Kazakhstan’s first computer station for data analysis launched at NU

Nazarbayev University has established the fastest computing station for the advanced development of artificial intelligence (AI). Two NVIDIA DGX-2 servers included in the workstation specialize in general-purpose computing on GPUs to accelerate research in the field of deep learning.

“The research group of the Institute of Smart Systems and Artificial Intelligence of the Nazarbayev University (ISSAI) was one of the first to introduce DGX systems in Kazakhstan, and gained considerable experience in using these systems. The computational potential of ISSAI will allow teachers at Nazarbayev University and researchers from other Kazakhstan universities to implement their research in the field of AI, ”said Professor Atakan Varol, director of ISSAI.

According to experts, NVIDIA DGX systems are based on the revolutionary architecture of NVIDIA Volta GPUs. With innovative GPU software and simplified management, NVIDIA DGX delivers high performance and speed. Since AI work involves big data processing, speed plays an important role. In ordinary servers, the access speed between the GPU, RAM and hard drive is very small, which is why there is a long processing of data (video, images, text, etc.). The new system uses NVLINK and NVSWITCH technologies that significantly accelerate the relationship between GPUs and, accordingly, accelerate the processes of working with Big Data.

In addition, NVIDIA DGX-2 systems provide scientists with the most powerful tools and programs for developing AI projects on any platform, from the desktop system to the data center and cloud systems. Each researcher gets access to the necessary technical programs and research tools (i.e., programs such as pytorch, tensorflow, etc.), thereby allowing one to find the best solutions for processing Big Data.

AI algorithms are becoming more complex and require an unrivaled level of computing power. NVIDIA DGX-2 combines the power of 16 of the most technologically advanced GPUs in the world to accelerate new types of AI models and simplifies the scaling of AI algorithms, thereby increasing the level of computing power and data processing efficiency.

For example, at Nazarbayev University, NVIDIA DGX-2 has already been successfully used to analyze ISSAI projects’ data. These projects include: Face recognition of conventional and thermal images, Voice Command Dataset, Localization using WiFi signals indoors, Chest x-ray analysis, Automatic brain tumor segmentation on MRI images, Optimal placement of city observation cameras.

Each of these projects uses a huge amount of data such as image, video and audio recordings. Without a high performance computational system like DGX, processing this data and training deep learning models is much more complicated and slows down in speed.

You can get access to computing power, for this you need to contact ISSAI at the following contact information: issai@nu.edu.kz ;.

For more information, you can visit the website of the Institute for Smart Systems and Artificial Intelligence, or the Facebook and LinkedIn pages.