Tianhe Computational Road Forum: CNMGE High-throughput Material Computing Technology Innovation and Application Forum-Review of wonderful reports by experts (2)

This issue interprets the content of the report entitled "Development and Thinking of Domestic Material Computing Platform" by Professor Yang Mingli of Sichuan University. Professor Yang gave a comprehensive introduction to the development status of foreign material computing platforms, the development of my country‘s material computing platforms during the "13th Five-Year Plan" period, requirements, problems and challenges, and suggestions for the construction of my country‘s material computing platforms.

Professor Yang Mingli

Doctor of Science, Ph.D. Supervisor, Deputy Director of Materials Genetic Engineering Research Center of Sichuan University. Member and Deputy Secretary-General of the Material Gene Components Committee of the Chinese Society for Materials Research; Member of the Special Expert Committee and Project Responsible Expert of the National Key Research and Development Program "Key Technologies and Supporting Platforms for Material Genetic Engineering"; Deputy Director of the Material Genetic Engineering Field Committee of the China Material and Experimental Association Standards Committee Member, Chairman of the Subcommittee on Materials High-throughput Computing Technology. Deputy Director of the Key Laboratory of High Energy Density Physics of the Ministry of Education, Director of the Sichuan Biomaterial Genetic Engineering Research Center.

01
The development status of foreign material computing platforms
Driven by the material genome and related projects in the United States and European countries, many typical material computing platforms have been formed abroad, such as Material Project, Aflow, AiiDA, ASE, etc. These platforms have fixed and expandable teams in R&D. After years of accumulation, they have core technologies for material high-throughput computing and data management; in terms of platform operations, they have application modularity, specialization, user experience visualization, customization, and One-click operation realizes professional mode and sharing mode; in terms of supporting support, relying on government policy support and high-performance computing system, it obtains industry support based on popular application fields. These characteristics and experience provide good experience for the development of domestic material computing platform.

02
The construction progress of my country‘s material computing platform during the 13th Five-Year Plan
During the "Thirteenth Five-Year Plan" period, China independently developed a variety of material high-throughput, concurrent, automatic process calculation methods and software, and achieved 10,000-level (104-level) high-throughput concurrent calculations, which is an efficient and rapid screening and discovery method. New materials and prediction of new effects have laid the foundation, and realized the whole process of automatic and visual operation without human intervention from modeling, calculation to data analysis.

In terms of application verification, a new type of Y-Sb2Te3 information storage material was designed based on high-throughput calculations, and after experimental verification, the performance of the memory prepared with this material has reached the international advanced level. The high-throughput concurrent first-principles calculation system for high-throughput superalloys with multiple complex structures developed within 18 days completed the calculation and screening task of traditional methods that required 440 days. In the multi-scale design of fiber-reinforced epoxy resin-based composites, it has achieved efficient screening capabilities from 1,000-100,000 examples, and the performance prediction error is within 20%. At present, some specialized application platforms have been initially established, such as ALKEMIE of Beihang University, Matcloud of the Network Information Center of Chinese Academy of Sciences, CNMGE of Tianjin Supercomputing, MIP of Shanghai University, Matgen of Guangzhou Supercomputing and so on.

03
Demands, problems and challenges of domestic material computing platforms

Professor Yang pointed out that China‘s materials genetic engineering and material computing platforms have the following problems and challenges: insufficient attention and insufficient investment in infrastructure construction in the field of new materials; basic software and application software are heavily dependent on foreign countries, and independent intellectual property rights are insufficient; target orientation and demand traction are insufficient , Production, education and research are disconnected; management and operation mechanisms have yet to be established, responsibilities and rights need to be clear, professional talents are lacking, and the current system is not conducive to the growth of talents; how to solve the problem of the survival and sustainable development of the material computing platform? And the new international situation poses new challenges to the software and hardware on which the material computing platform depends.

04
Suggestions on the construction of my country‘s material computing platform

Professor Yang put forward some suggestions for the construction of domestic material computing platform.

Government level: build a national-level material computing platform and guide governments at all levels to support the development of the platform.

Project level: Continue to set up material computing platform tasks in major national science and technology projects or key research and development plans, give full play to the demonstration role of platforms built during the "13th Five-Year Plan" period, further strengthen platform functions and applications, and take advantage of China‘s supercomputing advantages to be efficient and orderly Promote platform construction.

Industry level: Serving the national new material strategy, supporting material research and industry, undertaking major national and local key new material research projects, and playing the supporting role of the platform.

Technical level: Take the road of integration of "computing platform + data platform" and equal emphasis on "hardware + software + application", so that the platform will become the carrier of the new ecology of material computing software in my country.

Project undertaker level: it is necessary to continuously improve the technical level, form industry characteristics, expand the user base, and cautiously promote commercial operations.

05
Concluding remarks
Although China is a big manufacturing country, the problem of "stuck neck" in manufacturing in China still exists. Materials used in many manufacturing industries still rely heavily on imports, and some materials are even embargoed or exported.

In order to improve the research and development efficiency of new materials, shorten the research and development cycle, reduce the cost of research and development, solve the problem of "stuck neck" made in China, and promote the innovation of research and development in the field of material genetic engineering and industrialization service mode, the National Supercomputing Tianjin Center has created a high Flux calculation platform. Taking the platform as the starting point, guided by the material innovation needs of multi-industry and multi-link enterprises, integrate and gather domestic and foreign materials research and development teams to form efficient demand docking and coupling, break the constraints of traditional regions, industries and talent teams, and form industry-university-research institutes Use a collaborative innovation environment to efficiently solve the needs of enterprises and improve the efficiency and ability of innovation-driven materials in the field of materials.
This information is from the Internet for academic exchanges. If there is any infringement, please contact us and delete it immediately