SZShengguang Zhu

Shengguang Zhu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Shengguang Zhu has a total of 16 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • GeoProspect: A domain-specific geological large language model with enhanced continual learning

    by Gang Wu, Haitao Wang, Kunyan Zhang, Shengguang Zhu, Jian Cui, Mengzhe Fan, Hailiang Wang, Hengliang Guo, Gubin Zhang, Dujuan Zhang, et al.


  • Open Access

    Porting and Optimization of the Ipopt for the FT-M6678 Platform

    by Hengliang Guo, Lu Liu, Lin Han, Youcai Luo, Yang Guo, Yubo Han, Shengguang Zhu, Gang Wu

    Ipopt is one of the most frequently used nonlinear planning solvers, which is widely used in the fields of energy system optimization, aerospace and artificial intelligence, etc. Due to the differences in the underlying hardware structure and compilation environment, Ipopt cannot be directly used in the high-performance multicore DSP platform FT-M6678, and in order to solve this problem, this paper analyzes the adaptability of Ipopt and the third-party libraries it relies on to FT-M6678 platform, solves the problems such as compilation failure and Program run result errors during the porting process and then port Ipopt and its dependent third-party libraries to the FT-M6678 platform. On this basis, this paper analyzes the time-consuming part of the program execution process, combine it with the architectural features of FT-M6678 platform, proposes to reduce calls to functions that have no effect on the result of the calculation, uses optimization methods such as loop optimization and software pipeline optimization to efficiently combine with compilation optimization and Cache acceleration supported by the platform, supported by the platform to improve the execution efficiency of Ipopt on the FT-M6678 platform. The test cases provided by the official are selected for testing and verification, and the experimental results show that the operation results after porting are correct. Performance tests on the selected test cases show that the optimization method used increases the overall running speed by 15.930 to 19.753 times, and the performance is better than TMS320C6678 platform.