In 1973, the american scholar daniel bell introduced the concept of the “post-industrial age”, in which science, technology and innovation became increasingly important factors in the sustainable social and economic development of society. Science and technology in the future will focus more on the design and development of materials, and products will be more than just macro-design, but more than micro-materials. Technology for the development of industrial materials in the future is not based on newton mechanics, but on quantum mechanics, and a new research paradigm is urgently needed to develop a new approach。
The link between the post-industrial age and material calculations is mainly in two areas: material design and material manufacture. First, material design. The post-industrial age advocates the development of new materials using quantum mechanics, artificial intelligence, machine learning and data-driven methods. Quantum mechanics material computing is an important tool in that it uses computer simulations and predictions of the nature and behaviour of materials and provides data for artificial intelligence. By establishing formulae based on quantum mechanics, material simulations can guide the design and optimization of new materials by calculating and predicting parameters such as their structure, performance and stability. This method of calculation can significantly reduce the research and development cycle of materials, reduce trial error costs and provide greater possibilities for finding new materials. Second, material manufacturing. The post-industrial era emphasized efficient, flexible and sustainable manufacturing methods. Traditional material manufacturing processes usually require a large number of experiments and test errors, and material calculations can improve manufacturing efficiency by optimizing design and predictive performance。
Dr. Wang linwan, founder and chief scientist of yongtai technology ltd. In beijing, has been a leading member of the country's large-scale top-level computing industry, where he has worked for more than 30 years in material computing. Pwmat, the world's first fully functional, fully functional, with nearly 1 million lines of code, with the advantage of large-scale computing capabilities, has become the core product of a large system far faster than other similar software and is currently used by more than 200 subject groups and thousands of users in the country. The company has now become a leader in the development of software for computer simulation tools for domestic materials。
Long-tune leading materials computing software industry
First of all, there's a concept of cope, what's the first principle of calculation? Computation of the so-called primary principle is a method of calculation based on quantum mechanics, which allows predictions of the various properties of the material, such as energy-capable structure, charge density, mechanics, optical properties, electrical properties, etc. In response, dr. Wang linsheng stated that “in principle, as long as the earth's materials are available, they can be calculated without any empirical parameters, a very powerful tool. The substance is made of atoms, which follow the laws of quantum mechanics.”
In today's era, as material development gradually moves from process optimization to the design of molecular atomic structures at source, the simulation of microscale materials dominated by the first principle has become increasingly important, and the trend in first-level calculations over the past 20 years has been reflected in indices published in relevant academic papers. At the same time, trends can be observed in industry, such as the important predictive value of capture and analysis of the effects of process changes on the performance of the device, which is a major area of the eda card necks abroad, since when the process is below 7 nm, it enters a microscale where it is difficult to accurately describe the microscale of the device based on the traditional continuous medium drift equation, the single-atom effect, the material impact of quantum effects, and the performance of the device itself, and therefore the computational method based on quantum mechanics with the smallest atom unit has been increasingly applied to the actual semiconductor industrial research and development chain。
Computation material as a cross-cutting discipline in material science and computational science has a very high threshold and is the most hard-core direction within material disciplines, where practitioners often need solid physical background (quantitative mechanics and solid physics) and rigid programming capabilities, and where the knowledge, methods and methods of multiple disciplines need to be integrated and applied, often building competitive advantages over decades. In this regard, the research and development team of beijing dragon trend technology ltd, founded by dr. Wang lin, has received numerous recognitions, including experience-building and world-class awards. The technological advantage has broken through the quantum mechanics bottlenecks of atoms larger than millions of atoms and has gained the strength to advance from academia to industry。
