The end of the fiscal year is approaching, and the usage of general-purpose CPU nodes is increasing, resulting in longer wait times for execution.
Therefore, starting from 10:00 AM on Wednesday, January 15, after the scheduled system maintenance period in January
Downtime: January 12, 2025 (Sunday) 10:00 - January 15, 2025 (Wednesday) 10:00
The cloud bursting feature implemented in SQUID will be enabled, allowing some small-scale jobs to be executed using computing resources on Oracle Cloud Infrastructure (OCI), Oracle Corporation's public cloud.
By utilizing this feature, there is a possibility of executing small jobs with shorter wait times compared to usual. Those wishing to run jobs eligible for cloud bursting, please check the following notes and make use of the feature.
Target
Currently, this feature is available to users of the SQUID general-purpose CPU node share queue (SQUID-S), where offloading to OCI computational resources is being executed.
Point Consumption
When executing a job using OCI computational resources, the points consumed will be equivalent to the "general-purpose CPU node share queue (SQUID-S)." There will be no additional high charges, so you can use the feature with peace of mind. More details are provided below.
How to Use Cloud Bursting
To use OCI computational resources, add options to the job submission shell script (job script file) and submit it to the scheduler. Below is an example of a job script:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 |
#!/bin/bash #------- qsub options ----------- #PBS -q SQUID-S # The only queue available for cloud bursting is SQUID-S. #PBS --group=G01234 # Specify the group name you belong to. #PBS -l cpunum_job=36 # Request the number of CPU cores to use. #PBS -l elapstim_req=01:00:00 # Specify the maximum job execution time. *Jobs exceeding one hour are not eligible for cloud bursting. #PBS --enable-cloud-bursting=yes # Allow cloud bursting for this job. #PBS -U cloud_wait_limit=04:00:00 # If the wait time exceeds the specified limit (e.g., 4 hours), the job may run on the cloud. #------- Program execution ----------- module load BaseCPU # Load the base environment. cd $PBS_O_WORKDIR # Move to the directory where qsub was executed. ./a.out > result.txt # Execute the program and redirect the output to a file. |
Notes
1. The hardware configuration of the OCI compute nodes differs from that of the SQUID general-purpose CPU nodes, which may result in different computation times and results.
・SQUID general-purpose CPU node:
Processor: Intel Xeon Platinum 8368 (Icelake / 2.40 GHz, 38 cores) x 2
Memory: 256GB
・OCI general-purpose CPU node:
Processor: Intel Xeon 6354 (Icelake / 3.0 GHz, 18 cores) x 2
Memory: 512GB
2. Even if cloud bursting is enabled and OCI computational resources are used, the points consumed will be the same as the general-purpose CPU node share queue (SQUID-S).
That is, the calculation follows the formula:
Job Execution Time x General-purpose CPU Node Share Queue Consumption Coefficient (0.2248) x General-purpose CPU Node Fuel Coefficient x General-purpose CPU Node Seasonal Coefficient.
3. By enabling the cloud bursting feature, you agree to the following points:
- Computation may be performed using OCI computational resources.
- Depending on resource availability, OCI resources may not be used, and computation may only occur on SQUID CPU nodes.
4. The cloud node mounts Lustre via NFS, which results in lower I/O performance compared to SQUID general-purpose CPU nodes with Lustre mounted directly.
Therefore, programs with heavy I/O processing may experience delays in computation time.
We hope you can make use of the cloud bursting feature after confirming the above information.
Posted : January 10,2025