OCTOPUS will lower the upper limit of memory usage on May 10 to prevent abnormal system outages and stabilize the system. The following three node groups are targeted.

- OCTOPUS general-purpose CPU nodes
- OCTOPUS GPU nodes
- OCTOPUS Xeon Phi nodes

The memory usage limit for the above three node group is currently set at 190 GB, but this change will lower the limit to 185 GB. After this change, if the memory usage per node is specified as 190GB in the job script, an error will occur and the job cannot be submitted.
 

Please modify the specification in the job script to 185GB or remove this option.(If not specified, the maximum value is automatically set to 185GB.)
 

We apologize for the inconvenience.