The collaborative program between USU and Erasmus + CBHE resulted in the Master Program in Data Science and Artificial Intelligence. Students studying in this program are required to become experts in data analysis by mastering theory and practice. Therefore, a data processing practice platform is needed that is used by students and lecturers during this program. For now, at DS&AI USU, the Jupyter Hub platform is available which is installed on a server that has a GPU so that students are expected to be able to process and analyze data.USU procured the GPU server and other equipment in December 2020. The specification of the server and other equipment are as the following:
The equipment specification:
|
No. |
Specification |
Quantity |
|
1 |
SUPERMICRO SuperServer 7049GP-TRT • Processor: 2 x CLX-SP 4214 12C/24T (2.20 GHz) Cache 16.5MB 9.6GT 2UPI (Dual Socket) • Chipset: Intel C621 • Memory: 4 x 32 GB DDR4-2933 ECC REG DIMM (Max 2.048 GB/16 DIMMs) • Hard Disks: 2 x 8 TB 7.2K RPM 3.5” SATA 6Gbps Cache 256MB 512E Hot Swap • SSD: 1 x Samsung SM883 480GB SATA 6Gbps V4 MLC 2.5" 7mm 3.6DWPD • DVD-RW + Floppy: No• Expansion Slots: PCI-Express 3.0 x 16 (6 Slots) & PCI-Express 3.0 x 4 (1 Slot) • HDD Drive Bays: 8 x 3.5” Hot Swap • Intel C621 Controller for 10 SATA3 (6Gbps) Ports, RAID 0, 1 & 10 (Onboard) • LAN: Dual 10GBase-T LAN Ports with Intel X550 • A speed AST2500 BMC• IPMI 2.0 + KVM with Dedicated LAN: Support for Intelligent Platform Management Interface v.2.0, IPMI 2.0with virtual media over LAN and KVM-over-LAN support • GPU: 4 x NVIDIA PNY GeForce RTX 2080TI 11GB GDDR6 PCIe 3.0 • Software: 1 x SFT-DCMS-Single (Data Center Management Package) • Monitor & Keyboard/Mouse: No • Power Supply: Redundant 2200W |
1 |
|
2 |
PC AIO ASUS V272 1 |
1 |
|
3 |
Notebook ASUS S333JQ Standard Memory: 16GB LPDDR4X on boardStorage: "512GB M.2 NVMe™ PCIe® 3.0 SSDGraphics: Intel® Iris™ Plus GraphicsDisplay: 13.3" Full HD (1920 x 1080) |
1 |
|
4 |
Epson Printer WF-7711 USB, Ethernet, Wi-Fi, Wi-Fi Direct and NFCDuplex printing up to A3 sizeEpson DURABriteTM Ultra ink4.3” Colour LCD Touch PanelPrint, Scan, Copy Fax with ADFThis server is used by students and lecturers at DS&AI USU Master Programmeas a platform for practicing programming. Each student and lecturer are given access rights to log into the JupyterHub platform installed on this server. |
1 |
This server is used by students and lecturers at DS&AI USU Master Program as a platform for practicing programming. Each student and lecturer are given access rights to log into the Jupyter Hub platform installed on this server.
On this server, Linux Container (LXC / LCD) is installed first, so that later this server can be used for many applications that support data processing, for example by providing applications for learning Data Modeling and Management and Business Intelligence. This server also has CUDA installed to support parallel computing capabilities of the NVIDIA GPU. Various programming support libraries for Python-based data processing have also been installed on the Jupyter Hub platform so that users are not bothered by the libraries they don't need.
The GPU server is co-located at the University Data Center which is mounted on a 4U Rack. Pictures of the installation are given in the following figures, 4U Rackmount at USU Data Center.