Login
ssh -Y skx.supermuc.lrz.de -l ge54cel3
ssh - Secure Shell on LRZ HPC Systems
[1] https://doku.lrz.de/display/PUBLIC/ssh+-+Secure+Shell+on+LRZ+HPC+Systems
According to the webside, first we do the “Configuration of SSH public-key authentication for access to LRZ systems from outside”.
- open the “Public keys for SuperMUC-NG login nodes”, and copy the content;
- add into the ~/.ssh/known_hosts file (you may need to generate the .ssh directory if it does not exist yet) on your local machine;
- do according the Client (your desktop) is a LINUX system
Generating public/private ecdsa key pair.
ssh-keygen -t ecdsa -b 521 -a 100
Enter passphrase,e.g. hereiam
ssh-copy-id -i /home/songchen/.ssh/id_ecdsa ge54cel3@skx.supermuc.lrz.de
If successful, the command’s response will contain:
Number of key(s) added: 1
Check the status of the SuperMUC-NG
[2] https://doku.lrz.de/display/PUBLIC/SuperMUC-NG+Status
Upload or download files
- sent DATA FROM LRZ to local computer
scp ge54cel3@skx.supermuc.lrz.de:/dss/dsshome1/0E/ge54cel3/testcase/temp.plt ~/Downloads
- sent data from local computer to LRZ
scp -r ~/code ge54cel3@skx.supermuc.lrz.de:/dss/dsshome1/0E/ge54cel3/sparta
scp cannot be invoked on the LRZ system, because all outgoing connections are blocked. You have to use it on your local computer.
Compile the Sparta
- check module
module list module show mpi.intel/2019
- find the libraries
echo $MPI_LIB
- Set the path in the Makefile.foo
MPI_INC = -I/lrz/sys/intel/impi2019u6/impi/2019.6.154/intel64/include /usr/include/c++/4.8 MPI_PATH = -L/lrz/sys/intel/impi2019u6/impi/2019.6.154/intel64/lib/
MPI_LIB = /lrz/sys/intel/impi2019u6/impi/2019.6.154/intel64/lib/libmpicxx.so.12 - compile the code
make foo
The sbatch Command
[3] https://doku.lrz.de/display/PUBLIC/Job+Processing+with+SLURM+on+SuperMUC-NG
#!/bin/bash # Job Name and Files #SBATCH -o ./myjob_output.%j.%N.out #SBATCH -D ./ #SBATCH -J test # #Notification and type #send a notice to email when end: SBATCH --mail-type=END #SBATCH --mail-user=insert_your_email_here # #Setup of execution environment #SBATCH --export=NONE #SBATCH --get-user-env #SBATCH --account=insert your_projectID_here #SBATCH --partition=insert test, micro, general, large or fat # #SBATCH --nodes=10 # --- available cores per node 48 --- #SBATCH --ntasks-per-node=48 # # Wall clock limit: #SBATCH --time=04:00:00 #SBATCH --no-requeue # module load slurm_setup # #module load gcc/4.8 # #cd $SCRATCH/mydata # cd /dss/dsshome1/0E/ge54cel3/testcases/unsteady/case1 mpirun -n $SLURM_NTASKS ./spa_foo < in.config ---
List of relevant commands
submit a job script
sbatch jobname
delete or terminate a queued or running job
scancel job_id
print table of submitted jobs and their state.
squeue -p test
create an interactive SLURM shell
salloc
execute argument command on the resources assigned to a job. Note: must be executed inside an active job (script or interactive environment). mpiexec is an alternative and preferred on LRZ system
srun
Display various status information of a running job/step.
sstat
provide overview of cluster status
sinfo
uery and modify SLURM state
scontrol q
(sacct is not available for users. The detailed usage of these commands can be found in Webpage Ref.[3])