As we announced about changes to terms of service charge and seasonal coefficient(from Nov 1), soaring electricity bills due to soaring fuel costs are making it difficult to operate the center's large-scale computer system. Since the utilization rate of Large-scale shared-memory node (2 nodes) of OCTOPUS is low compared to other nodes, in order to more and more people can use other nodes (General purpose CPU node and GPU node) which the utilization rate is higher than others, we will operate Large-scale shared-memory node by application system as follows from Nov 4.
OCTOPUS Large-scale shared-memory node (from Nov 4, 2022)
- User ID
- Schedule of start day and end day of using (up to 2 weeks)
- A number of nodes you preferred (1 node or 2 nodes)
- Elapsed time per 1 job (please let us know the value specified for #PBS -l elapstim_req=)
* Until now, the upper limit was 120 hours per 1 job, but if you contact us, we will change it flexibly.
- Total number of jobs
You cannot use it without contacting us, since you can possess the node and use it practically during the period of application, please feel free to apply.
Please refer to the following for the consumption points of Large-scale shared-memory node.
In addition, since Large-scale shared-memory node includes not only rich memory but also rich CPU with "128 cores" per 1 node.
You can use it for test runs, so if you are interested, please feel free to contact us.
We are making every effort not to stop Supercomputers. Thank you for your understanding and cooperation. Also, please apply for additional resources.