Heterogeneous computing hardware energy consumption and performance optimization method and system and storage medium

A hardware and heterogeneous technology, applied in computing, computing models, energy-saving computing, etc., can solve problems affecting performance efficiency and performance degradation, and achieve optimal power consumption and performance, and reduce power consumption

Inactive Publication Date: 2021-11-19
睿识科技南京有限责任公司
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Simply reducing power consumption often impacts performance efficiency, and due to performance degradation, it takes more time to complete tasks

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Heterogeneous computing hardware energy consumption and performance optimization method and system and storage medium
  • Heterogeneous computing hardware energy consumption and performance optimization method and system and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0015] This embodiment provides a method for optimizing energy consumption and performance of heterogeneous computing hardware through reinforcement learning. Such as figure 1 As shown, the hardware frame structure applied by this method mainly includes:

[0016] PS: Processing System (Processing System), part of the SOC based on the ARM core.

[0017] PL: Programmable logic (Progarmmable Logic), FPGA part.

[0018] MCU: Microcontroller Unit Microcontroller Unit.

[0019] PS and PL together form a heterogeneous computing platform, that is, this heterogeneous platform includes ARM and FPGA computing hardware.

[0020] figure 1 The left part: is the target heterogeneous computing platform, that is, the hardware that needs to be optimized for power consumption. In the embodiment of the present invention, this is a hardware including ARM (PS side) and FPGA (PL side).

[0021] figure 1 Intermediate components: power management module, which is a hardware module connected to ...

Embodiment 2

[0073] This embodiment provides an embedded system, including a memory, a processor, and a program stored in the memory and operable on the processor. When the program is executed by the processor, the program described in Embodiment 1 is implemented. Steps of a method for realizing optimization of energy consumption and performance of heterogeneous computing hardware through reinforcement learning.

Embodiment 3

[0075] This embodiment provides a storage medium, the storage medium stores at least one program, and the at least one program can be executed by at least one processor, so as to realize the energy consumption of heterogeneous computing hardware through reinforcement learning described in Embodiment 1. And the steps of the method of performance optimization.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a heterogeneous computing hardware energy consumption and performance optimization method and system and a storage medium, and the method comprises the steps: carrying out the matching of a current operation state of a heterogeneous computing hardware platform with 32 preset operation states, and analyzing the frequency of an embedded processing system and the opening and closing time of a programmable logic computing module based on a reinforcement learning algorithm; and obtaining the optimal operation state of the embedded processing system and the programmable logic, and feeding theoptimal operation state back to the power management tool to realize hardware control. By adopting the method provided by the invention, the resource utilization rate of the heterogeneous platform can be dynamically adjusted, so that the power consumption is reduced to the greatest extent. The reinforcement learning algorithm is used for analyzing and optimizing the resource utilization rate of the state control function of the field programmable gate array, and as a hardware control algorithm, the PL power consumption can be accurately adjusted, it is guaranteed that the power consumption requirement can be met under any work load, and the optimal power consumption and performance are obtained.

Description

technical field [0001] The invention relates to an embedded system, in particular to a method for optimizing energy consumption and performance of heterogeneous computing hardware. Background technique [0002] In embedded systems, the conventional strategy for optimizing power consumption is to reduce power consumption by reducing the operating speed of the processor. Energy-efficient computing platforms require greater power and performance efficiency, not just energy savings. Simply reducing power consumption often impacts performance efficiency, and tasks take more time to complete due to performance degradation. Traditional dynamic voltage frequency scaling (DVFS) and adaptive voltage scaling (AVS) technologies focus on the collaborative architecture of the processing system (PS) of the ARM hardware and the programmable logic (PL) of the field programmable gate array (FPGA), usually relying on in feedback control. Currently, many heterogeneous computing hardware can ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06F1/3234G06F1/324G06F1/3287G06F11/30G06F11/34G06N20/00
CPCG06F1/3243G06F1/324G06F1/3287G06F11/3062G06F11/3447G06F11/3457G06N20/00Y02D10/00
Inventor 俞喆祺杜越刘念
Owner 睿识科技南京有限责任公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap