Multitask Oriented GPU Resource Sharing and Virtualization in Cloud Environment

Zhao, Xingfang; Zhang, Yujie; Su, Bing

doi:10.1007/978-3-319-27122-4_35

Cited by 2 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This framework allows the registration and the execution of GPU kernels from multiple applications concurrently on distributed GPUs. Zhoa et al 12 built a dynamic GPU resources allocation and management framework based on multiple load features of the GPU in the cloud using the virtual machines. This framework allows an automatic detection of the nodes in the cloud and estimate the computing capacity of GPU dynamically to reach the purpose of the rational allocation of the GPU in the cloud environments.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

KubCG: A dynamic Kubernetes scheduler for heterogeneous clusters

2020

View full text Add to dashboard Cite

SummaryContainer platforms are increasingly being used to deploy cloud‐based services. Nevertheless, many cloud services are also demanding graphics processing units (GPUs) to accelerate different applications that make use of their parallel architecture, such as deep learning or just video processing. Thus, different container technologies, such as Docker and Kubernetes, are implementing GPU support. Some effort is being devoted to design algorithms to schedule applications into heterogeneous computing systems that use CPUs and GPUs together. This article is part of this effort, and we describe how to build a dynamic scheduling platform for Kubernetes that is able to manage the deployment of Docker containers in a heterogeneous cluster, which we call KubCG. This platform implements a new scheduler that optimizes the deployment of new containers by taking into account the Kubernetes Pod timeline and the historical information about the execution of the containers. We have performed different tests to validate this new algorithm, and KubCG was able to reduce the time to complete different tasks down to a 64% of the original time in our different experiments.

show abstract

Section: Related Workmentioning

confidence: 99%

“…T Run ′ 1 in Equation 9and T Run ′ 2 in Equation 11represent, respectively, the total duration of the running time of the first and the second Pod. In addition, Equations (10) and (12) represent, respectively, the increase in the running time of Pods 1 and Pod 2 due to concurrency.…”

Section: Multiple Pods In the Same Nodementioning

confidence: 99%

KubCG: A dynamic Kubernetes scheduler for heterogeneous clusters

2020

View full text Add to dashboard Cite

show abstract

“…How to efficiently organize and share underlying resources on-demand for mission-oriented swarm intelligent systems is another vital problem [1,40]. As well known, technologies that are concerned with resource sharing, virtualization, and service are very useful for swarm systems, through which the cooperation will be enhanced efficiently from the resource level [41,42]. Essentially, distributed resource sharing is the foundation for distributed computing and cloud computing, and has been studied widely.…”

Section: Resource Sharing and Virtualizationmentioning

confidence: 99%

Multi-Constraint Optimized Planning of Tasks on Virtualized-Service Pool for Mission-Oriented Swarm Intelligent Systems

et al. 2019

View full text Add to dashboard Cite

With the emergence of swarm intelligent systems, especially the swarming of aircraft and ground vehicles, cooperation in multiple dimensions has becoming one of the great challenges. How to dynamically schedule the resources within a swarm intelligent system and optimize the execution of tasks are all vital aspects for such systems. Focusing on this topic, in this paper, one new task planning mechanism with multiple constraints is proposed to solve such dynamic programming problems. Concretely, several fundamental models, covering three-level task models and resource-service pool models, are put forward and defined first. Considering the limitations of swarm systems running within complicated cyber-physical space, multi-dimension constraints for tasks scheduling and execution are further modeled and established. On this basis, we mapped this planning problem to an optimization searching problem, and then proposed a Genetic-Algorithm-based mechanism. All these works have been verified with simulated cooperation scenes. Experimental results show that this new mechanism is efficient to solve such resource-related and mission-oriented cooperation problems in complicated environments.

show abstract

Multitask Oriented GPU Resource Sharing and Virtualization in Cloud Environment

Cited by 2 publications

References 14 publications

KubCG: A dynamic Kubernetes scheduler for heterogeneous clusters

KubCG: A dynamic Kubernetes scheduler for heterogeneous clusters

Multi-Constraint Optimized Planning of Tasks on Virtualized-Service Pool for Mission-Oriented Swarm Intelligent Systems

Contact Info

Product

Resources

About