Task Scheduler Device for Low-Latency Power Management

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In systems requiring real-time capability, such as base stations, the transition of CPU cores to deep idle states leads to extended latency times during task restoration, compromising low-latency operations while attempting power saving.

Innovation Solution

A task scheduler device that continuously allocates tasks to a core or core group of the processor based on the processor's use rate, preventing transitions to deep idle states and reducing restoration latency, while allowing other CPU cores to transition to deep idle states for power saving.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Use of energy by stationary object

If CPU cores transition to deep idle states for power saving, then power consumption is reduced, but restoration latency increases

Engineering Contradiction:
Improvepower consumptionVSAvoidrestoration latency
Core Design Contradiction:
Use of energy by stationary objectVSLoss of time

Solution Approach 1:

The patent segments CPU cores into two groups: active core groups that continuously execute tasks and remain in shallow idle states, and inactive cores that transition to deep idle states for power saving. This segmentation resolves the contradiction by allowing different idle state behaviors for different core groups based on their task allocation status.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The task scheduler continuously allocates tasks to active core groups to prevent them from transitioning to deep idle states. By maintaining continuous task allocation as a preliminary action, the system ensures that active cores remain in a ready state with minimal restoration latency while inactive cores can deeply sleep.

Inventive Principle:
Principle #10Preliminary action

2Loss of time

If CPU cores remain in active state to reduce restoration latency, then real-time capability is improved, but power consumption increases

Engineering Contradiction:
Improverestoration latencyVSAvoidpower consumption
Core Design Contradiction:
Loss of timeVSUse of energy by stationary object

Solution Approach 1:

The patent divides CPU cores into active and inactive groups, allowing active cores to remain in low-latency states while inactive cores enter deep power-saving states. This segmentation enables the system to achieve both low restoration latency for active cores and reduced power consumption for inactive cores simultaneously.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different idle state qualities are applied to different CPU core groups based on their local task allocation status. Active core groups maintain shallow idle states with fast restoration characteristics, while inactive cores use deep idle states with maximum power saving characteristics. This local quality differentiation resolves the global contradiction.

Inventive Principle:
Principle #3Local quality

3Loss of time

If tasks are continuously allocated to CPU cores, then restoration latency is reduced, but task scheduling complexity increases

Engineering Contradiction:
Improverestoration latencyVSAvoidtask scheduling complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The task scheduler manages complexity by segmenting CPU cores into active and inactive groups with different scheduling policies. Active core groups receive continuous task allocation to maintain low latency, while inactive cores are left unallocated to enable deep idle transitions. This segmented management approach keeps scheduling complexity manageable.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system dynamically adjusts the classification of CPU cores between active and inactive groups based on task arrival patterns and load conditions. This dynamic adaptation allows the scheduler to optimize the balance between restoration latency and scheduling complexity in real-time without requiring complex static configurations.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4557052A1Task scheduler device, calculation system, and task scheduling method, and program
Publication Date: 2025.05.21 NT T INC
  • EP4557052A1 patent drawingFigure 1
  • EP4557052A1 patent drawingFigure 2
  • EP4557052A1 patent drawingFigure 3

AI summary

A task scheduler device (100, 100A) that, in a calculation system (1000) that reduces power consumption by reducing stepwise an operation state of a processor according to a processing load, allocates tasks to a core group of the processor, task scheduler device (100, 100A) including: a task amount/period prediction part (122) that acquires a use rate of the processor, and a task CPU allocation part (123) that, based on the use rate of the processor acquired by the task amount/period prediction part (122), continuously allocates tasks to a core or a core group of the processor to be used at a predetermined frequency or more.