The Dortmund University of Applied Sciences and Arts needs a high-performance Deep Learning Server (DL Server) for the approved BMBF project ReMeDi:Blut. As part of the project, machine learning will be used to predict complications of blood transfusions, among other things. The server will contain 8 CUDA-capable 80GB GPGPUs and two 56 core (112 threads) CPUs and have 2 TB of main memory (RAM). CUDA-capable GPUs are required, as common machine learning software and optimized containers currently only support CUDA. Among other things, the project aims to analyze complete genome information, which is why multi-GPU models with maximum memory capacity and maximum interconnect are required.

The Dortmund University of Applied Sciences and Arts needs a high-performance Deep Learning Server (DL Server) for the approved BMBF project ReMeDi:Blut. As part of the project, machine learning will be used to predict complications of blood transfusions, among other things. The server will contain 8 CUDA-capable 80GB GPGPUs and two 56 core (112 threads) CPUs and have 2 TB of main memory (RAM). CUDA-capable GPUs are required, as common machine learning software and optimized containers currently only support CUDA. Among other things, the project aims to analyze complete genome information, which is why multi-GPU models with maximum memory capacity and maximum interconnect are required.