To access a system a secure method, such as ssh, has to be used. This is normally run from a text console:

ssh login.curta.zedat.fu-berlin.de

Normally programs are run in batch mode. This means that a program is not started to directly, but instead is sent to a so-called resource manager. This ensures that the program is only then started when enough resources are available.

A nightly backup of the data in /home is made. Data stored in the /scratch file systems is not backed up and cannot be recovered if deleted by accident or lost due to a file system failure.

Please note that even in /home, files contained within a directory called tmp or temp will excluded from the backup.

An incremental backup is performed every night and a full backup made every 6 weeks. If you need to recover lost data, please contact us at hpc@zedat.fu-berlin.de.

Occasionally work needs to be done on the HPC systems which leads to services being restricted

When your FU account expires or you leave the group you specified in your account application, then your access to the HPC systems will be deactivated. All data on the HPC systems will be deleted after 4 weeks.

Please ensure that you move data that you still require to another system. Data which is no longer needed should be deleted.

Email

Please send email only to hpc@zedat.fu-berlin.de. Your message will reach all HPC support staff. If you write to an individual member of the team and that person is unavailable, you will not receive a prompt reply.

If you have observed a specific error with a job sent to the queuing system, please fill in this error report.

Telephone

If you would like to discuss any issue related to our service with us, please call one of the numbers given on the contacts page.

Personal Consultation

You may also visit us personally in order to receive advice about using our systems. However, you must make an appointment with us in advance. In general, office hours are as follows:

Available disk space can be queried via the command df:

$ df -h
Filesystem            Size  Used Avail Use% Mounted on
rootfs                 19G  5.9G   12G  33% /
/dev/sda1              19G  5.9G   12G  33% /
udev                   24G  216K   24G   1% /dev
tmpfs                  24G  606M   24G   3% /dev/shm
/dev/sda6             192G  188M  182G   1% /local
/dev/sda3             1.9G   95M  1.7G   6% /tmp
/dev/sda2             1.9G  879M  947M  49% /var
none                   24G  606M   24G   3% /dev/shm
master:/cm/shared     247G  125G  109G  54% /cm/shared
master:/home           17T   11T  5.8T  65% /home
fhgfs_nodev           164T  106T   58T  65% /scratch

Here the entries /home and /scratch are particularly important.

Disk space usage can be queried via the command du:

$ du -sh /{home,scratch}/joeuser
3.6G	/home/joeuser
20K	/scratch/joeuser

This provides a summary of the disk usage in the users home directory.

The following text may be used:

The authors would like to thank the HPC Service of ZEDAT, Freie Universität Berlin, for computing time.

Please also use the following DOI as a reference for the system Curta:

10.17169/refubium-26754

If the HPC Service has been helpful for you in producing a publication or thesis, please send us the relevant DOI and, if possible, an appropriate image to hpc@zedat.fu-berlin.de, so that we can include this in our list https://www.fu-berlin.de/sites/high-performance-computing/Forschungsergebnisse.