
to the whole group.Similar to
broadcast()
, but Python objects can be passed in. Note that all objects inobject_list
must be picklable in order to be broadcasted.- Parameters
object_list (List[Any]) – List of input objects to broadcast. Each object must be picklable. Only objects on the
src
rank will be broadcast, but each rank must provide lists of equal sizes.src (int) – Source rank from which to broadcast
object_list
. Source rank is based on global process group (regardless ofgroup
argument)group – (ProcessGroup, optional): The process group to work on. If None, the default process group will be used. Default is
None
.device (
torch.device
, optional) – If not None, the objects are serialized and converted to tensors which are moved to thedevice
before broadcasting. Default isNone
.
- Returns
None
. If rank is part of the group,object_list
will contain the broadcasted objects fromsrc
rank.
Note
For NCCL-based process groups, internal tensor representations of objects must be moved to the GPU device before communication takes place. In this case, the device used is given by
torch.cuda.current_device()
and it is the user’s responsibility to ensure that this is set so that each rank has an individual GPU, viatorch.cuda.set_device()
.Note
Note that this API differs slightly from the
all_gather()
collective since it does not provide anasync_op
handle and thus will be a blocking call.Warning
broadcast_object_list()
usespickle
module implicitly, which is known to be insecure. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling. Only call this function with data you trust.Warning
Calling
broadcast_object_list()
with GPU tensors is not well supported and inefficient as it incurs GPU -> CPU transfer since tensors would be pickled. Please consider usingbroadcast()
instead.- Example::
>>> # Note: Process group initialization omitted on each rank. >>> import torch.distributed as dist >>> if dist.get_rank() == 0: >>> # Assumes world_size of 3. >>> objects = ["foo", 12, {1: 2}] # any picklable object >>> else: >>> objects = [None, None, None] >>> # Assumes backend is not NCCL >>> device = torch.device("cpu") >>> dist.broadcast_object_list(objects, src=0, device=device) >>> objects ['foo', 12, {1: 2}]
- torch.distributed.all_reduce(tensor, op=<RedOpType.SUM: 0>, group=None, async_op=False)[source]¶
Reduces the tensor data across all machines in a way that all get the final result.
After the call
tensor
is going to be bitwise identical in all processes.Complex tensors are supported.
- Parameters
tensor (Tensor) – Input and output of the collective. The function operates in-place.
op (optional) – One of the values from
torch.distributed.ReduceOp
enum. Specifies an operation used for element-wise reductions.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group
Examples
>>> # All tensors below are of torch.int64 type. >>> # We have 2 process groups, 2 ranks. >>> device = torch.device(f'cuda:{rank}') >>> tensor = torch.arange(2, dtype=torch.int64, device=device) + 1 + 2 * rank >>> tensor tensor([1, 2], device='cuda:0') # Rank 0 tensor([3, 4], device='cuda:1') # Rank 1 >>> dist.all_reduce(tensor, op=ReduceOp.SUM) >>> tensor tensor([4, 6], device='cuda:0') # Rank 0 tensor([4, 6], device='cuda:1') # Rank 1
>>> # All tensors below are of torch.cfloat type. >>> # We have 2 process groups, 2 ranks. >>> tensor = torch.tensor([1+1j, 2+2j], dtype=torch.cfloat, device=device) + 2 * rank * (1+1j) >>> tensor tensor([1.+1.j, 2.+2.j], device='cuda:0') # Rank 0 tensor([3.+3.j, 4.+4.j], device='cuda:1') # Rank 1 >>> dist.all_reduce(tensor, op=ReduceOp.SUM) >>> tensor tensor([4.+4.j, 6.+6.j], device='cuda:0') # Rank 0 tensor([4.+4.j, 6.+6.j], device='cuda:1') # Rank 1
- torch.distributed.reduce(tensor, dst, op=<RedOpType.SUM: 0>, group=None, async_op=False)[source]¶
Reduces the tensor data across all machines.
Only the process with rank
dst
is going to receive the final result.- Parameters
tensor (Tensor) – Input and output of the collective. The function operates in-place.
dst (int) – Destination rank on global process group (regardless of
group
argument)op (optional) – One of the values from
torch.distributed.ReduceOp
enum. Specifies an operation used for element-wise reductions.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group
- torch.distributed.all_gather(tensor_list, tensor, group=None, async_op=False)[source]¶
Gathers tensors from the whole group in a list.
Complex tensors are supported.
- Parameters
tensor_list (list[Tensor]) – Output list. It should contain correctly-sized tensors to be used for output of the collective.
tensor (Tensor) – Tensor to be broadcast from current process.
group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group
Examples
>>> # All tensors below are of torch.int64 dtype. >>> # We have 2 process groups, 2 ranks. >>> device = torch.device(f'cuda:{rank}') >>> tensor_list = [torch.zeros(2, dtype=torch.int64, device=device) for _ in range(2)] >>> tensor_list [tensor([0, 0], device='cuda:0'), tensor([0, 0], device='cuda:0')] # Rank 0 [tensor([0, 0], device='cuda:0'), tensor([0, 0], device='cuda:1')] # Rank 1 >>> tensor = torch.arange(2, dtype=torch.int64, device=device) + 1 + 2 * rank >>> tensor tensor([1, 2], device='cuda:0') # Rank 0 tensor([3, 4], device='cuda:1') # Rank 1 >>> dist.all_gather(tensor_list, tensor) >>> tensor_list [tensor([1, 2], device='cuda:0'), tensor([3, 4], device='cuda:0')] # Rank 0 [tensor([1, 2], device='cuda:1'), tensor([3, 4], device='cuda:1')] # Rank 1
>>> # All tensors below are of torch.cfloat dtype. >>> # We have 2 process groups, 2 ranks. >>> tensor_list = [torch.zeros(2, dtype=torch.cfloat, device=device) for _ in range(2)] >>> tensor_list [tensor([0.+0.j, 0.+0.j], device='cuda:0'), tensor([0.+0.j, 0.+0.j], device='cuda:0')] # Rank 0 [tensor([0.+0.j, 0.+0.j], device='cuda:1'), tensor([0.+0.j, 0.+0.j], device='cuda:1')] # Rank 1 >>> tensor = torch.tensor([1+1j, 2+2j], dtype=torch.cfloat, device=device) + 2 * rank * (1+1j) >>> tensor tensor([1.+1.j, 2.+2.j], device='cuda:0') # Rank 0 tensor([3.+3.j, 4.+4.j], device='cuda:1') # Rank 1 >>> dist.all_gather(tensor_list, tensor) >>> tensor_list [tensor([1.+1.j, 2.+2.j], device='cuda:0'), tensor([3.+3.j, 4.+4.j], device='cuda:0')] # Rank 0 [tensor([1.+1.j, 2.+2.j], device='cuda:1'), tensor([3.+3.j, 4.+4.j], device='cuda:1')] # Rank 1
- torch.distributed.all_gather_into_tensor(output_tensor, input_tensor, group=None, async_op=False)[source]¶
Gather tensors from all ranks and put them in a single output tensor.
