megengine.functional.distributed.scatter#

scatter(inp, group=WORLD, device=None, axis=0)[source]#

Split tensor in root process at first dimension.

Parameters:

  • inp (Tensor) – Input tensor.

  • group (Optional[Group]) – The process group to work on. The default group is WORLD which means all processes available. You can use a list of process ranks to create new group to work on it, e.g. [1, 3, 5].

  • device (Optional[str]) – The specific device to execute this operator. None default device means the device of inp will be used. Specify “gpu0:1” to execute this operator on diffrent cuda stream, 1 is stream id, and default stream id is 0.

  • axis – The concat axis for collective_comm result The default axis is 0

Return type:

Tensor

Returns:

Split tensor.

Examples

input = Tensor([0 1]) + rank*2
# Rank 0 # input: Tensor([0 1])
# Rank 1 # input: Tensor([2 3])
output = scatter(input)
# Rank 0 # output: Tensor([0])
# Rank 1 # output: Tensor([1])

input = Tensor([0 1]) + rank*2
group = Group([1, 0]) # first rank is root
output = scatter(input, group)
# Rank 0 # output: Tensor([3])
# Rank 1 # output: Tensor([2])