o Me ã@s4ddlmmZddlmZdd„Zd dd„ZdS) éN)Ú collectivecCsF|durt ¡n|}|r|j ||¡S|j |||¡}|r!| ¡|S)N)rZ_get_default_groupZ process_groupZbroadcast_on_calc_streamÚ broadcastÚwait)ÚtensorÚsrcÚgroupÚsync_opÚuse_calc_streamÚtask©r úaD:\Projects\ConvertPro\env\Lib\site-packages\paddle/distributed/communication/stream/broadcast.pyÚ_broadcast_in_dygraphsr TFcCsH|dur | ¡s tdƒ‚|s|rtdƒ‚t ¡r t|||||ƒStdƒ‚)a× Broadcast a tensor to all devices. Args: tensor (Tensor): The tensor to broadcast. Support float16, float32, float64, int32, int64, int8, uint8 or bool as its data type. src (int, optional): Rank of the source device. If none is given, use `0` as default. 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: .. code-block:: python # required: distributed import paddle import paddle.distributed as dist dist.init_parallel_env() local_rank = dist.get_rank() 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.broadcast(data, src=1, sync_op=False) task.wait() out = data.numpy() # [[1, 2, 3], [1, 2, 3]] (2 GPUs) NzjThe group should not be None and all ranks which invoke this operation should be the member of this group.z5use_calc_stream can only be True in sync op behavior.zJpaddle.distributed.stream.broadcast is only supported in dygraph mode now.)Z is_memberÚ RuntimeErrorÚ frameworkZin_dygraph_moder )rrrrr r r r rs%ÿÿ ÿÿr)rNTF)Zpaddle.fluid.frameworkZfluidrZpaddle.distributedrr rr r r r Ús