Architecture design, key technologies, and application research of GPU heterogeneous resource pool platform based on virtualization

The current challenges facing the field of artificial intelligence include high prices and market supply disruptions. The traditional single-card, single-use model results in low resource utilization and efficiency. Furthermore, existing technological research methods make it difficult to support th...

Full description

Saved in:
Bibliographic Details
Main Authors: ZHANG Wancai, ZHANG Nan, YANG Wenqing, WANG Tao, ZHANG Wenqiang
Format: Article
Language:zho
Published: Beijing Xintong Media Co., Ltd 2024-09-01
Series:Dianxin kexue
Subjects:
Online Access:http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2024216/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1841528875831525376
author ZHANG Wancai
ZHANG Nan
YANG Wenqing
WANG Tao
ZHANG Wenqiang
author_facet ZHANG Wancai
ZHANG Nan
YANG Wenqing
WANG Tao
ZHANG Wenqiang
author_sort ZHANG Wancai
collection DOAJ
description The current challenges facing the field of artificial intelligence include high prices and market supply disruptions. The traditional single-card, single-use model results in low resource utilization and efficiency. Furthermore, existing technological research methods make it difficult to support the efficient management and scheduling of diverse heterogeneous GPU resources. Based on this, a virtualization-based GPU heterogeneous resource pool platform was proposed. Firstly, the overall architecture, logical architecture, and functional architecture of the platform were planned and designed. Secondly, key technologies were studied, and a virtualization heterogeneous GPU resource pool framework and a scheduling model based on time slicing + load balancing were proposed. Finally, based on the methods described, various innovative application models were proposed, including multiservice single-card stacking, cross-pull, cross-machine integration, hybrid deployment, and time division multiplexing. The research method proposed provides enterprise-level AI applications with GPU computing resources that are compatible with multiple GPU manufacturers, support remote access, flexible partitioning and aggregation, and flexible scheduling. Following the completion of calculations and an in-depth analysis, it has been demonstrated that a reduction of up to 60% in the number of GPU cards can be achieved while simultaneously enhancing operational efficiency by a factor of four.
format Article
id doaj-art-a0c824c580274684ab7c853f2fdc4d89
institution Kabale University
issn 1000-0801
language zho
publishDate 2024-09-01
publisher Beijing Xintong Media Co., Ltd
record_format Article
series Dianxin kexue
spelling doaj-art-a0c824c580274684ab7c853f2fdc4d892025-01-15T03:34:03ZzhoBeijing Xintong Media Co., LtdDianxin kexue1000-08012024-09-014016217573366408Architecture design, key technologies, and application research of GPU heterogeneous resource pool platform based on virtualizationZHANG WancaiZHANG NanYANG WenqingWANG TaoZHANG WenqiangThe current challenges facing the field of artificial intelligence include high prices and market supply disruptions. The traditional single-card, single-use model results in low resource utilization and efficiency. Furthermore, existing technological research methods make it difficult to support the efficient management and scheduling of diverse heterogeneous GPU resources. Based on this, a virtualization-based GPU heterogeneous resource pool platform was proposed. Firstly, the overall architecture, logical architecture, and functional architecture of the platform were planned and designed. Secondly, key technologies were studied, and a virtualization heterogeneous GPU resource pool framework and a scheduling model based on time slicing + load balancing were proposed. Finally, based on the methods described, various innovative application models were proposed, including multiservice single-card stacking, cross-pull, cross-machine integration, hybrid deployment, and time division multiplexing. The research method proposed provides enterprise-level AI applications with GPU computing resources that are compatible with multiple GPU manufacturers, support remote access, flexible partitioning and aggregation, and flexible scheduling. Following the completion of calculations and an in-depth analysis, it has been demonstrated that a reduction of up to 60% in the number of GPU cards can be achieved while simultaneously enhancing operational efficiency by a factor of four.http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2024216/GPU heterogeneous resource poolcomputing power platformvirtualizationtime slicingload balancing
spellingShingle ZHANG Wancai
ZHANG Nan
YANG Wenqing
WANG Tao
ZHANG Wenqiang
Architecture design, key technologies, and application research of GPU heterogeneous resource pool platform based on virtualization
Dianxin kexue
GPU heterogeneous resource pool
computing power platform
virtualization
time slicing
load balancing
title Architecture design, key technologies, and application research of GPU heterogeneous resource pool platform based on virtualization
title_full Architecture design, key technologies, and application research of GPU heterogeneous resource pool platform based on virtualization
title_fullStr Architecture design, key technologies, and application research of GPU heterogeneous resource pool platform based on virtualization
title_full_unstemmed Architecture design, key technologies, and application research of GPU heterogeneous resource pool platform based on virtualization
title_short Architecture design, key technologies, and application research of GPU heterogeneous resource pool platform based on virtualization
title_sort architecture design key technologies and application research of gpu heterogeneous resource pool platform based on virtualization
topic GPU heterogeneous resource pool
computing power platform
virtualization
time slicing
load balancing
url http://www.telecomsci.com/zh/article/doi/10.11959/j.issn.1000-0801.2024216/
work_keys_str_mv AT zhangwancai architecturedesignkeytechnologiesandapplicationresearchofgpuheterogeneousresourcepoolplatformbasedonvirtualization
AT zhangnan architecturedesignkeytechnologiesandapplicationresearchofgpuheterogeneousresourcepoolplatformbasedonvirtualization
AT yangwenqing architecturedesignkeytechnologiesandapplicationresearchofgpuheterogeneousresourcepoolplatformbasedonvirtualization
AT wangtao architecturedesignkeytechnologiesandapplicationresearchofgpuheterogeneousresourcepoolplatformbasedonvirtualization
AT zhangwenqiang architecturedesignkeytechnologiesandapplicationresearchofgpuheterogeneousresourcepoolplatformbasedonvirtualization