- Parameters
output_tensor (Tensor) – Output tensor to accommodate tensor elements from all ranks. It must be correctly sized to have one of the following forms: (i) a concatenation of all the input tensors along the primary dimension; for definition of “concatenation”, see
torch.cat()
; (ii) a stack of all the input tensors along the primary dimension; for definition of “stack”, seetorch.stack()
. Examples below may better explain the supported output forms.input_tensor (Tensor) – Tensor to be gathered from current rank. Different from the
all_gather
API, the input tensors in this API must have the same size across all ranks.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group
Examples
>>> # All tensors below are of torch.int64 dtype and on CUDA devices. >>> # We have two ranks. >>> device = torch.device(f'cuda:{rank}') >>> tensor_in = torch.arange(2, dtype=torch.int64, device=device) + 1 + 2 * rank >>> tensor_in tensor([1, 2], device='cuda:0') # Rank 0 tensor([3, 4], device='cuda:1') # Rank 1 >>> # Output in concatenation form >>> tensor_out = torch.zeros(world_size * 2, dtype=torch.int64, device=device) >>> dist.all_gather_into_tensor(tensor_out, tensor_in) >>> tensor_out tensor([1, 2, 3, 4], device='cuda:0') # Rank 0 tensor([1, 2, 3, 4], device='cuda:1') # Rank 1 >>> # Output in stack form >>> tensor_out2 = torch.zeros(world_size, 2, dtype=torch.int64, device=device) >>> dist.all_gather_into_tensor(tensor_out2, tensor_in) >>> tensor_out2 tensor([[1, 2], [3, 4]], device='cuda:0') # Rank 0 tensor([[1, 2], [3, 4]], device='cuda:1') # Rank 1
Warning
The Gloo backend does not support this API.
- torch.distributed.all_gather_object(object_list, obj, group=None)[source]¶
Gathers picklable objects from the whole group into a list.
Similar to
all_gather()
, but Python objects can be passed in. Note that the object must be picklable in order to be gathered.- Parameters
object_list (list[Any]) – Output list. It should be correctly sized as the size of the group for this collective and will contain the output.
obj (Any) – Pickable Python object to be broadcast from current process.
group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used. Default is
None
.
- Returns
None. If the calling rank is part of this group, the output of the collective will be populated into the input
object_list
. If the calling rank is not part of the group, the passed inobject_list
will be unmodified.
Note
Note that this API differs slightly from the
all_gather()
collective since it does not provide anasync_op
handle and thus will be a blocking call.Note
For NCCL-based processed groups, internal tensor representations of objects must be moved to the GPU device before communication takes place. In this case, the device used is given by
torch.cuda.current_device()
and it is the user’s responsiblity to ensure that this is set so that each rank has an individual GPU, viatorch.cuda.set_device()
.Warning
all_gather_object()
usespickle
module implicitly, which is known to be insecure. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling. Only call this function with data you trust.Warning
Calling
all_gather_object()
with GPU tensors is not well supported and inefficient as it incurs GPU -> CPU transfer since tensors would be pickled. Please consider usingall_gather()
instead.- Example::
>>> # Note: Process group initialization omitted on each rank. >>> import torch.distributed as dist >>> # Assumes world_size of 3. >>> gather_objects = ["foo", 12, {1: 2}] # any picklable object >>> output = [None for _ in gather_objects] >>> dist.all_gather_object(output, gather_objects[dist.get_rank()]) >>> output ['foo', 12, {1: 2}]
- torch.distributed.gather(tensor, gather_list=None, dst=0, group=None, async_op=False)[source]¶
Gathers a list of tensors in a single process.
- Parameters
tensor (Tensor) – Input tensor.
gather_list (list[Tensor], optional) – List of appropriately-sized tensors to use for gathered data (default is None, must be specified on the destination rank)
dst (int, optional) – Destination rank on global process group (regardless of
group
argument). (default is 0)group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group
- torch.distributed.gather_object(obj, object_gather_list=None, dst=0, group=None)[source]¶
Gathers picklable objects from the whole group in a single process.
Similar to
gather()
, but Python objects can be passed in. Note that the object must be picklable in order to be gathered.- Parameters
obj (Any) – Input object. Must be picklable.
object_gather_list (list[Any]) – Output list. On the
dst
rank, it should be correctly sized as the size of the group for this collective and will contain the output. Must beNone
on non-dst ranks. (default isNone
)dst (int, optional) – Destination rank on global process group (regardless of
group
argument). (default is 0)group – (ProcessGroup, optional): The process group to work on. If None, the default process group will be used. Default is
None
.
- Returns
None. On the
dst
rank,object_gather_list
will contain the output of the collective.
Note
Note that this API differs slightly from the gather collective since it does not provide an async_op handle and thus will be a blocking call.
Note
For NCCL-based processed groups, internal tensor representations of objects must be moved to the GPU device before communication takes place. In this case, the device used is given by
torch.cuda.current_device()
and it is the user’s responsiblity to ensure that this is set so that each rank has an individual GPU, viatorch.cuda.set_device()
.Warning
gather_object()
usespickle
module implicitly, which is known to be insecure. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling. Only call this function with data you trust.Warning
Calling
gather_object()
with GPU tensors is not well supported and inefficient as it incurs GPU -> CPU transfer since tensors would be pickled. Please consider usinggather()
instead.- Example::
>>> # Note: Process group initialization omitted on each rank. >>> import torch.distributed as dist >>> # Assumes world_size of 3. >>> gather_objects = ["foo", 12, {1: 2}] # any picklable object >>> output = [None for _ in gather_objects] >>> dist.gather_object( ... gather_objects[dist.get_rank()], ... output if dist.get_rank() == 0 else None, ... dst=0 ... ) >>> # On rank 0 >>> output ['foo', 12, {1: 2}]
- torch.distributed.scatter(tensor, scatter_list=None, src=0, group=None, async_op=False)[source]¶
Scatters a list of tensors to all processes in a group.
