HPC Resources/Armcluster

From VI-SEEM Wiki
Jump to: navigation, search

In 2004, in the Institute for Informatics and Automation Problems of NAS RA (IIAP) the first high Performance computing cluster (Armenian Cluster - Armcluster) in Armenia had been developed, which consists of 128 Xeon 3.06 GHz (64 nodes) processors. Myrinet high bandwidth and Gigabit networks interconnect the nodes of the cluster. The Myrinet network is used for computation and Gigabit for task distribution and management. The cluster achieved 523.4 GFlops performance by HPL (High Performance Linpack) test. Many intellectual software packages to support the advance in the field of modeling and analysis of quantum systems, signal and image processing, theory of radiation transfer, calculation of time constants for bimolecular chemical reactions, a system of mathematically proved methods, fast algorithms and programs for solving of certain classes of problems in linear algebra, calculus, algebraic reconditibility, test-checkable design of the built-in control circuits have been developed. In addition, user friendly tools and scientific gateways have been developed for Armcluster.


Technical Specification

Administrative Data
Name Armcluster
Short Description Armenian Cluster
Owner Institute for Informatics and Automation Problems of NAS RA
Country Armenia
Computational Power
Number of servers 64
Server specification Intel Xeon 3.06.GHz
CPU per server 2 x Intel Xeon 3.06.GHz
RAM per server 2 GB
Total number of CPU-cores 128
Max number of parallel processes 128
Interconnect type Myrinet2000
Interconnect latency 3.2 us
Interconnect bandwidth 4 Gbps
Local filesystem type NFS
Total storage (TB) 40 GB
Accelerators type N/A
Number of cores N/A
Accelerators per server N/A
Servers equipped with accelerators N/A
Peak performance CPU (Tflops) 783.36
Peak performance accelerators (Tflops) N/A
Peak performance (Tflops) N/A
Real performance (Tflops) N/A
Operating system Scientific Linux
Version 6.5
Batch system/scheduler Torque/Maui
Development tools GNU Compiler, MPICH-2
Libraries Blas, ScalaPack, MKL
Dedication to VI-SEEM
CPU (percent) 10%
Storage (percent) 10%
Accelerators (percent) N/A
CPU (core-hours per year) 112,128
Storage in TB 2
Accelerators (hours per year) N/A
System operational since 2005
Available to the project from PM04
System to be phased out N/A
Interfaces SSH