Faculty of Civil and Environmental Engineering made the computing cluster available to GTS2 startup teams | Faculty of Civil and Environmental Engineering at the Gdańsk University of Technology

Page content

News

Date added: 2023-11-29

Faculty of Civil and Environmental Engineering made the computing cluster available to GTS2 startup teams

Computing cluster
The cluster available at the Faculty of Civil and Environmental Engineering   is equipped with 16-core Intel Xeon Gold 6346 processors, 1 TB of RAM per node and NVIDIA Tesla A100 80GB graphics cards and is designed to, m.in, support calculations performed using Artificial Intelligence. The NVIDIA A100 is designed to power the world's most efficient data centers, accelerate advanced analytics, and accelerate the power of Artificial Intelligence.

Clusters equipped with modern NVIDIA Tesla A100 graphics cards allow for efficient training of AI models in significantly less time than on standard CPU-based computing servers.

As part of a long-term cooperation, the Faculty of Civil and Environmental Engineering has made one of the cluster nodes available for the exclusive needs of start-up teams – participants of GTS2 programs.

Gdańsk Tech Startup School (GTS2) serves as an entrepreneurship centre at the Gdańsk University of Technology. It focuses on initiating and supporting the development of startup projects based on the results of research conducted, m.in, at the Faculty of Civil and Environmental Engineering.

As part of the Research University's program called Molybdenum Startup School, GTS2 implements two programs: Startup School One – "Test your idea", which is a pre-acceleration program that is currently in the course of the 4th Startup School. Edition, and Startup School Two – "Spread your wings", which is the second stage of the training cycle and has a grant character.

The first to use the cluster's capabilities is the 41hubs team, a graduate of the second edition of the Startup School One program. The team is developing an application that aims to generate expert industry reports on a given topic and answer questions from PDF documents.

A key element of the application is the natural language processing (NLP) model used to generate and compose text. Access to the Faculty of Civil and Environmental Engineering compute cluster will enable the team to run and train language models.

Gdańsk Tech Startup School


 
51 views