LAB

Multiprocessor Array Platform

Description

The Multiprocessor Array (MPA) Platform consists of a General-Purpose Graphics Processing Unit (GP-GPU) and a compatible host workstation/server, or multi-core processor technology featured for parallel computing such as Intel Knights Landing. The platform enables researchers to accelerate processing of algorithms and simulations, and explore topics such as multi-core software authoring, modeling, computer architecture, and novel approaches to power-efficient high-performance computing.

The first generation of the MPA includes the NVIDIA Tesla K20 GP-GPU that is powered by CUDA parallel-computing technology.

The second generation of the MPA includes two configurations: the NVIDIA Tesla K40 GP-GPU with compatible workstation and the Intel Xeon Phi 7120 coprocessor with compatible workstation.

The third generation of the MPA includes four configurations: NVIDIA Tesla K40, NVIDIA Tesla P100, Intel Phi 7120, and Intel Knights Landing, with compatible host computer.

Resources

Benefits

High performance and power efficiency
Accelerates research in the areas of massive data processing, image processing, scientific modeling, parallel computing architecture exploration etc.
Enables a faster path to commercialization by sourcing commercial-grade components and through compliance with commercial standards and interfaces ( PCI-express)

Features

MPA Generation I

NVIDIA Tesla K20 GPU card
Host workstation—Lenovo Thinkstation S30
- CPU: Intel Xeon E5 2620 Six Core 2.0 GHz Processor
- 16 Gbyte Memory
- Nvidia Quadro K600 Graghics Card w 1 Gbyte
- Two 1 Tbyte Hard Drives @ 7200 rpm
- Dual Integrated 100/1000 Ethernet
- Ten USB 2.0, two USB 3.0 Ports
- MS Windows 7 Professional

MPA Generation II

Intel Xeon Phi based configuration
- Intel Xeon Phi 7120 A Coprocessor
- Host workstation—Lenovo Thinkstation P900
  - CPU: two E5 2600v3 processors
  - 15 MB CPU cache per processor
  - 64 GB ECC 1600 MHz LP RDIMM system memory
  - Nvidia Quadro K420 Graghics Card
  - Storage: 500 GB SATA 1st Solid State (SSD), two 1TB SATA 7200 2nd HDDs
  - Gigabit Ethernet dual port
  - Three PCIe 3.0 x16
  - Linux Redhat Enterprise 6.6
NVIDIA Tesla K40 based configuration
- NVIDIA Tesla K40 GPU card
- Host workstation—custom built machine
  - CPU: E5 2630 v2 processor
  - 15 MB CPU cache
  - 32 GB ECC 1600MHz LP RDIMM system memory
  - eVga GeForce 210 Graghics Card
  - Storage: 500 GB SATA 1st Solid State (SSD), two 1TB SATA 7200 2nd HDDs
  - Gigabit Ethernet dual port
  - Three PCIe 3.0 x16, 3 x PCIe 3.0 x 8
  - Dual-boot: Windows 7 Pro. 64-bit and CentOS 6.x

MPA Generation III

Intel Xeon Phi 7120 based configuration
The Xeon Phi 7120 based MPA systems delivered in GIII have the same features as that were delivered in GII.
Intel Xeon Knights Landing based configuration
There are three configurations delivered based on Intel Xeon Knights Landing processor, in which all three configurations has the same type of processor (16GB, 1.30 GHz, 64 core), but with different size of memory, storages and power supply, and different type of NICs.
- Quad-node rackmount system (Colfax KNL Ninja Rackmount Developer Platform)
  - MEMORY: 96GB / Node
  - EXPANSION: Riser 1: 1x PCIe 3.0 x16, Riser 2: 1x PCIe Gen3 x 20 (x16 or x4) / Node
  - LAN: 2x Intel i210 Gigabit Ethernet / Node
  - STORAGE: one 240GB SSD and one 4TB SATA drive
  - POWER SUPPLY: 2x 2130W Common Redundant 80 Plus Platinum
  - CentOS 7.2 on all nodes
  - Intel® Parallel Studio XE Cluster Edition Named User License
- Tower system 1 & 2 (Colfax KNL Ninja Pedestal Developer Platform)
  - MEMORY: 96GB
  - EXPANSION: 2x PCIe 3.0 x16, 1x PCIe 3.0 x4
  - LAN: 2x Intel i350 Gigabit Ethernet
  - STORAGE for tower system 1: 2 x 1.6 TB SSDs and 2 x 6TB HDDs
  - STORAGE for tower system 2: 1 x 800GB SSD, 1 x 240GB SSD, 1 x 4TB HDD, and 1 x 2 TB HDDs
  - POWER SUPPLY: 1x 750W
  - CentOS 7.2
  - Intel Parallel Studio XE Cluster Edition Named User License
NVIDIA Tesla K40 based configuration
This Tesla K40 based system has similar configurations to the Tesla K40 based systems delivered in GII, except that the workstation delivered in GIII is HP Z840, and the OS pre-installed on it is CentOS 6.9.
NVIDIA Tesla P100 based configuration
- NVIDIA Tesla P100 GPU card
- Host Server—Dell PowerEdge R730
  - CPU: 2 x Intel Xeon E5-2637 v4
  - 15 MB CPU cache
  - 64GB 2400MT/s RDIMM system memory
  - Storage: 2 x 800 GB SATA 1st Solid State (SSD)
  - Gigabit Ethernet quad ports
  - Power supply: Redundant Power Supply (1+1), 1100W
  - CentOS 6.9

LAB Support

Does your research benefit from products and services provided by CMC Microsystems?

LAB

Multiprocessor Array Platform

Description

Resources

Benefits

Features

MPA Generation I

MPA Generation II

MPA Generation III

LAB Support

Open source design platforms for accelerated system development.

Gate-based Quantum
Computing Using IBM Q

Quantum Computing at your fingertips:

CMC Planned Service Disruption

LAB

Multiprocessor Array Platform

Description

Resources

Benefits

Features

MPA Generation I

MPA Generation II

MPA Generation III

LAB Support

Open source design platforms for accelerated system development.

Gate-based Quantum Computing Using IBM Q

Quantum Computing at your fingertips:

CMC Planned Service Disruption

Gate-based Quantum
Computing Using IBM Q