aiaccel.torch.lightning.ABCIEnvironment#

class aiaccel.torch.lightning.ABCIEnvironment[source]#

Environment class for ABCI.

This class provides methods to interact with the ABCI environment, such as retrieving the world size, global rank, node rank, and local rank.

__init__() → None[source]#

Methods

__init__()

detect()

Detects the environment settings corresponding to this cluster and returns True if they match.

global_rank()

The rank (index) of the currently running process across all nodes and devices.

local_rank()

The rank (index) of the currently running process inside of the current node.

node_rank()

The rank (index) of the node on which the current process runs.

set_global_rank(rank)

set_world_size(size)

teardown()

Clean up any state set after execution finishes.

validate_settings(num_devices, num_nodes)

Validates settings configured in the script against the environment, and raises an exception if there is an inconsistency.

world_size()

The number of processes across all devices and nodes.

Attributes

creates_processes_externally

Whether the environment creates the subprocesses or not.

main_address

The main address through which all processes connect and communicate.

main_port

An open and configured port in the main node through which all processes communicate.