From Montana Tech High Performance Computing

Revision as of 11:43, 23 August 2021 by Bdeng (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

HPC Architecture

The Montana Tech HPC (oredigger cluster) contains 1 management node, 26 compute nodes, and a total of 91 TB NFS storage systems. There is an additional computing server (copper). Twenty-two compute nodes contain two 8-core Intel Xeon 2.2 GHz Processors (E5-2660) and either 64 or 128 GB of memory. Two of these nodes are GPU Nodes, with three NVIDIA Tesla K20 accelerators and 128 GB of memory. Hyperthreading is enabled, so 704 threads can run simultaneously on just the XEON CPUs. The remaining four nodes feature the Intel 2nd Generation Xeon Scalable Processors (48 CPU Cores and 192 GB Ram per node). Internally, a 40 Gbps InfiniBand (IB) network interconnects the nodes and the storage system.

The system has a theoretical peak performance of 14.2 TFLOPS without the GPUs. The GPUs alone have a theoretical peak performance of 7.0 TFLOPS for double precision floating point operations. So the entire cluster has a theoretical peak performance of over 21 TFLOPS.

The operating system is Centos 7.6 and Penguin's Scyld ClusterWare is used to maintain and provision the compute nodes.

Head Node
CPU Dual E5-2660 (2.2 GHz, 2x 8-cores)
Disk 1 TB SSD
Copper Server
CPU Dual E5-2643 v3 (3.4 GHz, 2x 6-cores)
RAM 128 GB
Disk 1 TB
Other Specs
NFS storage nfs0 - 25 TB
nfs1 - 66 TB
Network Ethernet
40 Gbps InfiniBand
14 Compute Nodes
CPU Dual E5-2660 (2.2 GHz, 2x 8-cores)
Disk 450 GB
Nodes cn0~cn11, cn13, cn14
6 Compute Nodes
CPU Dual E5-2660 (2.2 GHz, 2x 8-cores)
RAM 128 GB
Disk 450 GB
Nodes cn12, cn15~cn19
2 GPU Nodes
CPU Dual E5-2660 (2.2 GHz, 2x 8-cores)
RAM 128 GB
Disk 450 GB
GPU Three nVidia Tesla K20
Nodes cn20, cn21


4 Compute Nodes
CPU Dual Xeon Platinum 8260 (2.40 GHz, 2x 24-cores)
RAM 192 GB
Disk 256 GB SSD
Nodes cn31~cn34

3D Visualization System

Montana Tech is developing two 3D data visualization systems. Both systems provide an immersive visualization experience (aka virtual reality) through 3D stereoscopic imagery and user tracking systems. These systems allow scientists to directly interact with their data and helps them gain a better understanding of their data generated modeling on the HPC Cluster or collected in the field. Remote data visualization is possible by running VisIt from the cluster's login node.

Windows Immersive 3D Visualization System
CPU Dual E5-2643 v4 (3.4 GHz, 2x 6-cores)
Disk 512 GB SSD + 1TB HD
GPU Dual nVidia Quadro K5000
OS Windows 7
Display 108" 3D projector screen
Linux IQ Station
CPU Dual E5-2670 (2.60 GHz, 8-cores)
RAM 128 GB
Disk 4 TB
GPU Dual nVidia Quadro 5000
OS CentOS 6.4
Display 70" 3D TV