Each process will receive exactly one tensor and store its data in the
tensor
argument.Complex tensors are supported.
- Parameters
tensor (Tensor) – Output tensor.
scatter_list (list[Tensor]) – List of tensors to scatter (default is None, must be specified on the source rank)
src (int) – Source rank on global process group (regardless of
group
argument). Default is 0group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group
Note
Note that all Tensors in scatter_list must have the same size.
- Example::
>>> # Note: Process group initialization omitted on each rank. >>> import torch.distributed as dist >>> tensor_size = 2 >>> t_ones = torch.ones(tensor_size) >>> t_fives = torch.ones(tensor_size) * 5 >>> output_tensor = torch.zeros(tensor_size) >>> if dist.get_rank() == 0: >>> # Assumes world_size of 2. >>> # Only tensors, all of which must be the same size. >>> scatter_list = [t_ones, t_fives] >>> else: >>> scatter_list = None >>> dist.scatter(output_tensor, scatter_list, src=0) >>> # Rank i gets scatter_list[i]. For example, on rank 1: >>> output_tensor tensor([5., 5.])
- torch.distributed.scatter_object_list(scatter_object_output_list, scatter_object_input_list, src=0, group=None)[source]¶
Scatters picklable objects in
scatter_object_input_list
to the whole group.Similar to
scatter()
, but Python objects can be passed in. On each rank, the scattered object will be stored as the first element ofscatter_object_output_list
. Note that all objects inscatter_object_input_list
must be picklable in order to be scattered.- Parameters
scatter_object_output_list (List[Any]) – Non-empty list whose first element will store the object scattered to this rank.
scatter_object_input_list (List[Any]) – List of input objects to scatter. Each object must be picklable. Only objects on the
src
rank will be scattered, and the argument can beNone
for non-src ranks.src (int) – Source rank from which to scatter
scatter_object_input_list
. Source rank is based on global process group (regardless ofgroup
argument).group – (ProcessGroup, optional): The process group to work on. If None, the default process group will be used. Default is
None
.
- Returns
None
. If rank is part of the group,scatter_object_output_list
will have its first element set to the scattered object for this rank.
Note
Note that this API differs slightly from the scatter collective since it does not provide an
async_op
handle and thus will be a blocking call.Warning
scatter_object_list()
usespickle
module implicitly, which is known to be insecure. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling. Only call this function with data you trust.Warning
Calling
scatter_object_list()
with GPU tensors is not well supported and inefficient as it incurs GPU -> CPU transfer since tensors would be pickled. Please consider usingscatter()
instead.- Example::
>>> # Note: Process group initialization omitted on each rank. >>> import torch.distributed as dist >>> if dist.get_rank() == 0: >>> # Assumes world_size of 3. >>> objects = ["foo", 12, {1: 2}] # any picklable object >>> else: >>> # Can be any list on non-src ranks, elements are not used. >>> objects = [None, None, None] >>> output_list = [None] >>> dist.scatter_object_list(output_list, objects, src=0) >>> # Rank i gets objects[i]. For example, on rank 2: >>> output_list [{1: 2}]
- torch.distributed.reduce_scatter(output, input_list, op=<RedOpType.SUM: 0>, group=None, async_op=False)[source]¶
Reduces, then scatters a list of tensors to all processes in a group.
- Parameters
output (Tensor) – Output tensor.
input_list (list[Tensor]) – List of tensors to reduce and scatter.
op (optional) – One of the values from
torch.distributed.ReduceOp
enum. Specifies an operation used for element-wise reductions.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op.
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group.
- torch.distributed.reduce_scatter_tensor(output, input, op=<RedOpType.SUM: 0>, group=None, async_op=False)[source]¶
Reduces, then scatters a tensor to all ranks in a group.
- Parameters
output (Tensor) – Output tensor. It should have the same size across all ranks.
input (Tensor) – Input tensor to be reduced and scattered. Its size should be output tensor size times the world size. The input tensor can have one of the following shapes: (i) a concatenation of the output tensors along the primary dimension, or (ii) a stack of the output tensors along the primary dimension. For definition of “concatenation”, see
torch.cat()
. For definition of “stack”, seetorch.stack()
.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op.
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group.
Examples
>>> # All tensors below are of torch.int64 dtype and on CUDA devices. >>> # We have two ranks. >>> device = torch.device(f'cuda:{rank}') >>> tensor_out = torch.zeros(2, dtype=torch.int64, device=device) >>> # Input in concatenation form >>> tensor_in = torch.arange(world_size * 2, dtype=torch.int64, device=device) >>> tensor_in tensor([0, 1, 2, 3], device='cuda:0') # Rank 0 tensor([0, 1, 2, 3], device='cuda:1') # Rank 1 >>> dist.reduce_scatter_tensor(tensor_out, tensor_in) >>> tensor_out tensor([0, 2], device='cuda:0') # Rank 0 tensor([4, 6], device='cuda:1') # Rank 1 >>> # Input in stack form >>> tensor_in = torch.reshape(tensor_in, (world_size, 2)) >>> tensor_in tensor([[0, 1], [2, 3]], device='cuda:0') # Rank 0 tensor([[0, 1], [2, 3]], device='cuda:1') # Rank 1 >>> dist.reduce_scatter_tensor(tensor_out, tensor_in) >>> tensor_out tensor([0, 2], device='cuda:0') # Rank 0 tensor([4, 6], device='cuda:1') # Rank 1
Warning
The Gloo backend does not support this API.
- torch.distributed.all_to_all_single(output, input, output_split_sizes=None, input_split_sizes=None, group=None, async_op=False)[source]¶
Split input tensor and then scatter the split list to all processes in a group.
