Class UPipelinedStepper

Class UPipelinedStepper

• Defined in File PipelinedStepper.h

Nested Relationships

Nested Types

• Struct UPipelinedStepper::FPipelinedStepperFrame

Inheritance Relationships

Base Types

• public UObject

• public IStepper (Class IStepper)

class UPipelinedStepper : public UObject, public IStepper

Source: Source/Schola/InferenceUtils/Public/Steppers/PipelinedStepper.h

Dependencies: IAgent, IPolicy

Pipelined stepper implementation for asynchronous inference.

A sophisticated stepper that pipelines agent observations and policy inference across multiple frames to hide inference latency. Uses a two-stage pipeline:

Frame N: Collect observations -> Dispatch async inference Frame N+1: Apply actions from N -> Collect observations -> Dispatch async inference

This allows the policy to perform inference on a background thread while the game continues running, significantly reducing effective latency for expensive inference operations. The stepper manages synchronization between the game thread and inference threads.

---

Public Functions

Symbol	Details
Init	Initialize the pipelined stepper with agents and policy.
Step	Execute one step of the pipelined agent-policy loop.
BeginDestroy	Override for object destruction.

Init

inline virtual bool Init(const TArray<TScriptInterface<IAgent>> &InAgents,
                         const TScriptInterface<IPolicy> &InPolicy) override

Initialize the pipelined stepper with agents and policy.

Sets up the pipeline frames and prepares for asynchronous operation.

Parameters

• InAgents – [in] The array of agents to manage

• InPolicy – [in] The policy to use for inference (must support batched operations)

Returns:

true if initialization succeeded (agents and policy are valid), false otherwise

#	Direction	Name	Type	Description
1	—	InAgents	const TArray< TScriptInterface< IAgent > > &	The array of agents to manage
2	—	InPolicy	const TScriptInterface< IPolicy > &	The policy to use for inference (must support batched operations)

Attributes: inline, virtual

Source: Source/Schola/InferenceUtils/Public/Steppers/PipelinedStepper.h

Step

void Step()

Execute one step of the pipelined agent-policy loop.

Must be called every tick on the Game Thread. Performs:

• Applies actions from the previous frame (if ready)

• Collects observations from all agents

• Dispatches asynchronous inference if policy is not busy

The inference runs on a background thread and results are applied in a subsequent frame once ready.

Source: Source/Schola/InferenceUtils/Public/Steppers/PipelinedStepper.h

Implementation: Source/Schola/InferenceUtils/Private/Steppers/PipelinedStepper.cpp

BeginDestroy

inline virtual void BeginDestroy() override

Override for object destruction.

Sets the shutdown flag to prevent background threads from accessing the object after it has been destroyed.

Attributes: inline, virtual

Source: Source/Schola/InferenceUtils/Public/Steppers/PipelinedStepper.h