local_comm_res

Applicability

Product	Supported
Atlas A3 training products/Atlas A3 inference products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas 200I/500 A2 inference products	x
Atlas inference products	x
Atlas training products	x

Note: For the Atlas A2 training products/Atlas A2 inference products, only the Atlas 800I A2 inference server, Atlas 300I A2 inference card, and A200I A2 Box heterogeneous components are supported.

Function Description

Configures local communication resources.

Prototype

local_comm_res(local_comm_res)

Parameters

Parameter	Data Type	Description
local_comm_res	str	Local communication resource information, formatted as a JSON string. Only the device information used by the current LLM-DataDist instance in ranktable needs to be configured; the server_count and rank_id fields are not required. For details about ranktable, see "Reference" > "Cluster Information Configuration" in HCCL User Guide. Example: { "server_list": [ { "device": [ { "device_id": "0", "device_ip": "x.x.x.x" }, ], "server_id": "xxxx" } ], "status": "completed", "version": "1.0" } If enable_cache_manager and enable_remote_cache_accessible are not set, configuring this option automatically sets both parameters to True. Note: This field can be left empty. The following is an example: local_comm_res = ""

Parameter

Data Type

Description

local_comm_res

str

Local communication resource information, formatted as a JSON string. Only the device information used by the current LLM-DataDist instance in ranktable needs to be configured; the server_count and rank_id fields are not required. For details about ranktable, see "Reference" > "Cluster Information Configuration" in HCCL User Guide. Example:

{
    "server_list": [
        {
            "device": [
                {
                    "device_id": "0",
                    "device_ip": "x.x.x.x"
                },
            ],
            "server_id": "xxxx"
        }
    ],
    "status": "completed",
    "version": "1.0"
}

If enable_cache_manager and enable_remote_cache_accessible are not set, configuring this option automatically sets both parameters to True.

Note:

This field can be left empty. The following is an example:

local_comm_res = ""

Example

from llm_datadist import LLMConfig
llm_config = LLMConfig()
llm_config.local_comm_res = '''{
    "status": "completed",
    "version": "1.0",
    "server_list": [
        {
            "server_id": "node_0",
            "device": [
                {
                    "device_id": "0",
                    "device_ip": "x.x.x.x"
                }
            ]
        }
    ]
}'''

Returns

In normal cases, no value is returned.

If a parameter is incorrect, a TypeError or ValueError may be thrown.

Restrictions

If this option is configured, the following restrictions apply:

When link_clusters or unlink_clusters is called, the append_local_ip_info field in the clusters parameter is not required if local_comm_res is not empty. If local_comm_res is empty, append_local_ip_info must be configured.
Currently, enable_cache_manager and enable_remote_cache_accessible cannot be set to False.

Parent topic: LLMConfig