HCCL_DEBUG_CONFIG

Description

Specifies which HCCL submodules output detailed runtime information to the runtime logs (that is, logs under the $HOME/ascend/log/run directory). Valid values include ALG (or alg) (algorithm orchestration module), TASK (or task) (task orchestration module), and RESOURCE (or resource) (resource management module, covering resource allocation and release operations).

This environment variable can be configured in either of the following ways:

Positive configuration: One or more modules can be configured. Use commas (,) to separate modules. ALG (or alg), TASK (or task), and RESOURCE (or resource) are case-insensitive.

# Record the runtime information about the task module in runtime logs.
export HCCL_DEBUG_CONFIG="TASK" 
# Record the runtime information about the alg, task, and resource modules in runtime logs.
export HCCL_DEBUG_CONFIG="alg,task,resource"

Negative configuration: Add a "^" prefix to the first module name, indicating that detailed runtime information of all submodules except the specified ones will be recorded in runtime logs.

# Record the runtime information about all modules except the task module in runtime logs. (In the current version, only the runtime information about the alg and resource modules is recorded.)
export HCCL_DEBUG_CONFIG="^task"
# Record the runtime information about all modules except the task and alg modules in runtime logs. (In the current version, only the runtime information about the resource module is recorded.)
export HCCL_DEBUG_CONFIG="^task,alg"

Note: When configuring environment variables, do not add redundant spaces. Otherwise, the configuration is invalid. For example, if there are redundant spaces before and after task in export HCCL_DEBUG_CONFIG="alg, task ", the environment variable configuration is invalid.

Example

export HCCL_DEBUG_CONFIG="ALG,TASK,RESOURCE"

Constraints

None

Applicability

Atlas A3 training product/Atlas A3 inference product

Atlas A2 training product/Atlas A2 inference 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.)

Parent topic: Debugging