Dynamic Task Scheduler for CPU Load Balancing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In real-time media processing systems, processing tasks often lead to potential overload of the processing unit during peak times, resulting in idling during non-peak times, and inefficient use of computing power.

Innovation Solution

A method of scheduling additional new processing tasks that determines the timeslots with the most available resources, associates performance consumption with each task, and schedules tasks to minimize the highest load in each timeslot, ensuring even distribution of packets across computing time-slots.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If processing tasks are executed in fixed time intervals regardless of load conditions, then the processing unit maintains regular operation rhythm, but the processing unit becomes overloaded during peak times and idle during non-peak times

Engineering Contradiction:
Improveprocessing throughputVSAvoidtask execution reliability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent applies dynamics by transitioning from fixed time interval scheduling to dynamic scheduling that adapts to current system load conditions. The scheduler monitors available processing capacity and adjusts task assignment accordingly, allowing the system to handle variable workloads effectively while preventing both overload and idle states.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the scheduling parameter from fixed time intervals to load-based decision making. By monitoring processing capacity parameters and using them to determine task assignment, the system optimizes both productivity and reliability under varying operational conditions.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If processing tasks are accumulated in short timeframes to maximize throughput, then more tasks can be processed, but packet delay increases and processing quality deteriorates

Engineering Contradiction:
Improvetask processing volumeVSAvoidpacket delay
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by proactively identifying and assigning tasks to optimal timeslots before peak load conditions occur. The scheduler analyzes future task requirements and pre-assigns them to appropriate time periods, preventing accumulation and delay rather than reacting to them after they occur.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces a new dimension to scheduling by considering not just immediate task execution but also future task requirements and optimal timeslot allocation. This multi-dimensional approach allows the system to distribute tasks across time in a way that maximizes throughput while minimizing delay.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Productivity

If the processing unit operates at high performance continuously, then all tasks can be processed efficiently, but the total computing power exceeds the actual performance needed causing resource waste

Engineering Contradiction:
Improveprocessing speedVSAvoidcomputing resource consumption
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The patent applies partial action by assigning only the necessary computing resources to each timeslot based on actual task requirements. Rather than maintaining high performance continuously, the system allocates processing capacity proportional to the workload, avoiding excessive resource consumption during low-demand periods while ensuring sufficient capacity when needed.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3805924B1Method of scheduling an additional new processing task to be executed by a processor, scheduler, and central processing unit
Publication Date: 2025.04.16 UNIFY BETEILIGUNGSVERWALTUNG GMBH & CO KG
  • EP3805924B1 patent drawingFigure 1
  • EP3805924B1 patent drawingFigure 2
  • EP3805924B1 patent drawing

AI summary

The present invention relates to a method of scheduling an additional new processing task to be executed by a processor which is already executing a predetermined number of processing tasks, wherein each running processing task is to be executed within a timeslot and in a periodic sequence of timeslots by the processor, and wherein each processing task causes a predetermined performance consumption when being executed by the processor, wherein the method comprises the steps of determining the periodicity of execution of each processing task, associating the performance consumption with each processing task, and scheduling the additional new processing task to be executed in a periodic sequence of timeslots so as to satisfy a predetermined performance consumption criterion. Further, the present invention relates to a scheduler and a Central Processing Unit with a scheduler.