mindspore.ops.communication.P2POp

class mindspore.ops.communication.P2POp(op, tensor, peer, group=None, tag=0)[source]

Object for batch_isend_irecv input, to store information of "isend" and "irecv".

Note

tensor will be modified in-place by final result when op is "irecv".

Parameters

op (Union[str, function]) – Only string of "isend" and "irecv" are allowed. Or function of ops.communication.isend and ops.communication.irecv are allowed.
tensor (Tensor) – tensor for sending/receiving.
peer (int) – remote global rank for send/receive.
group (str, optional) – The communication group to work on. Default: None, which means "hccl_world_group" in Ascend.
tag (int, optional) – currently not supported yet. Default: 0.

Returns

P2POp Object.

Raises

TypeError – when op is not string or function of 'isend' and 'irecv'.
TypeError – when tensor is not type of Tensor or peer is not int.
NotImplementedError – when tag is not 0.

Supported Platforms:: Ascend

Examples

Note

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

For Ascend devices, it is recommended to use the msrun startup method without any third-party or configuration file dependencies.

Please see the msrun startup for more details.

>>> import numpy as np
>>> import mindspore
>>> from mindspore.ops.communication import P2POp, isend, irecv
>>> from mindspore import Tensor
>>> # Launch 2 processes.
>>> send_tensor = Tensor(1.)
>>> send_op = P2POp('isend', send_tensor, 1)
>>> send_op = P2POp(isend, send_tensor, 1)
>>> recv_tensor = Tensor(0.)
>>> recv_op = P2POp('irecv', recv_tensor, 0)
>>> recv_op = P2POp(irecv, recv_tensor, 0)