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

An Optimization Method for Task-Oriented Parallel Programming Model in Virtualization Environment

A virtualization environment and programming model technology, applied in the optimization field of task-parallel programming model under virtualization environment, can solve the problems of lack of virtualization environment optimization, waste of computing resources, low performance, etc., achieve low overhead, reduce scheduling delay Time, the effect of reducing waste

Active Publication Date: 2017-09-12
HUAZHONG UNIV OF SCI & TECH
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

If a thief thread in the task parallel programming model continues to steal tasks and fails, the virtual CPU executing the thread will waste computing resources
Existing task-parallel programming models, such as Cilk++, TBB, and BWS, lack optimization for virtualized environments in terms of reducing the waste of computing resources by thief threads.
Therefore, the task-parallel programming model faces the problem of low performance in a virtualized environment

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
  • An Optimization Method for Task-Oriented Parallel Programming Model in Virtualization Environment
  • An Optimization Method for Task-Oriented Parallel Programming Model in Virtualization Environment
  • An Optimization Method for Task-Oriented Parallel Programming Model in Virtualization Environment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0059] Table 1 Example 1 Experiment configuration environment

[0060]

[0061] As listed in Table 1, Embodiment 1 deploys 8 customer virtual machines with 16 cores on a single 16-core physical server; respectively start and run 1, 2, 4, and 8 customer virtual machines to simulate a single physical core Scenes shared by 1, 2, 4, and 8 virtual CPUs; Embodiment 1 runs the Conjugate Gradient (CG) application based on the task parallel programming model in client virtual machines 1 to 8, and tests CG in Cilk++, The running time under the support of BWS and the present invention; wherein, CG is derived from a set of application program sets representing fluid dynamics calculations developed by NASA; Cilk++ is the most general task parallel programming model; BWS is realized based on Cilk++ , which is the best task parallel programming model in the current traditional stand-alone multi-core multi-application environment; in embodiment 1, acceleration optimization is performed for...

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 an optimization method for a task parallel programming model in a virtualized environment; the stealing operation that fails in the task parallel programming model is obtained in the front-end monitoring part of the client virtual machine; According to the running state of the acceleration initiator and the accelerated candidate and the information of the physical CPU where they are located, a decision is made whether to perform acceleration. If acceleration is performed, the remaining time slice of the acceleration initiator is provided to the accelerated candidate. When the accelerated candidate is preempted by the virtual machine monitor (the time slice is exhausted or blocked), if the accelerated candidate is in the runnable state , restore its original scheduling path; the present invention adds optimization for the virtualization environment to the existing task parallel programming model, reduces the waste of computing resources for virtual CPUs running thief threads, and reduces the scheduling of virtual CPUs running useful threads Delay, so as to maximize the input of physical computing resources into effective computing.

Description

technical field [0001] The invention belongs to the technical field of virtualization and parallel program optimization, and more specifically relates to a task-oriented parallel programming model optimization method under a virtualization environment. Background technique [0002] With the increase in the number of processor cores, the computer programming model has changed from the traditional serial programming model to the new parallel programming model to achieve the same practical effect as the number of cores increases; in recent years, the task parallel programming model has been widely used in development Parallel applications designed to simplify parallel programming and improve multi-core utilization. [0003] The core technology of the task-parallel programming model is task-stealing scheduling, that is, each processor core corresponds to a thread, and each thread maintains a double-ended queue. The tail of the queue is used to push ready tasks or pop out tasks t...

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 Patents(China)
IPC IPC(8): G06F9/44G06F9/455
Inventor 吴松金海彭亚琼
Owner HUAZHONG UNIV OF SCI & TECH
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