In addition to Weibo, there is also WeChat
Please pay attention
WeChat public account
Shulou
2025-01-16 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > Servers >
Share
Shulou(Shulou.com)06/02 Report--
With the advent of Industry 4.0, major automobile manufacturers are also facing unprecedented challenges. Faced with the pressure of performance growth, cost and challenges from new entrants, automobile manufacturers have used CAE simulation technology to enhance their core competitiveness. CAE simulation technology has become a favorable tool for automobile passive safety research. This technology can not only achieve early feedback in the design process, reduce or avoid product design errors, but also greatly reduce costs and shorten development cycles.
For example, a collision of automobile head needs 1 million data to support, and the completion of vehicle development needs to experience about 2000 collisions. After CAE technology is used, the time is saved thousands of times compared with the real vehicle collision, and the consumption of the original vehicle model is reduced.
In order to further improve efficiency, large-scale automobile enterprises adopt high-performance computing clusters to carry CAE simulation technology. According to an IDC survey, automakers using HPC programs have reduced the average time to develop a new model from 60 months to 24 months. However, the huge investment in the early HPC environment has slowed down the pace of the auto giants.
Changan Automobile's Original HPC System Rebuilding Challenges
Chongqing Changan Automobile Co., Ltd., referred to as Changan Automobile, is the core vehicle enterprise under China Changan Automobile Group Co., Ltd., and Changan officially entered the automobile field in the early 1980s.
With the continuous evolution of new technologies, in order to adapt to the development of business, the original HPC system of Chang 'an Automobile is difficult to meet, and needs to be comprehensively upgraded from the aspects of architecture, management and construction mode. In terms of computing, a unified computing resource pool based on x86 architecture should be built to improve the computing performance of HPC platform, so as to further enhance the market competitiveness of Changan Automobile and bring the business to a higher level. However, the road to HPC system transformation will not be smooth and will encounter many challenges.
Challenge 1: How to build high-performance computing systems? Chongqing Chang 'an Automobile Research Institute is mainly responsible for vehicle design-related business. The simulation work of vehicle structure design, collision simulation, fluid simulation, NVH and so on has a large amount of calculation, which requires high-performance computer to assist design. However, the calculation performance of the existing HPC platform cannot meet the business requirements.
Challenge 2: How to design a structure to improve management efficiency? How to meet the computing resource needs of different departments is an issue that needs to be considered early in the architecture and management planning process. Customers hope that after determining the architecture and management mode, they can simply expand the software and hardware without changing the management mode. Due to insufficient management planning in the early stage, each research institute cannot reasonably utilize computing resources, and customers urgently need to re-plan from management architecture and computing platform to cope with high performance computing scenarios under current and even future business pressures.
Huawei Helps Changan Automobile Remodel High Performance Computing Platform
In response to the challenges posed by the above customers, Huawei carefully designed and planned the high-performance computing cluster system for Changan Automobile, and provided architecture and management mode consulting services. Huawei has helped Chang 'an make detailed consultation planning and preparation from design, management planning, calculation test and other aspects, so that Chang' an Automobile can clearly define the goal and direction when building high-performance computing cluster.
The HPC system of the original customer network is seriously insufficient in computing capacity and cannot meet the business requirements of each research institute, which greatly prolongs the design cycle and test cycle of the automobile. After cooperation with Huawei, we will investigate the requirements of each research institute, calculate the size of computing resources required by the service, and consider the allocation of computing resources by core services and general services to provide a suitable architecture and configuration for the current situation.
Efficient and stable computing resources: This solution uses Huawei converged architecture E9000 to form a computing cluster, and Huawei OceanStor 9000 fully symmetric distributed storage to meet the IOPS requirements of computing clusters. It gradually replaces the existing network GPFS architecture and builds a HPC solution based on Infiniband 100GB EDR network.
Huawei E9000 Converged Architecture Blade Server is a new generation of powerful infrastructure platform launched by Huawei. It integrates computing, storage, switching and management. It has strong competitiveness in hardware computing platform availability, computing density, energy saving and emission reduction, backplane bandwidth, intelligent management and service, elastic configuration and flexible expansion of computing and storage, low network latency and application acceleration.
Huawei OceanStor 9000 is a scale-out file storage system for mass unstructured data storage applications. Fully symmetric distributed architecture to provide users with unstructured data shared resource storage with industry-leading performance, massive scale-out capabilities and a very large single file system. Improve IT system storage efficiency, simplify workloads and migration processes, and cope with the growth and evolution of unstructured data.
Reasonable management methods ensure the computing of business departments: Huawei conducts business research in depth into customer networks, starts with management and architecture, designs the computing requirements of business departments in time division and fragmentation, uses the short board principle, independently designs high-pressure nodes, and performs time division multiplexing. For general high-performance computing services, flexible scheduling through management means enables all business departments to use resources, make good use of resources, and use up resources.
At the same time, the computing node + distributed storage architecture solution provided by Huawei is adopted to cut down the GPFS nodes in the original architecture, and the failure rate is reduced to less than 3%(one year). The high bandwidth characteristics of distributed storage greatly improve the access efficiency of large file examples.
Computational power = market competitiveness
Huawei deploys customers 'typical algorithms such as StarCCM and Dyna in the same business environment for testing, in which the calculation cycle of fluid examples is shortened in a linear manner, and the overall efficiency is improved by 10%-30%; the business efficiency of vehicle design and vehicle testing is improved, shortening the vehicle launch cycle and greatly improving the market competitiveness of Changan Automobile.
Huawei provides a full set of high-performance computing planning and design services for the characteristics of Changan Automobile, which greatly improves its work efficiency; deploys a new high-performance simulation platform, which greatly improves computing performance; replaces the GPFS storage architecture and adopts professional distributed storage to make data storage more convenient and reduce the failure rate by 70%.
Huawei's high-performance simulation platform adopts Huawei's integrated solution, which has good scalability and subsequent expansion is convenient and fast. Through the management planning of business departments, business departments no longer need to seize high-performance computing resources. Each business department can flexibly use the computing platform. Through the scheduling system, for heavy I/O or high-frequency I/O cases, no longer let a single department monopolize, and time-sharing multiplexing is carried out, thus improving the satisfaction of business design departments.
In conclusion, the current digital transformation of industrial manufacturing requires smarter, more efficient, and more convenient enterprise-class IT solutions, especially in the area of high-performance computing. Huawei advocates the core values of openness, cooperation and win-win, fully integrates the resources and capabilities of various industry partners, and provides innovative industrial manufacturing solutions for customers to help them succeed in business.
Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.
Views: 0
*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.
Continue with the installation of the previous hadoop.First, install zookooper1. Decompress zookoope
"Every 5-10 years, there's a rare product, a really special, very unusual product that's the most un
How to build a Windows virtual machine
© 2024 shulou.com SLNews company. All rights reserved.