Parts/CPUs/IBM PowerPC 440 700MHz
Why this part
An embedded core, of the kind used in routers and disk controllers, running at 700 MHz when desktop parts were passing 3 GHz. That was the entire point. BlueGene/L was designed against a power budget rather than a per-node performance target, on the observation that performance per watt improves as you move down the frequency curve, and that a machine which draws less can simply contain more processors in the same building. 106,496 of these beat every faster chip on the list. The argument BlueGene made in 2004 is the argument every hyperscale design has made since.
Systems using this part
- Systems
- 11
- Rmax underneath
- 767.28 TFlop/s
- Units recorded
- 133,120
| System | Role | Country | Units | System Rmax | Supplied by | Rank |
|---|---|---|---|---|---|---|
| New York Blue 2007-06 | CPU | USA | - | 82.16 TFlop/s | IBM | #10 → |
| KEK Blue Gene System B 2006-03 | CPU | Japan | - | - | IBM | - |
| CyBlue 2006-01 | CPU | USA | 2,048 | 4.71 TFlop/s | IBM | #99 → |
| JUBL 2006 | CPU | Germany | 16,384 | 37.33 TFlop/s | IBM | #8 → |
| CERT Blue Gene 2006 | CPU | United Arab Emirates | - | 4.73 TFlop/s | IBM | #139 → |
| Stella 2005-04 | CPU | Netherlands | - | 27.45 TFlop/s | IBM | #6 → |
| BGW 2005 | CPU | USA | - | 91.29 TFlop/s | IBM | #2 → |
| Blue Protein 2005 | CPU | Japan | 8,192 | 18.20 TFlop/s | IBM | #8 → |
| Intimidata 2004-12 | CPU | USA | - | 13.78 TFlop/s | IBM | #42 → |
| BlueGene/L 2004-11 | CPU | USA | 106,496 | 478.20 TFlop/s | IBM | #1 → |
| Blue Fern | CPU | New Zealand | - | 9.43 TFlop/s | IBM | #99 → |
“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.