Later the received tensors are concatenated from all the processes in the group and returned as a single output tensor.
Complex tensors are supported.
- Parameters
output (Tensor) – Gathered concatenated output tensor.
input (Tensor) – Input tensor to scatter.
output_split_sizes – (list[Int], optional): Output split sizes for dim 0 if specified None or empty, dim 0 of
output
tensor must divide equally byworld_size
.input_split_sizes – (list[Int], optional): Input split sizes for dim 0 if specified None or empty, dim 0 of
input
tensor must divide equally byworld_size
.group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op.
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group.
Warning
all_to_all_single is experimental and subject to change.
Examples
>>> input = torch.arange(4) + rank * 4 >>> input tensor([0, 1, 2, 3]) # Rank 0 tensor([4, 5, 6, 7]) # Rank 1 tensor([8, 9, 10, 11]) # Rank 2 tensor([12, 13, 14, 15]) # Rank 3 >>> output = torch.empty([4], dtype=torch.int64) >>> dist.all_to_all_single(output, input) >>> output tensor([0, 4, 8, 12]) # Rank 0 tensor([1, 5, 9, 13]) # Rank 1 tensor([2, 6, 10, 14]) # Rank 2 tensor([3, 7, 11, 15]) # Rank 3
>>> # Essentially, it is similar to following operation: >>> scatter_list = list(input.chunk(world_size)) >>> gather_list = list(output.chunk(world_size)) >>> for i in range(world_size): >>> dist.scatter(gather_list[i], scatter_list if i == rank else [], src = i)
>>> # Another example with uneven split >>> input tensor([0, 1, 2, 3, 4, 5]) # Rank 0 tensor([10, 11, 12, 13, 14, 15, 16, 17, 18]) # Rank 1 tensor([20, 21, 22, 23, 24]) # Rank 2 tensor([30, 31, 32, 33, 34, 35, 36]) # Rank 3 >>> input_splits [2, 2, 1, 1] # Rank 0 [3, 2, 2, 2] # Rank 1 [2, 1, 1, 1] # Rank 2 [2, 2, 2, 1] # Rank 3 >>> output_splits [2, 3, 2, 2] # Rank 0 [2, 2, 1, 2] # Rank 1 [1, 2, 1, 2] # Rank 2 [1, 2, 1, 1] # Rank 3 >>> output = ... >>> dist.all_to_all_single(output, input, output_splits, input_splits) >>> output tensor([ 0, 1, 10, 11, 12, 20, 21, 30, 31]) # Rank 0 tensor([ 2, 3, 13, 14, 22, 32, 33]) # Rank 1 tensor([ 4, 15, 16, 23, 34, 35]) # Rank 2 tensor([ 5, 17, 18, 24, 36]) # Rank 3
>>> # Another example with tensors of torch.cfloat type. >>> input = torch.tensor([1+1j, 2+2j, 3+3j, 4+4j], dtype=torch.cfloat) + 4 * rank * (1+1j) >>> input tensor([1+1j, 2+2j, 3+3j, 4+4j]) # Rank 0 tensor([5+5j, 6+6j, 7+7j, 8+8j]) # Rank 1 tensor([9+9j, 10+10j, 11+11j, 12+12j]) # Rank 2 tensor([13+13j, 14+14j, 15+15j, 16+16j]) # Rank 3 >>> output = torch.empty([4], dtype=torch.int64) >>> dist.all_to_all_single(output, input) >>> output tensor([1+1j, 5+5j, 9+9j, 13+13j]) # Rank 0 tensor([2+2j, 6+6j, 10+10j, 14+14j]) # Rank 1 tensor([3+3j, 7+7j, 11+11j, 15+15j]) # Rank 2 tensor([4+4j, 8+8j, 12+12j, 16+16j]) # Rank 3
- torch.distributed.all_to_all(output_tensor_list, input_tensor_list, group=None, async_op=False)[source]¶
Scatters list of input tensors to all processes in a group and return gathered list of tensors in output list.
Complex tensors are supported.
- Parameters
output_tensor_list (list[Tensor]) – List of tensors to be gathered one per rank.
input_tensor_list (list[Tensor]) – List of tensors to scatter one per rank.
group (ProcessGroup, optional) – The process group to work on. If None, the default process group will be used.
async_op (bool, optional) – Whether this op should be an async op.
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group.
Warning
all_to_all is experimental and subject to change.
Examples
>>> input = torch.arange(4) + rank * 4 >>> input = list(input.chunk(4)) >>> input [tensor([0]), tensor([1]), tensor([2]), tensor([3])] # Rank 0 [tensor([4]), tensor([5]), tensor([6]), tensor([7])] # Rank 1 [tensor([8]), tensor([9]), tensor([10]), tensor([11])] # Rank 2 [tensor([12]), tensor([13]), tensor([14]), tensor([15])] # Rank 3 >>> output = list(torch.empty([4], dtype=torch.int64).chunk(4)) >>> dist.all_to_all(output, input) >>> output [tensor([0]), tensor([4]), tensor([8]), tensor([12])] # Rank 0 [tensor([1]), tensor([5]), tensor([9]), tensor([13])] # Rank 1 [tensor([2]), tensor([6]), tensor([10]), tensor([14])] # Rank 2 [tensor([3]), tensor([7]), tensor([11]), tensor([15])] # Rank 3
>>> # Essentially, it is similar to following operation: >>> scatter_list = input >>> gather_list = output >>> for i in range(world_size): >>> dist.scatter(gather_list[i], scatter_list if i == rank else [], src=i)
>>> input tensor([0, 1, 2, 3, 4, 5]) # Rank 0 tensor([10, 11, 12, 13, 14, 15, 16, 17, 18]) # Rank 1 tensor([20, 21, 22, 23, 24]) # Rank 2 tensor([30, 31, 32, 33, 34, 35, 36]) # Rank 3 >>> input_splits [2, 2, 1, 1] # Rank 0 [3, 2, 2, 2] # Rank 1 [2, 1, 1, 1] # Rank 2 [2, 2, 2, 1] # Rank 3 >>> output_splits [2, 3, 2, 2] # Rank 0 [2, 2, 1, 2] # Rank 1 [1, 2, 1, 2] # Rank 2 [1, 2, 1, 1] # Rank 3 >>> input = list(input.split(input_splits)) >>> input [tensor([0, 1]), tensor([2, 3]), tensor([4]), tensor([5])] # Rank 0 [tensor([10, 11, 12]), tensor([13, 14]), tensor([15, 16]), tensor([17, 18])] # Rank 1 [tensor([20, 21]), tensor([22]), tensor([23]), tensor([24])] # Rank 2 [tensor([30, 31]), tensor([32, 33]), tensor([34, 35]), tensor([36])] # Rank 3 >>> output = ... >>> dist.all_to_all(output, input) >>> output [tensor([0, 1]), tensor([10, 11, 12]), tensor([20, 21]), tensor([30, 31])] # Rank 0 [tensor([2, 3]), tensor([13, 14]), tensor([22]), tensor([32, 33])] # Rank 1 [tensor([4]), tensor([15, 16]), tensor([23]), tensor([34, 35])] # Rank 2 [tensor([5]), tensor([17, 18]), tensor([24]), tensor([36])] # Rank 3
