One of the great merits of the new pretty girl of artificial intelligence, Deepseek, is having reduced their prices, both of development and use, without losing their capacity. One of the keys to this was to have an expert engineering team and the other use the microchips available in the market. But do it creativity.
The graphic processing unit, or GPU, was initially the brain of video games: the “hemisphere” that processed the representation of graphics and video games. But for a decade, GPUs are becoming more popular for use in creative production and artificial intelligence (AI).
The developers also began to take advantage of the power of the GPUs to drastically accelerate the additional workloads of other parts of the microchip and use them in high performance computing, deep learning and more. Now a team of scientists from the Shenzhen MSU-bit University, co-founded by the State University of Moscow Lomonosov and the Beijing Institute of Technology, has developed a new algorithm that improves the computational efficiency of the GPU.
The advance opens new possibilities to Solve complex mechanical problems in various industries, including aerospace and military applicationsin widely available chips that are low cost and are not subject to US sanctions.
The algorithm focuses on the theory of Peridinamics (basically, the analysis of the materials and their possible failures) a science that was so far limited for not being able to perform large -scale simulations, since needed high memory use and processing speeds.
To address these challenges, the team led by Yang Yang, He took advantage of NVIDIA CUDA programming technology to optimize algorithm design and memory managementwhich resulted in a remarkable increase in performance. The results were published in the Chinese Journal of Computational Mechanics.
“This efficient computational power allows researchers reduce calculations that would normally take days to only a few hours, or even minutes, using a common domestic gpuwhich is a significant advance, ”concludes the study.
The interesting thing about this advance is not only the remarkable increase in performance, or that it has been carried out in a commercially available microchip, rather that it is about An advance that unites Russia in a technology such as artificial intelligence.