all_gather

paddle.distributed.communication.stream. all_gather ( tensor_or_tensor_list, tensor, group=None, sync_op=True, use_calc_stream=False ) [source]

Gather tensors across devices to a correctly-sized tensor or a tensor list.

Parameters
  • tensor_or_tensor_list (Union[Tensor, List[Tensor]]) – The output. If it is a tensor, it should be correctly-sized. If it is a list, it should be empty or contain correctly-sized tensors.

  • tensor (Tensor) – The input tensor on each rank. The result will overwrite this tenor after communication. Support float16, float32, float64, int32, int64, int8, uint, bool, complex64 or complex128 as the input data type.

  • group (Group, optional) – Communicate in which group. If none is given, use the global group as default.

  • sync_op (bool, optional) – Indicate whether the communication is sync or not. If none is given, use true as default.

  • use_calc_stream (bool, optional) – Indicate whether the communication is done on calculation stream. If none is given, use false as default. This option is designed for high performance demand, be careful to turn it on except you are clearly know its meaning.

Returns

Return a task object.

Warning

This API only supports the dygraph mode now.

Examples

>>> 
>>> import paddle
>>> import paddle.distributed as dist

>>> dist.init_parallel_env()
>>> local_rank = dist.get_rank()
>>> tensor_list = []
>>> if local_rank == 0:
...     data = paddle.to_tensor([[4, 5, 6], [4, 5, 6]])
>>> else:
...     data = paddle.to_tensor([[1, 2, 3], [1, 2, 3]])
>>> task = dist.stream.all_gather(tensor_list, data, sync_op=False)
>>> task.wait()
>>> print(tensor_list)
[[[4, 5, 6], [4, 5, 6]], [[1, 2, 3], [1, 2, 3]]] (2 GPUs)