>>> # Another example with tensors of torch.cfloat type. >>> input = torch.tensor([1+1j, 2+2j, 3+3j, 4+4j], dtype=torch.cfloat) + 4 * rank * (1+1j) >>> input = list(input.chunk(4)) >>> input [tensor([1+1j]), tensor([2+2j]), tensor([3+3j]), tensor([4+4j])] # Rank 0 [tensor([5+5j]), tensor([6+6j]), tensor([7+7j]), tensor([8+8j])] # Rank 1 [tensor([9+9j]), tensor([10+10j]), tensor([11+11j]), tensor([12+12j])] # Rank 2 [tensor([13+13j]), tensor([14+14j]), tensor([15+15j]), tensor([16+16j])] # Rank 3 >>> output = list(torch.empty([4], dtype=torch.int64).chunk(4)) >>> dist.all_to_all(output, input) >>> output [tensor([1+1j]), tensor([5+5j]), tensor([9+9j]), tensor([13+13j])] # Rank 0 [tensor([2+2j]), tensor([6+6j]), tensor([10+10j]), tensor([14+14j])] # Rank 1 [tensor([3+3j]), tensor([7+7j]), tensor([11+11j]), tensor([15+15j])] # Rank 2 [tensor([4+4j]), tensor([8+8j]), tensor([12+12j]), tensor([16+16j])] # Rank 3
- torch.distributed.barrier(group=None, async_op=False, device_ids=None)[source]¶
Synchronize all processes.
This collective blocks processes until the whole group enters this function, if async_op is False, or if async work handle is called on wait().
- Parameters
- Returns
Async work handle, if async_op is set to True. None, if not async_op or if not part of the group
Note
ProcessGroupNCCL now relies on stream synchronization instead of device synchronization to block the CPU. Thus, please do not assume that barrier() would perform a device synchronization.
- torch.distributed.monitored_barrier(group=None, timeout=None, wait_all_ranks=False)[source]¶
Synchronize processes similar to
torch.distributed.barrier
, but consider a configurable timeout.It is able to report ranks that did not pass this barrier within the provided timeout. Specifically, for non-zero ranks, will block until a send/recv is processed from rank 0. Rank 0 will block until all send /recv from other ranks are processed, and will report failures for ranks that failed to respond in time. Note that if one rank does not reach the monitored_barrier (for example due to a hang), all other ranks would fail in monitored_barrier.
This collective will block all processes/ranks in the group, until the whole group exits the function successfully, making it useful for debugging and synchronizing. However, it can have a performance impact and should only be used for debugging or scenarios that require full synchronization points on the host-side. For debugging purposes, this barrier can be inserted before the application’s collective calls to check if any ranks are desynchronized.
Note
Note that this collective is only supported with the GLOO backend.
- Parameters
group (ProcessGroup, optional) – The process group to work on. If
None
, the default process group will be used.timeout (datetime.timedelta, optional) – Timeout for monitored_barrier. If
None
, the default process group timeout will be used.wait_all_ranks (bool, optional) – Whether to collect all failed ranks or not. By default, this is
False
andmonitored_barrier
on rank 0 will throw on the first failed rank it encounters in order to fail fast. By settingwait_all_ranks=True
monitored_barrier
will collect all failed ranks and throw an error containing information about all failed ranks.
- Returns
None
.
- Example::
>>> # Note: Process group initialization omitted on each rank. >>> import torch.distributed as dist >>> if dist.get_rank() != 1: >>> dist.monitored_barrier() # Raises exception indicating that >>> # rank 1 did not call into monitored_barrier. >>> # Example with wait_all_ranks=True >>> if dist.get_rank() == 0: >>> dist.monitored_barrier(wait_all_ranks=True) # Raises exception >>> # indicating that ranks 1, 2, ... world_size - 1 did not call into >>> # monitored_barrier.
- class torch.distributed.Work¶
A Work object represents the handle to a pending asynchronous operation in PyTorch’s distributed package. It is returned by non-blocking collective operations, such as dist.all_reduce(tensor, async_op=True).
- class torch.distributed.ReduceOp¶
An enum-like class for available reduction operations:
SUM
,PRODUCT
,MIN
,MAX
,BAND
,BOR
,BXOR
, andPREMUL_SUM
.BAND
,BOR
, andBXOR
reductions are not available when using theNCCL
backend.AVG
divides values by the world size before summing across ranks.AVG
is only available with theNCCL
backend, and only for NCCL versions 2.10 or later.PREMUL_SUM
multiplies inputs by a given scalar locally before reduction.PREMUL_SUM
is only available with theNCCL
backend, and only available for NCCL versions 2.11 or later. Users are supposed to usetorch.distributed._make_nccl_premul_sum
.Additionally,
MAX
,MIN
andPRODUCT
are not supported for complex tensors.The values of this class can be accessed as attributes, e.g.,
ReduceOp.SUM
. They are used in specifying strategies for reduction collectives, e.g.,reduce()
.This class does not support
__members__
property.
