Samsung Patents a Power-Aware Task Scheduler for Multi-Core Chips

By Patentlyze Team · Updated Jul 24, 2026

Every chip in your phone has multiple processing cores, and the question of which core handles which task has a surprisingly large impact on battery life. Samsung's latest patent describes a scheduling system that tries to answer that question more efficiently than current approaches.

How Samsung's chip scheduler picks the right core for the job

Imagine your phone's chip as a team of workers, each with different strengths. Right now, when your phone gets a new task (opening an app, fetching email), it needs to figure out which worker to hand it to. Do it wrong and you waste power; a worker who's bad at that job burns more energy than one who's well-suited for it.

Samsung's patent describes a method where the system scores each core twice before assigning any task: once for how well that core would handle the incoming task, and again for how well it's currently handling whatever it's already doing. A core only becomes a candidate if it would do a better job on the new task than it's doing on the current one.

From those candidates, the core with the highest score for the new task wins the assignment. The goal is to keep work on the cores that can run it most efficiently, which in practice means squeezing more battery life out of the same hardware.

How affinity scores route tasks to the best available core

The patent introduces a two-score affinity system for deciding where to run a task on a multi-core processor.

When a new task becomes ready to run, the scheduler collects three data points:

• First scheduling parameter: properties of the incoming (ready) task, such as its workload characteristics
• Second scheduling parameter: properties of whatever task is currently running on each core
• First running parameter: the current state of each core, including how long it has been sitting idle

Using those inputs, the system calculates two scores per core. The ready-task affinity (how efficiently a given core would run the new task, expressed as a power-consumption optimization score) and the current-task affinity (how well that core is already matched to its current workload). A core is only eligible to receive the new task if its ready-task affinity score is higher than its current-task affinity score, meaning displacing the current task would actually be a net efficiency gain.

Among all eligible cores, the one with the highest ready-task affinity wins and becomes the target processor core for the new task. The continuous idle duration of each core feeds into the running parameter, so cores that have been sitting unused are weighted into the calculation.

What this means for battery life in Samsung devices

Modern mobile chips, including Samsung's own Exynos line, already use heterogeneous core designs where big, fast cores and small, efficient cores coexist. The scheduling logic that decides which core handles which job is a major factor in how long your battery lasts during everyday use. A smarter scheduler that avoids mismatching tasks to cores could extend screen-on time without any hardware changes.

For Samsung, filing this as a method patent rather than a hardware patent means it could apply broadly across any multi-core chip architecture. That matters both for phones and for chips in televisions, tablets, and other Samsung products where thermal management and battery efficiency are design priorities.

Editorial commentary on a publicly published patent application. Not legal advice.