In 2015, dr. Wang lin-kwan founded yong-tai tsing corporation as his own chief scientist. As the founder of the company, he led the company's research and development team in the development of the material computing simulation tool software, applying atomic-level calculations based on quantum mechanics to a range of research and development production chains, such as new material design processing and optimization, and moving the calculation of primary principle materials to industry. At present, pwmat, a self-developed home-grown home-based material computational software with full-owned intellectual property rights (a flat wave power method and based on gpu acceleration), allows for electro-structure calculations and ground-to-ground molecular dynamics simulations, applicable to crystals, defective systems, semiconductor systems, metal systems, nanosystems, quantum points, clusters and molecular systems. Internationally leading technological advantages in the search for algorithms, isomer acceleration and large-scale parallel optimization, with a significant increase in computing scale and computing efficiency while ensuring precision, allow for the application of atomic-level calculations based on quantum mechanics to a range of r & d production chains such as new material design, processing, optimization, etc., and provide revolutionary drivers of innovation for the development of microelectronics, chemicals, alloys, new energy sources, aerospace, automobiles, biomedicines, etc。
Today, under the leadership of dr. Wang lin, yongtai, combined with his own academic and scientific strength, has developed a multi-dimensional market-based campaign involving training in the subject area, hands-on training in software teaching, co-construction of internships in higher schools, provision of case-entry materials, etc., all-round tailor-made programmes to assist in the training of specialists in the various disciplines; a platform for integrated solutions has been set up in conjunction with the national institute of priority laboratories and authorities; an industry summit has been held with the national top design unit, the national brand group, and industry-led enterprises to develop the capacity of the industrial market with the strength of science and technology from the academic market service to provide guidance on the techniques of material computing hard technology and a strong post-sales team to resolve the technical problems of clients on a 24-hour basis。
The vision of yongtome corporation is to develop second-generation industrial software at the atomic level, using the first principle to calculate optimized material design, and has become a leader in the development of innovations in domestic materials computing simulation software。
Dragon-temperature to create a large-scale micro-integrator computing platform
As mentioned earlier, future science and technology is more focused on material design and development, and products are not just macro-design, but also the micro- nature of materials, and future technology for the development of industrial materials is not based on newton mechanics, but on quantum mechanics. As a result, the q-cad concept was first introduced in the country, and the development of the q-cad at the atomic level was a channel for opening a new paradigm of research. Dr. Wang linsheng described q-cad as a deep integrated platform. Existing quantum mechanics can already calculate the nature of the local domain, but not of the whole, while q-cad integrated quantum calculations, artificial intelligence, high performance calculations, monte carlo, etc., employ the concept of partitioning and treating problems into many parallel tasks, parallel calculations, quantum mechanics to calculate each small block and then integrated into larger systems, long-range cologne interactions in large systems, feedback from different regions in different ways, which can be built together in a short period of time to improve the efficiency of material computing.” in today's age of billions of numbers per second, if an industrial problem cannot be calculated, it is probably not because computer capacity is inadequate, but because we do not have the best software to address it. The development of software lags far behind the growth of hardware, as dr. Wang linsheng said。
At present, the three technologies on which the q-cad is based are unique algorithms: pwmat, high performance gpu acceleration, and the learning power field of artificial intelligence machines, overcoming the difficulties of precision, space and timescale, and providing the technical conditions needed to extend the first principle to industry。
First, the unique algorithm, pwmat, is a material computing software based on the density pancreatic theory, which calculates at a much faster rate than similar software. Pwmat allows for fast hse, tddft, namd, solvent model, constant power, acoustic coupling, acoustic spectroscopy of the defective system, non-radiation decay, charge capture, electrical defects of 2d materials, virtual approximation of vca, etc. It is widely used in different areas of material research and is one of the mainstream software for material computing。
Second, high performance calculation gpu acceleration, with pwmat being the world's first to use the gpu-accelerated dft computing software, achieving over 30 times faster performance than the cpu version. Based on this unique performance advantage, pwmat is well suited to large-system calculations, and the computational simulation of the 10,000-atomic scale system can be achieved on the gpu cluster with the first principle of precision. The 1000 atomic system can be easily calculated even on a single-point gpu server。
Finally, the pwmlff machine learning field, which uses the results of calculations of existing first-level principles, is designed to combine the field of power with large system energy and power without subsequent first-level principles. For systems close to the equilibrium, the machine learning field is expected to accelerate molecular dynamics calculations significantly, increasing the size and length of simulators within the limited computing resources. This approach could be used to address industrial problems in different areas, such as semiconductor film growth, electrolyte in lithium batteries and interface problems, small molecular drug screening and metallic alloys. Pwmlff uses a deep neural network (deep nn) model superseding the kalmann filter optimizer and a dp-torch model, which significantly reduces training time. Pwmlff is an open-source platform for gnu license with flexible and scalable structures. In addition, with pwmat software's unique altruistic altimeter, machine learning power fields can use the energy of individual atoms as training data to improve training accuracy and reliability. In addition, the machine learning field can train the power field based on charge density to describe long-range static and non-key interactions。