Profiling Collective Communication¶
Note that you can use torch.profiler
(recommended, only available after 1.8.1) or torch.autograd.profiler
to profile collective communication and point-to-point communication APIs mentioned here. All out-of-the-box backends (gloo
,
nccl
, mpi
) are supported and collective communication usage will be rendered as expected in profiling output/traces. Profiling your code is the same as any regular torch operator:
import torch
import torch.distributed as dist
with torch.profiler():
tensor = torch.randn(20, 10)
dist.all_reduce(tensor)
Please refer to the profiler documentation for a full overview of profiler features.
Multi-GPU collective functions¶
Warning
The multi-GPU functions (which stand for multiple GPUs per CPU thread) are deprecated. As of today, PyTorch Distributed’s preferred programming model is one device per thread, as exemplified by the APIs in this document. If you are a backend developer and want to support multiple devices per thread, please contact PyTorch Distributed’s maintainers.
Third-party backends¶
Besides the builtin GLOO/MPI/NCCL backends, PyTorch distributed supports
third-party backends through a run-time register mechanism.
For references on how to develop a third-party backend through C++ Extension,
please refer to Tutorials - Custom C++ and CUDA Extensions and
test/cpp_extensions/cpp_c10d_extension.cpp
. The capability of third-party
backends are decided by their own implementations.
The new backend derives from c10d::ProcessGroup
and registers the backend
name and the instantiating interface through torch.distributed.Backend.register_backend()
when imported.
When manually importing this backend and invoking torch.distributed.init_process_group()
with the corresponding backend name, the torch.distributed
package runs on
the new backend.
Warning
The support of third-party backend is experimental and subject to change.
Launch utility¶
The torch.distributed package also provides a launch utility in torch.distributed.launch. This helper utility can be used to launch multiple processes per node for distributed training.
Module torch.distributed.launch
.
torch.distributed.launch
is a module that spawns up multiple distributed
training processes on each of the training nodes.
Warning
This module is going to be deprecated in favor of torchrun.
The utility can be used for single-node distributed training, in which one or more processes per node will be spawned. The utility can be used for either CPU training or GPU training. If the utility is used for GPU training, each distributed process will be operating on a single GPU. This can achieve well-improved single-node training performance. It can also be used in multi-node distributed training, by spawning up multiple processes on each node for well-improved multi-node distributed training performance as well. This will especially be beneficial for systems with multiple Infiniband interfaces that have direct-GPU support, since all of them can be utilized for aggregated communication bandwidth.
In both cases of single-node distributed training or multi-node distributed
training, this utility will launch the given number of processes per node
(--nproc-per-node
). If used for GPU training, this number needs to be less
or equal to the number of GPUs on the current system (nproc_per_node
),
and each process will be operating on a single GPU from GPU 0 to
GPU (nproc_per_node - 1).
How to use this module:
Single-Node multi-process distributed training
python -m torch.distributed.launch --nproc-per-node=NUM_GPUS_YOU_HAVE
YOUR_TRAINING_SCRIPT.py (--arg1 --arg2 --arg3 and all other
arguments of your training script)
Multi-Node multi-process distributed training: (e.g. two nodes)
Node 1: (IP: 192.168.1.1, and has a free port: 1234)
python -m torch.distributed.launch --nproc-per-node=NUM_GPUS_YOU_HAVE
--nnodes=2 --node-rank=0 --master-addr="192.168.1.1"
--master-port=1234 YOUR_TRAINING_SCRIPT.py (--arg1 --arg2 --arg3
and all other arguments of your training script)
Node 2:
python -m torch.distributed.launch --nproc-per-node=NUM_GPUS_YOU_HAVE
--nnodes=2 --node-rank=1 --master-addr="192.168.1.1"
--master-port=1234 YOUR_TRAINING_SCRIPT.py (--arg1 --arg2 --arg3
and all other arguments of your training script)
To look up what optional arguments this module offers:
python -m torch.distributed.launch --help
Important Notices:
1. This utility and multi-process distributed (single-node or multi-node) GPU training currently only achieves the best performance using the NCCL distributed backend. Thus NCCL backend is the recommended backend to use for GPU training.
2. In your training program, you must parse the command-line argument:
--local-rank=LOCAL_PROCESS_RANK
, which will be provided by this module.
If your training program uses GPUs, you should ensure that your code only
runs on the GPU device of LOCAL_PROCESS_RANK. This can be done by:
Parsing the local_rank argument
>>> import argparse
>>> parser = argparse.ArgumentParser()
>>> parser.add_argument("--local-rank", type=int)
>>> args = parser.parse_args()
Set your device to local rank using either
>>> torch.cuda.set_device(args.local_rank) # before your code runs
or
>>> with torch.cuda.device(args.local_rank):
>>> # your code to run
>>> ...
3. In your training program, you are supposed to call the following function
at the beginning to start the distributed backend. It is strongly recommended
that init_method=env://
. Other init methods (e.g. tcp://
) may work,
but env://
is the one that is officially supported by this module.
>>> torch.distributed.init_process_group(backend='YOUR BACKEND',
>>> init_method='env://')
4. In your training program, you can either use regular distributed functions
or use torch.nn.parallel.DistributedDataParallel()
module. If your
training program uses GPUs for training and you would like to use
torch.nn.parallel.DistributedDataParallel()
module,
here is how to configure it.
>>> model = torch.nn.parallel.DistributedDataParallel(model,
>>> device_ids=[args.local_rank],
>>> output_device=args.local_rank)
Please ensure that device_ids
argument is set to be the only GPU device id
that your code will be operating on. This is generally the local rank of the
process. In other words, the device_ids
needs to be [args.local_rank]
,
and output_device
needs to be args.local_rank
in order to use this
utility
5. Another way to pass local_rank
to the subprocesses via environment variable
LOCAL_RANK
. This behavior is enabled when you launch the script with
--use-env=True
. You must adjust the subprocess example above to replace
args.local_rank
with os.environ['LOCAL_RANK']
; the launcher
will not pass --local-rank
when you specify this flag.
Warning
local_rank
is NOT globally unique: it is only unique per process
on a machine. Thus, don’t use it to decide if you should, e.g.,
write to a networked filesystem. See
https://github.com/pytorch/pytorch/issues/12042 for an example of
how things can go wrong if you don’t do this correctly.
Spawn utility¶
The Multiprocessing package - torch.multiprocessing package also provides a spawn
function in torch.multiprocessing.spawn()
. This helper function
can be used to spawn multiple processes. It works by passing in the
function that you want to run and spawns N processes to run it. This
can be used for multiprocess distributed training as well.
