Background

Driven by the rapid evolution of artificialintelligence (AI) technology, the education industry is ushered inunprecedented opportunities for innovation. Through the collection ofbehavioral data, management data and digital content in all teaching links byusing artificial intelligence technology in the education industry, and theanalysis of teaching content and teaching behavior, educational institutionswill be able to provide more advanced educational tools and methods for theeducators. Plus, shape more efficient, fair and individualizeddevelopment education models, and help the educators to improve theireducational opportunities and then transform them into qualified social roles.

Challenge: How to Improve Reasoning Efficiency and Save Costs inthe Application of Smart Education

Cloud classroom scheme is a solutionspecially built for computer classroom in primary and secondary schools. It issuitable for teaching scenes such as information technology class,human-computer interaction test, AI/ programming / multimedia applicationcourse in primary and secondary schools. It has the characteristics ofmulti-load fusion, easy management, easy maintenance, easy application and soon. In the practical application scene, the backend supporter of smartclassroom will pre-enter the list of students who should be in the classroom.The on-site camera completes the attendance by intelligently “scanning theface” of the students in the classroom. For supporting the AI application of"face scanning" attendance, expression, action analysis and so on,this needs enough backend operation force to support. Based on the pre-trainingdeep learning model, the additional graphical processor (GPU) is initiallyconfigured in the server, but GPU as a dedicated hardware will significantlyincrease the cost of hardware procurement, which is not conducive to thewidespread landing of cloud classroom schemes. Besides, the deployment of GPUwill also bring related problems such as late maintenance, and increase theburden of school operation and maintenance.

Solution profile: low-cost,efficient AI reasoning based on Intel ® architecture

On the basis of the second-generation IntelXeon Scalable processors and open VINO tool kits (no GPU), Trusme’s 2 U high performanceserver SW G220-P12I create "cloud classroom AI upgrade" program, whileensuring ultra-high energy ratios and easy deployment with significantlyimproved AI reasoning. Moreover, it supports teaching institutions to conductnear real-time and efficient analysis of teaching quality. Intelligently guideteachers to optimize the difficult points of the courseware, similarly guidestudents to strengthen learning and practice on weak points, which improve thequality of classroom teaching greatly.

Cloud classroomplan AI upgrade deployment chart

The second-generation Intel Xeon Scalable processornot only has powerful general computing power, but also integrates innovativetechnologies such as enhanced single-instruction majority data stream (SingleInstruction Multiple Data, SIMD), Intel AVX-512, etc. to achieve a balancebetween general computing power and parallel computing power, providing anexcellent performance basis for artificial intelligence training. This solutionnot only can widely improve the operational efficiency of AI applications, butalso can support the smart cloud classroom for a wide range of high-performanceapplications, with excellent flexibility and performance. It can help schoolsbetter integrate hardware resources and reduce the overall cost of ownership(TCO).

Trusme’s high performance server SW G220-P12I isoriented to high performance computing (HPC), storage and general applications.It is designed with the mainstream frame; Standard 2 U height makes SWG220-P12I easier to deploy. In hardware configuration, SW G220-P 12I comes withtwo of latest Intel XEON Scalable processors and a full load of up to 3TB2933MHz DDR4 memory; it can be extended up to 6 PCIe cards. Can install storagedrive and heat exchange fan module in simple 2 U chassis without tool, highdensity and scalability, suitable for a variety of virtualization solutions, bigdata analysis, and other application scenarios that require higher flexibility,higher performance, and greater storage capabilities.

Trusme 2U High Performanceserver SW G220-P12I

Customer benefits

Thanks to the significant improvement in theAVX512_VNNI instruction set of the second generation Intel Xeon Scalableprocessor in INT8 FT32 hybrid precision reasoning capabilities, and theoptimization of the Open VINO tool suite, coupled with the high reliability andeasy maintenance features of the high integration and modular design of the Trusme’s SW G220-P12I server, the AI reasoningcapability of the cloud classroom scheme is greatly improved. While reducingthe TCO, it satisfies the harsh requirements of cloud classroom AI applicationssuch as face recognition, facial expression analysis and action detection forcomputational power, and accelerates deep learning reasoning. Promote schoolsto make full use of classroom data, quickly gain insight, and then optimize theteaching model, targeted teaching according to their aptitude.

Company profile

Guangzhou Trusme Electronic Technology Co.,Ltd.

Guangzhou Trusme Electronic Technology Co.,Ltd., abbreviated as Trusme shares (stock symbol :836447) was established in2005, it is mainly engaged in the integration of cloud computing and edgecomputing ecosystems. It is a comprehensive high-tech enterprise integratingserver/storage production customization, supply chain services, and after-salesoperation and maintenance, and leader of the domestic server and ITinfrastructure product industry customization.

Add: No. 11th Kai Yuan Avenue, UniteA4, 4^th Floor, Guangzhou, P.R. China 510535

Tel: 86-13148986323(Angelia Zhou)   0086-20-37267442