HPC_facilities hardware and operating system

The GCP High Performance Computing (HPC) facility functions as a webservice and consists of a global grid of 4 cluster systems (brochure).
The GCP HPC consists of a global grid of 4 cluster systems based on Paracel dual-processor nodes, 1.8 GHz 64-bit AMD Opteron CPUs, 4GB memory. All systems are configured with Rocks Linux ( a variant of Red Hat Linux) and (commercial) Platform LSF cluster and multicluster (grid) management software. The management node of each system includes the basic Linux LAMP program suite.
The GCP high performance computing (HPC) cluster/grid facilities are installed at CIP (Peru), IRRI Philippines), ICRISAT (India) and at ILRI, Nairobi. A joint purchase and commissioning of Paracel Linux cluster computers was made in 2004, largely funded by Generation CP funds (plus some site specific funding, e.g. at IRRI and ILRI).

• CIP: 4 nodes (1 master, 3 slave), 1.8 GHz dual AMD Opteron, 4 GB memory per node

• IRRI: 9 nodes(1 master + mirror, 7 slave), 2.0 GHz dual AMD Opteron, 4 GB memory per node

• ICRISAT: 4 nodes (1 master, 3 slave), 1.8 GHz dual AMD Opteron, 4 GB memory per node

• ILRI: 33 nodes (1 master, 32 slaves), 1.8 GHz dual AMD Opteron, 4 GB memory per node

HPC Project Manager is Anthony Collins at CIP a.collins@cgiar.org.


HPC FAQ

HPC software faclities

back
back to the GCP bioinformatics portal page