For references on how to use it, please refer to PyTorch example - ImageNet implementation
Note that this function requires Python 3.4 or higher.
Debugging torch.distributed
applications¶
Debugging distributed applications can be challenging due to hard to understand hangs, crashes, or inconsistent behavior across ranks. torch.distributed
provides
a suite of tools to help debug training applications in a self-serve fashion:
Python Breakpoint¶
It is extremely convenient to use python’s debugger in a distributed environment, but because it does not work out of the box many people do not use it at all. PyTorch offers a customized wrapper around pdb that streamlines the process.
torch.distributed.breakpoint makes this process easy. Internally, it customizes pdb’s breakpoint behavior in two ways but otherwise behaves as normal pdb. 1. Attaches the debugger only on one rank (specified by the user). 2. Ensures all other ranks stop, by using a torch.distributed.barrier() that will release once the debugged rank issues a continue 3. Reroutes stdin from the child process such that it connects to your terminal.
To use it, simply issue torch.distributed.breakpoint(rank) on all ranks, using the same value for rank in each case.
Monitored Barrier¶
As of v1.10, torch.distributed.monitored_barrier()
exists as an alternative to torch.distributed.barrier()
which fails with helpful information about which rank may be faulty
when crashing, i.e. not all ranks calling into torch.distributed.monitored_barrier()
within the provided timeout. torch.distributed.monitored_barrier()
implements a host-side
barrier using send
/recv
communication primitives in a process similar to acknowledgements, allowing rank 0 to report which rank(s) failed to acknowledge
the barrier in time. As an example, consider the following function where rank 1 fails to call into torch.distributed.monitored_barrier()
(in practice this could be due
to an application bug or hang in a previous collective):
import os
from datetime import timedelta
import torch
import torch.distributed as dist
import torch.multiprocessing as mp
def worker(rank):
dist.init_process_group("nccl", rank=rank, world_size=2)
# monitored barrier requires gloo process group to perform host-side sync.
group_gloo = dist.new_group(backend="gloo")
if rank not in [1]:
dist.monitored_barrier(group=group_gloo, timeout=timedelta(seconds=2))
if __name__ == "__main__":
os.environ["MASTER_ADDR"] = "localhost"
os.environ["MASTER_PORT"] = "29501"
mp.spawn(worker, nprocs=2, args=())
The following error message is produced on rank 0, allowing the user to determine which rank(s) may be faulty and investigate further:
RuntimeError: Rank 1 failed to pass monitoredBarrier in 2000 ms
Original exception:
[gloo/transport/tcp/pair.cc:598] Connection closed by peer [2401:db00:eef0:1100:3560:0:1c05:25d]:8594
TORCH_DISTRIBUTED_DEBUG
¶
With TORCH_CPP_LOG_LEVEL=INFO
, the environment variable TORCH_DISTRIBUTED_DEBUG
can be used to trigger additional useful logging and collective synchronization checks to ensure all ranks
are synchronized appropriately. TORCH_DISTRIBUTED_DEBUG
can be set to either OFF
(default), INFO
, or DETAIL
depending on the debugging level
required. Please note that the most verbose option, DETAIL
may impact the application performance and thus should only be used when debugging issues.
Setting TORCH_DISTRIBUTED_DEBUG=INFO
will result in additional debug logging when models trained with torch.nn.parallel.DistributedDataParallel()
are initialized, and
TORCH_DISTRIBUTED_DEBUG=DETAIL
will additionally log runtime performance statistics a select number of iterations. These runtime statistics
include data such as forward time, backward time, gradient communication time, etc. As an example, given the following application:
import os
import torch
import torch.distributed as dist
import torch.multiprocessing as mp
class TwoLinLayerNet(torch.nn.Module):
def __init__(self):
super().__init__()
self.a = torch.nn.Linear(10, 10, bias=False)
self.b = torch.nn.Linear(10, 1, bias=False)
def forward(self, x):
a = self.a(x)
b = self.b(x)
return (a, b)
def worker(rank):
dist.init_process_group("nccl", rank=rank, world_size=2)
torch.cuda.set_device(rank)
print("init model")
model = TwoLinLayerNet().cuda()
print("init ddp")
ddp_model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[rank])
inp = torch.randn(10, 10).cuda()
print("train")
for _ in range(20):
output = ddp_model(inp)
loss = output[0] + output[1]
loss.sum().backward()
if __name__ == "__main__":
os.environ["MASTER_ADDR"] = "localhost"
os.environ["MASTER_PORT"] = "29501"
os.environ["TORCH_CPP_LOG_LEVEL"]="INFO"
os.environ[
"TORCH_DISTRIBUTED_DEBUG"
] = "DETAIL" # set to DETAIL for runtime logging.
