2.5 D multi-core graphics processor cluster task mapping method

By employing a two-stage task mapping algorithm and an adaptive controller in a multi-core GPU cluster, the problem of combining inter-core and intra-core resource mapping is solved, resource utilization and energy consumption are optimized, adapting to diverse task scenarios and supporting large-scale integration.

CN121326580APending Publication Date: 2026-01-13BEIJING UNIV OF TECH
0 Cites 0 Cited by

Patent Information

Application Number
CN202511591167.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-10-31
Publication Date
2026-01-13

AI Technical Summary

Technical Problem

Existing technologies fail to effectively combine inter-core mapping and intra-core mapping in multi-core GPU clusters, ignoring the scenario characteristics and resource preferences of the task itself, resulting in high communication overhead and low resource utilization.

Method used

A two-stage task mapping algorithm is proposed. First, the GPU chip queue is divided into four virtual queues using a scene-aware heuristic method, and mapping is performed based on task nature and communication cost. Second, an adaptive proportional-integral-derivative controller is used to dynamically allocate resources within the chip, optimizing resource utilization and energy consumption.

Benefits of technology

It achieves low-complexity task mapping, improves overall system performance and resource utilization, reduces system energy consumption, adapts to diverse task scenarios, and supports large-scale chip integration.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to a 2.5 D-based multi-core graphics processor cluster task mapping method, and aims to flexibly schedule computing resources in a multi-core integrated system, improve task execution efficiency and reduce system energy consumption. The invention provides a two-stage task mapping algorithm, the first stage is a scene-aware inter-core-particle mapping method, the scene-aware inter-core-particle mapping method performs scene-based classification management on tasks, and comprehensively considers the computing resource isomerism and inter-core-particle communication overhead in the mapping process, so that a heuristic mapping strategy with low complexity and weak information dependence is realized. In the second stage, a dynamic task mapping method in a core grain based on a proportional-integral-differential controller is provided, and optimal resource allocation in the core grain is guided to be realized in a mode of combining task initial mapping, periodic sampling, control algorithm decision and task remapping, so that the resource utilization rate is improved, and the service quality is guaranteed. The method has the characteristics of good expandability, low delay and low power consumption, and is suitable for a large-scale core particle integrated system.
Need to check novelty before this filing date? Find Prior Art