AI Supercomputer Sirius-325

AI Supercomputer Sirius-325

JPryamid has implemented its proprietary ‘AI in a Box’ to establish a learning environment to accelerate the up-skilling of your workforce in the latest techniques in machine learning and deep learning.  The environment provides a multi-user platform to provide ‘in context’ problems that provide the ideal scenarios to teach your people how to use Artificial Intelligence for business benefit.  Teaching the end to end process – from data ingestion, through data orchestration, into the key processes required to build and deploy AI models.

Take advantage of our in-depth expertise to accelerate the training of your next generation data engineers and data scientists.

 

Software Specifications:

Programming Languages: Python 2.7.12/3.5.2, Java 1.8.0_162, Scala 2.11.12, R 3.1.1

Frameworks : TensorRT 3.0 GA, cuDNN 7.0.5, VisionWorks 1.6, CUDA 9.0, Multimedia API, L4T, OpenGL ES 2.0, 3.0, 3.1, 3.2, OpenGL 4.3-4.6

Operating System: Linux Ubuntu 18.04

 

Hardware Specifications:

Cyber-Security

SonicWall NSA 2650 High Availability – security appliance

NAT support, VPN support, load balancing, Stateful Packet Inspection (SPI), DoS attack prevention, Intrusion Prevention System (IPS), URL filtering, deep inspection, DDos attack prevention, Stateful switchover (SSO), anti-spam protection, anti-malware protection, Quality of Service (QoS), RADIUS support, DHCP server, DNS proxy, DHCP relay, NetFlow, IPFIX, 2 fans

DES, Triple DES, MD5, SHA-1, 128-bit AES, 192-bit AES, 256-bit AES

RADIUS, internal user database, XAUTH authentication

IEEE 802.3, IEEE 802.1p

 

TPU-GOOGLE

12 TPU X Google Coral Dev Board Edge TPU Module (SOM) with ARM-64 Bit Quad symmetric Cortex-A53 processors,
Google Edge TPU ML Accelerator Coprocessor & Cryptographic coprocessor with 32 GFLOPs 32-bit or 64 GFLOPs 16-bit

12 X 1GB LPDDR4 Memory.

12 X 8GB eMMC Storage.

 

GPU-ARM

ARM 64-bit – NVIDIA Xavier™ Architecture GPU with unparalleled 32 TeraOPS (TOPS) of peak compute
& 750Gbps of high-speed I/O in a compact 100x87mm form-factor.

4 GPU x 512= 2054 GPU core Volta with 64 Tensor Cores.

4 CPU x 8-core ARM v8.2 64-bit CPU, 8MB L2 + 4MB L3 @ 2265MHz.

4 x 16GB 256-bit LPDDR4x @ 2133MHz | 137GB/s Memory.

4 X Dual NVIDIA Deep Learning Accelerators (DLAs) with 10W / 15W / 30W profiles, 9.0V-20VDC input.

 

CPU-INTEL

8 CPU X Intel 8th Generation Core, 4 Core, 3400 MHz.

8 x 4 CPU Cores, 32GB RAM DDR4 Memory.

8 x 500 SSD

 

Storage:

4 X 32 GB eMMC Flash Storage

12 X 14 TB HDD Storage (IronWolf 3.5 Inch 7200 RPM Internal Hard Drive (256 MB Cache, 180 TB/Year Workload Rate, Up to 210 MB/s)), 168 TB HDD

8 x 500 GB SSD, 4TB SSD

1 X 1TB SSD

 

Network:

All units have WLAN, Bluetooth and RJ45

10/100/1000 BASE-T Ethernet

 

Other:

1 X 27” Monitor

1 x Keyboard and Mouse

2 x Cooling Fans (120mm Fan with Speed Control Silent 5V Computer Fan Portable USB Cooler Dual Ball Bearing Ventilator Radiator)

 

Date

06/06/2019

Category

AI Supercomputer