mp.spawn(worker, nprocs=2, args=())
The following logs are rendered at initialization time:
I0607 16:10:35.739390 515217 logger.cpp:173] [Rank 0]: DDP Initialized with:
broadcast_buffers: 1
bucket_cap_bytes: 26214400
find_unused_parameters: 0
gradient_as_bucket_view: 0
is_multi_device_module: 0
iteration: 0
num_parameter_tensors: 2
output_device: 0
rank: 0
total_parameter_size_bytes: 440
world_size: 2
backend_name: nccl
bucket_sizes: 440
cuda_visible_devices: N/A
device_ids: 0
dtypes: float
master_addr: localhost
master_port: 29501
module_name: TwoLinLayerNet
nccl_async_error_handling: N/A
nccl_blocking_wait: N/A
nccl_debug: WARN
nccl_ib_timeout: N/A
nccl_nthreads: N/A
nccl_socket_ifname: N/A
torch_distributed_debug: INFO
The following logs are rendered during runtime (when TORCH_DISTRIBUTED_DEBUG=DETAIL
is set):
I0607 16:18:58.085681 544067 logger.cpp:344] [Rank 1 / 2] Training TwoLinLayerNet unused_parameter_size=0
Avg forward compute time: 40838608
Avg backward compute time: 5983335
Avg backward comm. time: 4326421
Avg backward comm/comp overlap time: 4207652
I0607 16:18:58.085693 544066 logger.cpp:344] [Rank 0 / 2] Training TwoLinLayerNet unused_parameter_size=0
Avg forward compute time: 42850427
Avg backward compute time: 3885553
Avg backward comm. time: 2357981
Avg backward comm/comp overlap time: 2234674
In addition, TORCH_DISTRIBUTED_DEBUG=INFO
enhances crash logging in torch.nn.parallel.DistributedDataParallel()
due to unused parameters in the model. Currently, find_unused_parameters=True
must be passed into torch.nn.parallel.DistributedDataParallel()
initialization if there are parameters that may be unused in the forward pass, and as of v1.10, all model outputs are required
to be used in loss computation as torch.nn.parallel.DistributedDataParallel()
does not support unused parameters in the backwards pass. These constraints are challenging especially for larger
models, thus when crashing with an error, torch.nn.parallel.DistributedDataParallel()
will log the fully qualified name of all parameters that went unused. For example, in the above application,
if we modify loss
to be instead computed as loss = output[1]
, then TwoLinLayerNet.a
does not receive a gradient in the backwards pass, and
thus results in DDP
failing. On a crash, the user is passed information about parameters which went unused, which may be challenging to manually find for large models:
RuntimeError: Expected to have finished reduction in the prior iteration before starting a new one. This error indicates that your module has parameters that were not used in producing loss. You can enable unused parameter detection by passing
the keyword argument `find_unused_parameters=True` to `torch.nn.parallel.DistributedDataParallel`, and by
making sure all `forward` function outputs participate in calculating loss.
If you already have done the above, then the distributed data parallel module wasn't able to locate the output tensors in the return value of your module's `forward` function. Please include the loss function and the structure of the return va
lue of `forward` of your module when reporting this issue (e.g. list, dict, iterable).
Parameters which did not receive grad for rank 0: a.weight
Parameter indices which did not receive grad for rank 0: 0
Setting TORCH_DISTRIBUTED_DEBUG=DETAIL
will trigger additional consistency and synchronization checks on every collective call issued by the user
either directly or indirectly (such as DDP allreduce
). This is done by creating a wrapper process group that wraps all process groups returned by
torch.distributed.init_process_group()
and torch.distributed.new_group()
APIs. As a result, these APIs will return a wrapper process group that can be used exactly like a regular process
group, but performs consistency checks before dispatching the collective to an underlying process group. Currently, these checks include a torch.distributed.monitored_barrier()
,
which ensures all ranks complete their outstanding collective calls and reports ranks which are stuck. Next, the collective itself is checked for consistency by
ensuring all collective functions match and are called with consistent tensor shapes. If this is not the case, a detailed error report is included when the
application crashes, rather than a hang or uninformative error message. As an example, consider the following function which has mismatched input shapes into
torch.distributed.all_reduce()
:
import torch
import torch.distributed as dist
import torch.multiprocessing as mp
def worker(rank):
dist.init_process_group("nccl", rank=rank, world_size=2)
torch.cuda.set_device(rank)
tensor = torch.randn(10 if rank == 0 else 20).cuda()
dist.all_reduce(tensor)
torch.cuda.synchronize(device=rank)
if __name__ == "__main__":
os.environ["MASTER_ADDR"] = "localhost"
os.environ["MASTER_PORT"] = "29501"
os.environ["TORCH_CPP_LOG_LEVEL"]="INFO"
os.environ["TORCH_DISTRIBUTED_DEBUG"] = "DETAIL"
mp.spawn(worker, nprocs=2, args=())
With the NCCL
backend, such an application would likely result in a hang which can be challenging to root-cause in nontrivial scenarios. If the user enables
TORCH_DISTRIBUTED_DEBUG=DETAIL
and reruns the application, the following error message reveals the root cause:
work = default_pg.allreduce([tensor], opts)
RuntimeError: Error when verifying shape tensors for collective ALLREDUCE on rank 0. This likely indicates that input shapes into the collective are mismatched across ranks. Got shapes: 10
[ torch.LongTensor{1} ]
Note
For fine-grained control of the debug level during runtime the functions torch.distributed.set_debug_level()
, torch.distributed.set_debug_level_from_env()
, and
torch.distributed.get_debug_level()
can also be used.
In addition, TORCH_DISTRIBUTED_DEBUG=DETAIL can be used in conjunction with TORCH_SHOW_CPP_STACKTRACES=1 to log the entire callstack when a collective desynchronization is detected. These
collective desynchronization checks will work for all applications that use c10d
collective calls backed by process groups created with the
torch.distributed.init_process_group()
and torch.distributed.new_group()
APIs.
Logging¶
In addition to explicit debugging support via torch.distributed.monitored_barrier()
and TORCH_DISTRIBUTED_DEBUG
, the underlying C++ library of torch.distributed
also outputs log
messages at various levels. These messages can be helpful to understand the execution state of a distributed training job and to troubleshoot problems such as network connection failures. The
following matrix shows how the log level can be adjusted via the combination of TORCH_CPP_LOG_LEVEL
and TORCH_DISTRIBUTED_DEBUG
environment variables.
|
|
Effective Log Level |
---|---|---|
|
ignored |
Error |
|
ignored |
Warning |
|
ignored |
Info |
|
|
Debug |
|
|
Trace (a.k.a. All) |
Distributed components raise custom Exception types derived from RuntimeError:
torch.distributed.DistError: This is the base type of all distributed exceptions.
torch.distributed.DistBackendError: This exception is thrown when a backend-specific error occurs. For example, if the NCCL backend is used and the user attempts to use a GPU that is not available to the NCCL library.
torch.distributed.DistNetworkError: This exception is thrown when networking libraries encounter errors (ex: Connection reset by peer)
torch.distributed.DistStoreError: This exception is thrown when the Store encounters an error (ex: TCPStore timeout)
- class torch.distributed.DistError¶
Exception raised when an error occurs in the distributed library
- class torch.distributed.DistBackendError¶
Exception raised when a backend error occurs in distributed
- class torch.distributed.DistNetworkError¶
Exception raised when a network error occurs in distributed
- class torch.distributed.DistStoreError¶
Exception raised when an error occurs in the distributed store
If you are running single node training, it may be convenient to interactively breakpoint your script. We offer a way to conveniently breakpoint a single rank: