Parts/CPUs/Intel Xeon E5-2670
CPUs · Sandy Bridge-EP · 2012
Intel Xeon E5-2670
8-core Sandy Bridge Xeon at 2.60GHz, the volume server part that SDSC's Gordon and UNAM's Miztli both converged on within a year of each other simply because it was the merchant x86 chip a cluster bought in 2011 and 2012.
Why this part
Launched Q1 2012 at a $1,552 to $1,556 recommended price and formally end-of-lifed by 2015, it is the chip generation just before the E5 v2/v3/v4 refreshes that dominate the rest of this dataset's early-2010s academic clusters. Gordon and Miztli bought it independently an ocean apart, one for a flash-heavy NSF data-intensive system, the other for UNAM's national cluster, which says less about the chip than about how few merchant options a cluster buyer had in that window.
Systems using this part
- Systems
- 27
- Rmax underneath
- 13.08 PFlop/s
- Units recorded
- 10,178
| System | Role | Country | Units | System Rmax | Supplied by | Rank |
|---|---|---|---|---|---|---|
| Sky Bridge 2015-01 | CPU | USA | - | 532.90 TFlop/s | Intel | #123 → |
| Moordiup 2013-11 | CPU | Australia | 110 | - | Intel | - |
| TEIDE-HPC 2013-11 | CPU | Spain | - | 273.97 TFlop/s | Intel | #169 → |
| Conte 2013-07-30 | CPU | USA | - | 976.76 TFlop/s | Intel | #34 → |
| Spirit (AFRL) 2013-06 | CPU | USA | 9,216 | 1.42 PFlop/s | Intel | #15 → |
| OCuLUS (Paderborn) 2013-03 | CPU | Germany | - | 188.72 TFlop/s | Intel | #174 → |
| PARAM Yuva II 2013-02 | CPU | India | - | 388.44 TFlop/s | Intel | #70 → |
| Miztli 2013-01 | CPU | Mexico | - | 92.28 TFlop/s | Intel | #348 → |
| Aaditya 2013 | CPU | India | - | 719.22 TFlop/s | Intel | #37 → |
| Argus 2013 | CPU | Canada | - | 136.22 TFlop/s | Intel | #377 → |
| Cascade (PNNL) 2013 | CPU | USA | - | 2.54 PFlop/s | Intel | #14 → |
| Makman-1 2013 | CPU | Saudi Arabia | - | 441.82 TFlop/s | Intel | #61 → |
| Keeneland 2012-11 | CPU | USA | - | 319.56 TFlop/s | Intel | #75 → |
| Raijin (NCI) 2012-11 | CPU | Australia | - | 978.60 TFlop/s | Intel | #24 → |
| Midway (UChicago) 2012-11 | CPU | USA | - | 130.16 TFlop/s | Intel | #303 → |
| Gordon 2012-01 | CPU | USA | - | 218.10 TFlop/s | Intel | #48 → |
| Cab 2012 | CPU | USA | - | 347.40 TFlop/s | Intel | #63 → |
| HA-PACS 2012 | CPU | Japan | 268 | 421.60 TFlop/s | Intel | #41 → |
| MareNostrum 3 2012 | CPU | Spain | - | 925.06 TFlop/s | Intel | #30 → |
| Vilje 2012 | CPU | Norway | - | 396.70 TFlop/s | Intel | #56 → |
| Abel (UiO) 2012 | CPU | Norway | - | 178.60 TFlop/s | Intel | #134 → |
| Nimbus 2012 | CPU | USA | - | 82.69 TFlop/s | Intel | #433 → |
| Virgo (IIT Madras) 2012 | CPU | India | 584 | 91.13 TFlop/s | Intel | #224 → |
| SuperMike-II 2012 | CPU | USA | - | 110.93 TFlop/s | Intel | #250 → |
| Carter (Purdue) 2011-11 | CPU | USA | - | 186.90 TFlop/s | Intel | #54 → |
| Zin 2011 | CPU | USA | - | 773.70 TFlop/s | Intel | #15 → |
| Maia (NAS) | CPU | USA | - | 212.91 TFlop/s | Intel | #117 → |
“Units recorded” sums only the edges where a public source states a quantity. It is a floor, not a total, and should never be read as installed base.