Past systems
This section introduces the changes to the large-scale general-purpose computer systems and supercomputer systems that have been in service at the Information Technology Center (Computer Center) from the past.
System configuration
Systems
General-purpose computer systems
| Operational period | Model (main storage capacity) |
|---|---|
| 1966 ～ 1973 | HITAC 5020 (16 KW) / 5020E (65 KW) + HITAC 5020 (32 KW) |
| 1973 ～ 1980 | HITAC 8800 × 3 + HITAC 8700 × 1 (8 MB) |
| 1980 ～ 1982 | HITAC M-200H × 8 (64 MB) |
| 1982 ～ 1985 | HITAC M-280H × 6 (192 MB) |
| 1986 ～ 1990 | HITAC M-680H (128 MB) + HITAC M-682H (384 MB) |
| 1991 ～ 1996 |
HITAC M-880/310 (MS: 1 GB, ES: 4 GB) + HITAC M-682H (256MB) |
| 1996 ～ 2001 | HITACHI MP5800/320 (MS + ES : 10GB) |
| 2001 ～ 2007 | Operational support system (MS + ES: 10 GB) |
*1W = 4 byte, MS = Main storage device, ES = Extended storage device
Vector-supercomputers
| Operational period | Model (main storage capacity) | Theoretical computational performance |
|---|---|---|
| 1983 ～ 1987 |
HITAC S-810/20 (MS : 256 MB, ES : 512 MB) |
800 MFlops |
| 1988 ～ 1993 | HITAC S-820/80 (MS : 512 MB, ES : 6 GB) | 4 GFlops |
| 1993 ～ 1999 | HITAC S-3800/480 (MS : 2 GB, ES : 16 GB) | 32 GFlops |
(Massively) Parallel supercomputers
| Operational period | Model (main storage capacity) | Theoretical computational performance |
|---|---|---|
| 1996 ～ 2001 |
HITACHI SR2201 (1024 nodes, 256 MB/nodes) | 307.2 GFlops |
| 1999 ～ 2005 |
HITACHI SR8000 (128 nodes, 8 GB/nodes) |
1,024.0 GFlops |
| 2001 ～ 2007 | HITACHI SR8000/MPP (144 nodes, 16 GB/nodes) | 2,073.6 GFlops |
| 2005 ～ 2007 | HITACHI SR11000/J1 (44 nodes, 128 GB/nodes) | 5,350.4 GFlops |
| 2007 ～ 2011 | HITACHI SR11000/J2 (128 nodes, 128 GB/nodes) | 18,841.6 GFlops |
| 2008 ～ 2014 | HITACHI HA8000-tc/RS425 (952 nodes) | 140,134.4 GFlops |
| 2011 ～ 2017 | HITACHI SR16000/M1 (Compute nodes: 56 nodes, 200 GB/node) | 54,906.4 GFlops |
| 2012 ～ 2018 | FUJITSU PRIMEHPC FX10 (Compute nodes: 4800 nodes) | 1.135 PFlops |
| 2014 ～ 2018 | FUJITSU PRIMEHPC FX10 (Compute nodes: 576 nodes) | 136.2 TFlops |
| 2016 ～ 2020 | SGI Rackable C2112-4GP3 (Compute nodes: 420 nodes, 256 GB/node) | 508.03 TFlops |
| 2017 ～ 2021 | SGI Rackable C1102-GP8 (Compute nodes: 120 nodes(GPU 240), 256 GB/node) |
1418.2 TFlops |
| 2017 ～ 2021 | SGI Rackable C1102-GP8 (Compute nodes: 64 nodes(GPU 256), 256 GB/node) |
1435.3 TFlops |
| 2016 ～ 2022 | Fujitsu PRIMERGY CX1640 M1 (Compute nodes: 8,208 nodes, 96 GB(DDR4)+16 GB(MCDRAM)/node) |
25.004 PFlops |