You are located in service: RWTH High Performance Computing (Linux)

Using partitions

Using partitions

In general, you do not need to explicitly request a partition in your job script with the #SBATCH -p tag. Typically, the appropriate partition is automatically selected based on your computing time project defaults and other specifications in your job script.

However, in the following cases you might need to specify the partition explicitly:

Special Memory Requirements on Claix-2023

Jobs scheduled on Claix-2023 are usually assigned to the c23ms partition. The nodes in this partition have 2625 MB per core. If your job requires more memory per core and you do not want to pay extra, you can specify the c23mm for 5300 MB per core or c23ml for 10650 MB per core. Note that this may result in increased waiting times.

Testing Applications

The devel partition is designed for quick testing of computing jobs, offering short wait times but limited run times. You can find more information here.

GPU Partition

Usually, specifying the GPU partition is unnecessary, as your job will be automatically scheduled there when GPU resources are requested.

 

Overview

The following table gives an overview of the currently available partitions:

 

partition

#nodes#cores per node#mem per noderemarks
c23ms63296256Claix-2023 (small memory partition)
c23mm16096512Claix-2023 (medium memory partition)
c23ml2961024Claix-2023 (large memory partition)
c23g5296256Claix-2023 GPU partition with four H100 GPUs per node
c18m124048192default partition for the "default" project
c18g54481922 V100 gpus, request of volta gpu needed to submit to this partition
devel848192Designed for testing jobs and programs. Maximum runtime:  1 Hour
Has to be used without an project!
Maximum of Simultanous Jobs!
Further information can be found here
 

last changed on 04/29/2024

How did this content help you?

Creative Commons Lizenzvertrag
This work is licensed under a Creative Commons Attribution - Share Alike 3.0 Germany License