Product Overview
Featuring up to eight NVIDIA® Pascal P100 GPU’s, connected via SXM2 and NVLink in a 4U rackmount server, the Boston ANNA Pascal XL is perfect for high performance computing & deep learning. Dual Xeon E5-2600 v4 processors and a huge 3TB DDR-4 registered ECC memory capacity make this the perfect choice for intensive processing tasks, like neural network training, or scientific simulation.
Featuring up to eight NVIDIA Pascal P100 GPU's, connected via NVLink in a 4U rackmount server, the Boston ANNA Pascal XL is perfect for high performance computing & deep learning. Dual Xeon E5-2600 v4 processors and a huge 3TB DDR-4 registered ECC memory capacity make this the perfect choice for intensive processing tasks, like neural network training, or scientific simulation. A compliment of 16 x 2.5" hot pluggable disk bays allow for a range of high performance storage options too, including SAS, S-ATA, or even up to eight NVMe based devices for extreme IO requirements. With dual 10 Gigabit Ethernet ports on board, plus add on card options to include four 100GbE, EDR InfiniBand or Omni-Path ports with non-blocking PCI-Express connectivity, there are multiple networking options. Eight SXM2 slots support the latest generation of NVIDIA GPU's, including the Pascal based P100. NVLink connectivity of 80GB/s per link ensures that there is plenty of bandwidth between GPU's for even the most IO bound applications, reducing latency and the CPU as a bottleneck for transfers. Powered by four ground-breaking technologies with discontinuous jumps in performance, Tesla P100 enables lightning fast nodes to deliver the highest absolute performance for HPC and deep learning workloads with infinite computing needs. One of the most interesting features of Pascal is the new NVLink interconnect system. Today's leading deep learning models typically take days to weeks to train, forcing data scientists to make compromises between accuracy and time to deployment. The Boston ANNA Pascal XL, features 8 x NVIDIA® Tesla® P100 GPU accelerators for deep learning training, purpose-built to dramatically reduce training time. Running Caffe and Torch on the Tesla P100 delivers the same model within days versus weeks on CPU based compute systems.
Boston Ltd are also offering customers a dedicated optimised cloud platform for their deep learning analytics solutions - removing the initial capital expenditure of a physical on premise system, and enabling you to immediately gain insight through processing and analysing your data in the cloud. Our cloud platform enables agile sys-admin teams to quickly deploy scalable, production-ready deep learning environments in its own public cloud.
Key Features
Deep Learning & HPC GPU Accelerated Compute Solution 8 x SXM2 connected NVIDIA P100 GPU/s a in 4U NVLink connectivity between GPU's as standard Dual Intel Xeon E5-2600 v4 processors, up to 22 cores each Up to 3TB DDR-4 Registered ECC Memory EDR InfiniBand or Omni-Path interconnect options 16 x SAS/SATA or 8 x NVMe HDD / SSD support
-
Applications
Advanced Neural Network Analysis
Computational Fluid Dynamics
Deep Learning
High Performance Computing
Molecular Dynamics
-
BIOS Features
ACPI 1.0 / 2.0 / 3.0
APM 1.2
DMI 2.3
PCI 2.3
Plug and Play (PnP)
SMBIOS 2.7.1
UEFI
USB keyboard support
-
BIOS Type
128MB SPI Flash with AMI UEFI
-
Certifications
80 plus Titanium
-
Chipset
Intel C612
-
Colour
Black
-
CPU Cores
20
-
CPU Family
Intel Xeon E5
-
CPU Manufacturer
Intel
-
CPU Model
E5-2698 v4
-
CPU Quantity (Installed)
2
-
CPU Series
Intel Xeon E5-2600 v4
-
CPU Socket
Dual Socket R3 (LGA 2011)
-
CPU Speed
2.2 GHz
-
Drive Bays
16 x 2.5" Hot-swap SAS/SATA
-
Error Detection
Corrects single-bit errors
-
Expansion Slots
4 x Mellanox ConnectX-4 VPI adapter card, EDR IB (100Gb/s) and 100GbE, single-port QSFP, PCIe 3.0 x16
-
Form Factor
4U
-
GPU Family
NVIDIA Tesla Series
-
GPU Manufacturer
NVIDIA
-
GPU Memory Sizes
16MB
-
GPU Quantity
8
-
Input/Output
1 x RJ45 Dedicated IPMI LAN port
1 x Serial Header
1 x VGA Connector
2 x RJ45 10GBase-T port
4 x USB 3.0 ports
10 x SATA3 (6Gbps) ports
-
IPMI
Aspeed AST2400 BMC
IPMI 2.0 with virtual media over LAN and KVM-over-LAN support
Support for Intelligent Platform Management Interface v.2.0
-
Manufacturer
Boston
-
Memory (Installed)
512GB
-
Memory Sizes
32GB
-
Memory Slots
16
-
Memory Type
DDR4-2400
-
Memory Voltage
1.2V
-
Power Supply
2200W Redundant Titanium Level Power Supplies with PMBus
-
SATA Controller
SATA 3.0 6Gb/s with RAID 0, 1, 5, 10
-
SSD's (Installed)
4 x Samsung SSD 1.9Tb 2.5" SATA PM863a
-
System Bus
QPI up to 9.6 GT/s
-
System Cooling
1 x Air Shroud (MCP-310-84808-0B)
8 x 92mm Cooling Fans
-
Warranty
Mellanox 2 Year Extended Warranty for a total of 3 years Bronze for Adaptors