[newsflash] The configuration of "Large-scale Parallel Computing System for Nation-wide Shared Use" system
The configuration of "Large-scale Parallel Computing System for Nation-wide Shared Use" system is following:
Large-scale Parallel Computing System for Nation-wide Shared Use (codename: octopus) | ||
---|---|---|
Theoretical Computing Speed | approximately 1.446 PFLOPS | |
Compute Node | CPU node | Intel CPU(Skylake)x2, 192GB memory |
GPU node | Intel CPU(Skylake)x2, NVIDIA GPU(P100)x4, 192GB memory | |
XeonPhi node | Intel XeonPhi(Knights Landing), 192GB memory | |
Large memory node | Intel CPU(Skylake)x8, 6TB memory | |
Interconecct | InfiniBand EDR | |
Strage | DDN Lustre(approximately 3 PB) |
* This is the newsflash on June 1, 2017. Therefore, there may a bit of a discrepancy about performance values. Thank you for your understanding.
We will announce on the following page, if there is any progress about contract or development.
Octopus
Posted : June 01,2017