mindspore.ops.ReduceScatter

class mindspore.ops.ReduceScatter(op=ReduceOp.SUM, group=GlobalComm.WORLD_COMM_GROUP)[source]

Reduces and scatters tensors from the specified communication group. For more details about it, please refer to Distributed Set Communication Primitives - ReduceScatter .

Note

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

Parameters

  • op (str) – Specifies an operation used for element-wise reductions, like SUM and MAX. Default: ReduceOp.SUM.

  • group (str) – The communication group to work on. Default: “GlobalComm.WORLD_COMM_GROUP”.

Inputs:

  • input_x (Tensor) - Input Tensor, suppose it has a shape \((N, *)\), where * means any number of additional dimensions. N must be divisible by rank_size. rank_size refers to the number of cards in the communication group.

Outputs:

Tensor, it has the same dtype as input_x with a shape of \((N/rank\_size, *)\).

Raises

  • TypeError – If any of operation and group is not a string.

  • ValueError – If the first dimension of the input cannot be divided by the rank_size.

Supported Platforms:

Ascend GPU

Examples

Note

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

For the Ascend devices, users need to prepare the rank table, set rank_id and device_id. Please see the Ascend tutorial for more details.

For the GPU devices, users need to prepare the host file and mpi, please see the GPU tutorial .

This example should be run with 2 devices.

>>> import mindspore as ms
>>> from mindspore import Tensor
>>> from mindspore.communication import init
>>> from mindspore.ops import ReduceOp
>>> import mindspore.nn as nn
>>> import mindspore.ops as ops
>>> import numpy as np
>>>
>>> ms.set_context(mode=ms.GRAPH_MODE)
>>> init()
>>> class Net(nn.Cell):
...     def __init__(self):
...         super(Net, self).__init__()
...         self.reducescatter = ops.ReduceScatter(ReduceOp.SUM)
...
...     def construct(self, x):
...         return self.reducescatter(x)
...
>>> input_ = Tensor(np.ones([8, 8]).astype(np.float32))
>>> net = Net()
>>> output = net(input_)
>>> print(output)
[[2. 2. 2. 2. 2. 2. 2. 2.]
 [2. 2. 2. 2. 2. 2. 2. 2.]
 [2. 2. 2. 2. 2. 2. 2. 2.]
 [2. 2. 2. 2. 2. 2. 2. 2.]]