In order to break the western countries' blockade and restrictions on China in the field of chip technology, the country has initiated major special projects of 'nuclear high-tech' (core electronic devices, high-end general-purpose chips and basic software products), concentrating on conquering core key technologies in related fields, and building The independent innovation system frees China from the embarrassment of being stuck in the neck and becomes a powerful competitor in the world's information industry. After years of unremitting efforts, China has successively introduced a series of self-developed domestic chips such as the Godson series. , has achieved world-renowned achievements.
In the top 500 global supercomputers list in 2016, supercomputers using domestically produced chips, the “Sweet Lake Lights” supercomputer, topped the list, indicating that our country’s information industry is undergoing a historic transformation, and the level of innovation is mainly tracking. A new historical stage of tracking, running, and co-existing. In March 2018, the Chinese Academy of Sciences announced that a Mapu algebraic microprocessor, a new concept microprocessor with completely independent intellectual property rights and leading international standards, will be officially released during the year. , Turned on the "China Core" corner overtaking.
The so-called MaPU algebraic micro-processor, in short, the Chinese Academy of Sciences' demand for computationally intensive applications, the world's first concept of the 'Mathematic Processing Unit' for high performance algebraic microprocessors, parallel algebraic operations, parallel memory system instruction systems. Major innovations in hardware architecture and hardware support from scalar/superscalar operations to algebraic operations, including supercomputing high-performance algebraic microprocessors, general-purpose communications algebraic microprocessors, general-purpose multimedia algebraic micro Processors and so on.
The MaPU algebraic micro-processor combines the flexibility of the CPU, the parallel storage characteristics of the FPGA, and the high efficiency of the ASIC. Through the change of the architecture, the system hardware level is improved, and the optimal space-time decomposition strategy is combined with the super-variable algebra pipeline. The method can directly map the optimized algorithm and fill the application computing gap. It is a revolutionary innovation in the field of microprocessors. At present, it has completed the design of the architecture and the design of the processor chip front and back end, and applied for more than 40 invention patents. , Including 8 international patents.