High -performance algorithm that can solve the problem of consumer complex material design GPU Developed by Chinese researchers, the speed was 800 times higher than conventional methods.
Developed by A the study The Team of the Shenzhen MSU-BIT university, co-founded by Lomonosov Moscow State University and Beijing Institute of Technology, enhances the most advanced local theory with a new algorithm (PD). Crack, damage, fracture.
Open new possibilities to solve complex mechanical problems in various industries, including aerospace and aerospace Army Applications, low cost, widely available chips that are not eligible US sanctions。
Peliginemics has been proven to be advantageous in modeling of materials damage, but due to the highly calculated complexity, traditionally large -scale simulation is inefficient and the memory usage and processing speed are used. There were problems such as late.
In order to deal with these issues, Yan Yang, an associate professor, leverage nvidia’s CUDA programming technology Create a PD General Framework. By performing detailed analysis of the unique structure of the chip, her team has optimized the design and memory management of algorithms that have led to surprising performance improvements. Their research was published in CHINESE JOURNAL OF COMPUTATIONAL MECHANICS on January 8th.
“With this efficient calculation ability, researchers can usually reduce the calculation that takes a few hours, and even a few minutes. This is an important progress in PD research. Is using “
