Overview
High Performance Compute Cluster - hpc.uea.ac.uk
The cluster consists of the following components:
Node Names | Node Type | CPU | Cores (Total) | Clock Speed | Memory | Interconnect | IB Fabric | Queues | Usage |
e0001 to e0092 | Standard | Haswell (v3) | 16 | 2.6Ghz | 64GB DRR3 | 1Gb/s Eth | NA | *-eth | Standard workflows |
e0093 to e0216 | Standard | Broadwell (v4) | 16 | 2.1Ghz | 64GB DDR4 | 1Gb/s Eth | NA | *-eth | Standard workflows |
e0221 to e0236 |
Standard | Skylake | 24 | 2.1Ghz | 98GB DDR4 | 1Gb/s Eth | NA | eth-sky | Standard workflows |
i0001 to i0040 | Infiniband (IB) | Haswell (v3) | 24 | 2.3GHz | 128GB DDR3 |
1 x 40Gb/s (IB) 1 x 1Gb/s Eth |
Qlogic | *-ib | Parallel jobs and or I/O intensive/threaded jobs |
i0041 to i0081 | Infiniband (IB) | Broadwell (v4) | 28 | 2.4GHz | 128GB DDR4 |
1 x 56Gb/s (IB) 1 x 1Gb/s Eth |
Mellanox | mellanox-ib | Parallel jobs and or I/O intensive/threaded jobs |
i0082 to i0109 | Infiniband (IB) | Skylake | 24 | 2.1GHz | 98GB DDR4 |
1 x 56Gb/s (IB) 1 x 1Gb/s Eth |
Mellanox | sky-ib | Parallel jobs and or I/O intensive/threaded jobs |
g0001 to g0008 | GPU (Tesla K40) | Haswell | 16 | 2.6GHz | 64GB DDR4 | 1 x 1Gb/s Eth | NA | gpu | GPU aware jobs |
hm001 | Huge memory | Broadwell | 16 | 2.1GHz | 128GB | 1 x 1Gb/s Eth | NA | huge-mem | Large memory jobs |
hm002 | Huge memory | Broadwell | 16 | 2.1GHz | 512GB | 1 x 1GHz Eth | NA | huge-mem | Large memory jobs |
The HPC cluster was installed in the autumn of 2015, and is due to be decommissioned in 2020. For this cluster we partnered with OCF.
There have been phased introduction of new hardware every year - see the history section.