Deep Learning Lab

The Deep Learning Lab has been established in the CREDIT Center at Prairie View A&M University since 2017. The Deep Learning Lab features four NVIDIA DGX-1 Deep Learning systems totaling 32 Tesla P100 GPUs with 114,688 CUDA cores, 2,752 GB memory, and 244 TB HDD. Each DGX-1 system has eight Tesla P100 GPU accelerators connected through NVLink, the NVIDIA high-performance GPU interconnect, in a hybrid cube-mesh network. Together with dual socket Intel Xeon CPUs and four 100 Gb InfiniBand network interface cards, DGX-1 provides unprecedented performance for deep learning training. Moreover, the DGX-1 system software and powerful libraries are tuned for scaling deep learning on its network of Tesla P100 GPUs to provide a flexible and scalable platform for deep learning. In addition, the Deep Learning Lab also contains one Dell R920 server for user and resource management, and four Dell R730 storage servers to maintain large datasets.

A newly established Deep Learning Lab
  • 4 NVIDIA DGX-1 systems: Each DGX-1 is equivalent of 250 servers in a box: 7TB of SSD storage, 8 Tesla P100 / V100 GPUs (28672 CUDA cores) and two Xeon processors
  • 1 Dell R920 management node
  • 4 Dell R730/R730xd storage servers (> 400TB)
  • GPU workstations: 2 EXXACT Valence Workstation with GeForce Titan V; 2 Dell Precision T7600 with GeForce Titan X; 1 Supermicro GPU SuperWorkstation 7048GR-T with 3 Tesla K10
  • Embedded Devices: 7 NVIDIA Jetson TX2 Developer Kits; 7 Xilinx PYNQ-Z1 FPGA Board; 10 Raspberry Pi 3
Cloud Computing Lab established since 2011
  • 24 SMP nodes with 12 cores, 128GB memory each (4 GPU nodes)
  • 8 SMP nodes with 16 cores, 128GB memory each
  • 4 SMP nodes with 32 cores, 64GB memory each
  • 54 IBM blade servers with 4 cores, 4GB memory each