mindspore.ops.AllGather

class mindspore.ops.AllGather(group=GlobalComm.WORLD_COMM_GROUP)[source]

Gathers tensors from the specified communication group and returns the tensor which is all gathered.

Note

  • The tensors must have the same shape and format in all processes of the collection.

Parameters

group (str, optional) – The communication group to work on. Default: GlobalComm.WORLD_COMM_GROUP , which means "hccl_world_group" in Ascend, and "nccl_world_group" in GPU.

Inputs:

  • input_x (Tensor) - The shape of tensor is \((x_1, x_2, ..., x_R)\).

Outputs:

Tensor. If the number of devices in the group is N, then the shape of output is \((N, x_1, x_2, ..., x_R)\).

Raises

  • TypeError – If group is not a str.

  • ValueError – If the local rank id of the calling process in the group is larger than the group's rank size.

  • RuntimeError – If device target is invalid, or backend is invalid, or distributed initialization fails.

Supported Platforms:

Ascend GPU

Examples

Note

Before running the following examples, you need to configure the communication environment variables.

For Ascend/GPU/CPU devices, it is recommended to use the msrun startup method without any third-party or configuration file dependencies. Please see the msrun start up for more details.

This example should be run with 2 devices.

>>> import numpy as np
>>> import mindspore as ms
>>> from mindspore import ops
>>> import mindspore.nn as nn
>>> from mindspore.communication import init
>>> from mindspore import Tensor
>>>
>>> ms.set_context(mode=ms.GRAPH_MODE)
>>> init()
>>> class Net(nn.Cell):
...     def __init__(self):
...         super(Net, self).__init__()
...         self.allgather = ops.AllGather()
...
...     def construct(self, x):
...         return self.allgather(x)
...
>>> input_x = Tensor(np.ones([2, 8]).astype(np.float32))
>>> net = Net()
>>> output = net(input_x)
>>> print(output)
[[1. 1. 1. 1. 1. 1. 1. 1.]
 [1. 1. 1. 1. 1. 1. 1. 1.]
 [1. 1. 1. 1. 1. 1. 1. 1.]
 [1. 1. 1. 1. 1. 1. 1. 1.]]
Tutorial Examples: