Point Consumption Criteria
Please click here for more information on the point consumption guideline.
Criteria for Standard Queues
System B has a CPU node with a typical configuration and a FAT node with large memory capacity, each with cpu and fat included in the queue name. The number of points consumed when using each queue is described as follows.
Queue name | Execution time limit | Number of nodes (minimum to maximum) | Points consumed per day |
---|---|---|---|
i8cpu | 30 minutes | 1〜8 | 1 point/node |
F/L1cpu | 24/120 hours | 1〜1 | 1 point/node |
F/L4cpu | 24/120 hours | 2〜4 | 1 point/node |
F/L16cpu | 24/120 hours | 5〜16 | 1 point/node |
F/L36cpu | 24/120 hours | 17〜36 | 1 point/node |
F/L72cpu | 24/120 hours | 72〜72 | 1 point/node |
F/L144cpu | 24/120 hours | 144〜144 | 1 point/node |
i1fat | 30 minutes | 1〜1 | 4 point/node |
F/L2fat | 24/120 hours | 1〜2 | 4 point/node |
Disk Charges
Points are required depending on the amount of disk space used on the system.
Points are not consumed when usage remains within the standard quota (default quota).
However, should you raise the quota above the standard value and the usage exceeds the default quota, points will be consumed on a daily basis for any amount in excess of the standard value.
Quota changes can be set in units of 1 GB, and point consumption is also calculated in units of 1 GB/day.
The billing factor for points is different in the /home and /work areas.
System B
Queue name | Standard value | Maximum value | Points consumed per day |
---|---|---|---|
Long term (/home) | 300 GB | 9 TB | 1.0 point/TB |
Short term (/work) | 6 TB | 240 TB | 0.1 point/TB |
Standard point consumption guidelines
Points are distributed to each assignment based on these point consumption standards and the total compute capacity, with the expectation that the supercomputer will be in continuous operation throughout the year.
For System B:
(4pt [per Fat node per day] x 8 Fat Node+1pt [per CPU node per day] x 1680 CPU Nodes) x 365 days ~ 625 kpt