Applicability

Product	Supported
Atlas 350 Accelerator Card	√
Atlas A3 training product/Atlas A3 inference product	√
Atlas A2 training product/Atlas A2 inference product	√
Atlas 200I/500 A2 inference product	☓
Atlas inference product	√
Atlas training product	√

For Atlas A2 training product/Atlas A2 inference product, only the Atlas 800T A2 training server, Atlas 900 A2 PoD cluster basic unit, and Atlas 200T A2 Box16 heterogeneous subrack are supported.

For the Atlas inference products, only the Atlas 300I Duo inference card is supported.

Function

Obtains the number of local ranks on the server where the devices in the group are located.

Prototype

def get_local_rank_size(group="hccl_world_group")

Parameters

Parameter	Input/Output	Description
group	Input	A string containing a maximum of 128 bytes, including the end character. If this parameter is not set, the default value hccl_world_group is used.

Parameter

Input/Output

Description

group

Input

A string containing a maximum of 128 bytes, including the end character.

If this parameter is not set, the default value hccl_world_group is used.

Returns

An int for the number of local ranks on the server where the device is located

Constraints

This API must be called after the initialization of collective communication is complete.
The caller rank must be within the range defined by the group argument passed to this API call. Otherwise, the API call fails.
After the create_group call, call this API to obtain the number of local ranks in the current group.
If hccl_world_group is passed, the number of local ranks in world_group is returned.

Call Example

from hccl.manage.api import create_group
from hccl.manage.api import get_local_rank_size
create_group("myGroup", 4, [0, 1, 2, 3])  
localRankSize = get_local_rank_size("myGroup")