Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

GPU thread load balancing method and device

A load balancing and threading technology, applied in multi-program devices, program control design, processor architecture/configuration, etc., can solve problems such as wasted computing resources, excessive, and lack of memory access optimization for load and storage units

Active Publication Date: 2020-04-28
INSPUR SUZHOU INTELLIGENT TECH CO LTD
View PDF6 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the direct difference in architecture between the GPU and the CPU (central processing unit), the GPU needs to continuously improve the architecture design and optimize the scheduling mechanism and strategy in terms of hardware, so as to fully ensure the full utilization of computing resources and avoid excessive hardware. overhead
The GPU structure of the prior art has an unbalanced load on thread computing tasks, causing some computing units to be busy or idle for a long time, affecting computing efficiency, or causing a lot of waste of computing resources due to deadlock problems; there is a lack of corresponding access to load and store units. Memory optimization, easy to cause unbalanced hardware loss
[0003] There is currently no effective solution to the problem of uneven load on GPU hardware resources in the prior art

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
  • GPU thread load balancing method and device
  • GPU thread load balancing method and device
  • GPU thread load balancing method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0056] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0057] It should be noted that all the expressions "first" and "second" in the embodiments of the present invention are used to distinguish two entities with the same name but not the same or parameters that are not the same, as shown in "first" and "second" Only for the convenience of presentation, it should not be construed as a limitation to the embodiments of the present invention, and subsequent embodiments will not describe this one by one.

[0058] Based on the foregoing objective, the first aspect of the embodiments of the present invention proposes an embodiment of a method capable of balancing the load of each thread of the GPU. figure 1 Shown is a schematic flowchart of the GPU thread load ba...

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 GPU thread load balancing method and device. The GPU thread load balancing method comprises the following steps: polling and accessing a computing unit and a loading storageunit of each thread to obtain register state information of each thread and form a register state table; analyzing the type of an instruction being executed according to the register state informationof each thread, determining the type of an execution instruction of each thread, and updating a historical storage load table; and allocating a new work task to each thread according to the registerstate table and the historical storage load table. According to the invention, each thread of the GPU is enabled to balance the load, the working efficiency and stability are improved, and the servicelife of hardware is prolonged.

Description

Technical field [0001] The present invention relates to the field of load balancing, and more specifically, to a method and device for GPU thread load balancing. Background technique [0002] Large-scale parallel processors such as GPUs (graphics processing units) stack a large number of computing units in space, and improve computing performance by increasing the degree of parallelism. However, due to the direct difference in architecture between GPU and CPU (central processing unit), GPU hardware needs to continuously improve its architecture design and optimize scheduling mechanisms and strategies to fully ensure the full use of computing resources and avoid excessive hardware Overhead. The GPU structure of the prior art has an unbalanced load of thread computing tasks, causing some computing units to be busy or idle for a long time, affecting computing efficiency, or due to deadlock problems, resulting in a lot of waste of computing resources; lack of corresponding access to...

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): G06F9/50G06T1/20
CPCG06F9/505G06F9/5083G06T1/20
Inventor 王凯周玉龙
Owner INSPUR SUZHOU INTELLIGENT TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products