...
| Installation Phase | Phase 1 (decomissioned end of 2018) | Phase 2 (decomissioned end of 20182019) | ||
|---|---|---|---|---|
| Installation Date | 2011 | 2012 | 2013 | 2015 |
Islandtype | Fat Nodes | Thin Nodes | Many Cores Nodes | Haswell Nodes |
| System | BladeCenter HX5 | IBM System x iDataPlex dx360M4 | IBM System x iDataPlex dx360M4 | Lenovo NeXtScale nx360M5 WCT |
| Processor Type | Westmere-EX Xeon E7-4870 10C | Sandy Bridge-EP Xeon E5-2680 8C | Ivy-Bridge and Xeon Phi 5110P | Haswell Xeon Processor E5-2697 v3 |
| Nominal Frequency [GHz] | 2.4 | 2.7 | 1.05 | 2.62 |
| Performance per core | 4 DP Flops/cycle = 9.6 DP GFlop/s 2-wide SSE2 add + 2-wide SSE2 mult | 8 DP Flops/cycle = 21.6 DP GFlops/s 4-wide AVX add + 4-wide AVX mult | 16 DP Flops/cycle = | 16 DP Flops/cycle = |
| Total Number of nodes | 205 | 9216 | 32 | 3072 |
| Total Number of cores | 8,200 | 147,456 | 3,840 (Phi) | 86,016 |
| Total Peak Performance [PFlop/s] | 0.078 | 3.2 | 0.064 (Phi) | 3.58 |
| Total Linpack Performance [PFlop/s] | 0.065 | 2.897 | n.a. | 2.814 |
| Total size of memory [TByte] | 52 | 288 | 2.56 | 194 |
| Total Number of Islands | 1 | 18 | 1 | 6 |
| Typical Power Consumption [MW] | < 2.3 | ~1.1 | ||
| Components | ||||
| Nodes per Island | 205 | 512 | 32 | 512 |
| Processors per Node | 4 | 2 | 2 (IvyB) 2.6 GHz + 2 Phi 5110P | 2 |
| Cores per Processor | 10 | 8 | 8 (IvyB) + 60 (Phi) | 14 |
| Cores per Node | 40 | 16 | 16 (host) + 120 (Phi) | 28 |
| Logical CPUs per Node (Hyperthreading) | 80 | 32 | 32 (host) + 480 (Phi) | 56 |
| Memory and Caches | ||||
| Memory per Core [GByte] (typically available for applications) | 6.4 (~6.0) | 2 (~1.5) | 4 (host) + 2 x 0.13 (Phi) | 2.3 (2.1) |
| Graphical representation of processor topology | westmere.png | sandbridge.png | host.png phi.png | haswell.png haswell.big.png |
| Size of shared Memory per node [GByte] | 256 | 32 | 64 (host) + 2 x 8 (Phi) | 64 |
| Bandwidth to Memory per node [Gbyte/s] | 136.4 | 102.4 | Phi: 384 | 137 |
| Level 3 Cache Size (shared) [Mbyte] | 4x30 | 2x20 | 4x18 | |
| Level 2 Cache Size per core [kByte] | 256 | 256 | Phi: 512 | 256 |
| Level 1 Cache Size [kByte] | 32 | 32 | 32 | 32 |
| Latency Access Memory [cycles] / Bandwidth per core [GB/s] | ~160 /8.8 | ~200 / 6.7 | ||
| Level 3 Latency [cycles] /BW per Core [GB/s] | ~ 30 / 31 | 36 / 39 | ||
Level 2 Latency [cycles]1 /BW per Core [GB/s] | 12 / 42 | 12 / 92 | ||
Level 1 Latency [cycles]1 /BW per Core [GB/s] | 4 | 4 /130 | 4 / 343 | |
| Interconnect | ||||
| Technology | Infiniband QDR | Infiniband FDR10 | Infiniband FDR10 | Infiniband FDR14 |
| Intra-Island Topology | non-blocking Tree | non-blocking Tree | ||
| Inter-Island Topology | Pruned Tree 4:1 | n.a. | Pruned Tree 4:1 | |
| Bisection bandwidth of Interconnect [TByte/s] | 12.5 | 5.1 | ||
| Servers | ||||
| Login Servers for users | 2 | 7 | 1 | 5 |
| Storage | ||||
| Size of parallel storage (SCRATCH/WORK) [Pbyte] | 15 | |||
| Size of NAS storage (HOME) [PByte] | 3.5 (+ 3.5 for replication) | |||
| Aggregated bandwidth to/from parallel storage [GByte/s] | 250 | |||
| Aggregated bandwidth to/from NAS storage [GByte/s] | 12 | |||
| Capacity of Archive and Backup Storage [PByte] | > 30 | |||
| System Software | ||||
| Operating System | Suse Linux Enterprise Server (SLES) | |||
| Batchsystem | IBM Loadleveler | |||
| Parallel Filesystem for SCRATCH and WORK | IBM GPFS | |||
| File System for HOME | NetApp NAS | |||
| Archive and Backup Software | IBM TSM | |||
| System Management | xCat from IBM | |||
| Monitoring | Icinga, Splunk | |||
...