A success story in the tremor industry
At the present time, the solution of the telecommunications scrambling has been applied to many areas such as new energy sources, semiconductors, and alloy materials, where the high-throughput advantage has been found to be fairly stable in three-dollar orthogonal materials, revealing more possibilities for charging, thus completing the calculation of the electro-voltage platform; similarly, the high-intensity manganese phosphate design of the current heat optimally allows for the optimal matching, which also allows for the selection of suitable mixing elements, the provision of ion-conductor rates and the exploration of intermediate metal solution mechanisms; and the unique fixed electrical pattern (fpm) of the telecommunications screforming can be seen in the study of the mechanism of the formation of the sei film, where it is difficult to see some reactions before it is possible to find a new pattern in battery studies. Similar studies have created a platform for the development and design of the battsim batteries, open to fixed clients, a database of embedded million-grade molecules, selection of elements such as solvents, additives and lithium salt, which can begin to simulate and quickly produce the results of the analysis, which undoubtedly lowers the learning threshold and the trial costs of technicians at the enterprise level, and creates the flexible form of the workflow in which bottom-level data links are built autonomously, untying the difficulties that professionals need to learn cross-cutting knowledge。
In addition to this, it embeds pwmat into the platform for the development of agricultural equipment materials, and works with the institute of agricultural machinery to create a platform for the development of industrial chain materials, which improves the efficiency of material development and development. Key farm equipment materials are being optimized in terms of high-strength, grinding, corrosive, autonomous gear-conformity, etc., and require effective breakthroughs in key common technologies in production applications to upgrade the overall technological level of our agricultural machine manufacturing industry. Long-tune pwmat is able to produce material performance data (mechanical, formative, compositional), cross-comparison of performance data and material process production information from material enterprises, and accelerate design and validation of new materials. Pwmat embedding the r & d platform not only allows for a summary of common performance requirements and common technologies for a wide range of agro-mechanical materials, improving the efficiency of r & d for agro-mechanical materials, but also explores the possibility of integrating r & d for material information platforms and working with other enterprises in the industrial chain to develop a full-process solution for material development。
At the same time, the pwmat was embedded in the research platform for the steel industry to calculate the high temperature and high pressure phase and to obtain a picture of metal metal and titanium that is closest to the results of the experiment. Because of the high and difficult phase costs of experimental measurements of high-temperature high-pressure environmental metals, the industry considers complete and systematic understanding through the thermodynamic properties of the computing material, with the initial phase being mapped and the key points being tested. The rationale for calculating solvency is the quantum mechanic equation (calculation of the first principle) that solves different structures, the gibbs free energy of the material, and the structure of the map. In the computational practice, the elasticization of the material, taking into account its own energy, zero-point energy, material volume changes, atomic vibration entropy, electron vortex direction entropy, electro-orbit detached entropy, etc., constructs the best integrated formula for calculating the phase in different environments, such as pressure, temperature, magnetic canvass. In the china steel research institute project, the only calculations obtained by using the above calculation process were those of ti and fe, which were close to the experimental mapping trend, without using any experimental parameters. The first principle calculations of the density general theory and the error of the experiment are modified by using the very few experimental phase variable point data, the calibration of general letters lda-pbe exchange, the dft-u calibration, etc., and the calculation of the first principle calculations of the density general theory, which are closest to the experimental data in all literature。
Concluding remarks
Long-term enterprises have developed second-generation industrial software at the atomic level, with the vision of advancing the first principle of calculation of applications in industrial enterprise material development and optimizing material design using first principle calculations。
In the hope of the future, the burgeoning news can pragmatically provide industry enterprises with more intelligent, efficient and flexible solutions to meet changing market needs and technological challenges. Moving second-generation industrial software to industry through continuous innovation and technology integration will play an important role in material science, energy, electronics and many other industrial areas to accelerate innovation and promote sustainable development. (van/lijay)
