Sample SLURM Scripts

From UFRC
Jump to: navigation, search
Hpg2 wiki logo.png

HiPerGator 2.0 documentation

Sample SLURM Scripts

Below are a number of sample scripts that can be used as a template for building your own SLURM submission scripts for use on HiPerGator 2.0. These scripts are also located at: /ufrc/data/training/SLURM/, and can be copied from there. If you choose to copy one of these sample scripts, please make sure you understand what each line of the sbatch directives before using it to submit your jobs. Otherwise, you may not get the result you want and may waste valuable computing resources.

Note: There is a maximum limit of 3000 jobs per user.

Memory requests

A large number of users request far more memory than their jobs use (100-10,000 times!). As an example, since August 1st, looking at groups that have run over 1,000 jobs, there are 28 groups whose users have requested 100x the memory used in over half of those jobs. Groups often find themselves with jobs pending due to having reached their memory limits (QOSGrpMemLimit).

While it is important to request more memory than will be used (10-20% is usually sufficient), requesting 100x, or even 10,000x, more memory only reduces the number of jobs that a group can run as well as overall throughput on the cluster. Many groups, and our overall user community, will be able to run far more jobs if they request more reasonable amounts of memory.

The email sent when a job finishes shows users how much memory the job actually used and can be used to adjust memory requests for future jobs. The SLURM directives for memory requests are the --mem or --mem-per-cpu. It is in the user’s best interest to adjust the memory request to a more realistic value.

Requesting more memory than needed will not speed up analyses. Based on their experience of finding their personal computers run faster when adding more memory, users often believe that requesting more memory will make their analyses run faster. This is not the case. An application running on the cluster will have access to all of the memory it requests, and we never swap RAM to disk. If an application can use more memory, it will get more memory. Only when the job crosses the limit based on the memory request does SLURM kill the job.

Basic, single-processor job

This script can serve as the template for many single-processor applications. The mem-per-cpu flag can be used to request the appropriate amount of memory for your job. Please make sure to test your application and set this value to a reasonable number based on actual memory use. The %j in the -o (can also use --output) line tells SLURM to substitute the job ID in the name of the output file. You can also add a -e or --error with an error file name to separate output and error logs.

Download the single_core_job_annotated.sh script

#!/bin/bash
#SBATCH --job-name=serial_job_test    # Job name
#SBATCH --mail-type=END,FAIL          # Mail events (NONE, BEGIN, END, FAIL, ALL)
#SBATCH --mail-user=email@ufl.edu     # Where to send mail	
#SBATCH --ntasks=1                    # Run on a single CPU
#SBATCH --mem=1gb                     # Job memory request
#SBATCH --time=00:05:00               # Time limit hrs:min:sec
#SBATCH --output=serial_test_%j.log   # Standard output and error log
pwd; hostname; date
 
module load python
 
echo "Running plot script on a single CPU core"
 
python /ufrc/data/training/SLURM/plot_template.py
 
date


Threaded or multi-processor job

This script can serve as a template for applications that are capable of using multiple processors on a single server or physical computer. These applications are commonly referred to as threaded, OpenMP, PTHREADS, or shared memory applications. While they can use multiple processors, they cannot make use of multiple servers and all the processors must be on the same node.

These applications required shared memory and can only run on one node; as such it is important to remember the following:

  • You must set --ntasks=1, and then set --cpus-per-task to the number of OpenMP threads you wish to use.
  • You must make the application aware of how many processors to use. How that is done depends on the application:
    • For some applications, set OMP_NUM_THREADS to a value less than or equal to the number of cpus-per-task you set.
    • For some applications, use a command line option when calling that application.


Download the multicore_job.sh script

#!/bin/bash
#SBATCH --job-name=parallel_job      # Job name
#SBATCH --mail-type=END,FAIL         # Mail events (NONE, BEGIN, END, FAIL, ALL)
#SBATCH --mail-user=email@ufl.edu    # Where to send mail	
#SBATCH --ntasks=1                   # Run a single task	
#SBATCH --cpus-per-task=4            # Number of CPU cores per task
#SBATCH --mem=1gb                    # Job memory request
#SBATCH --time=00:05:00              # Time limit hrs:min:sec
#SBATCH --output=parallel_%j.log     # Standard output and error log
pwd; hostname; date
 
echo "Running prime number generator program on $SLURM_CPUS_ON_NODE CPU cores"
 
module load gcc/5.2.0 
 
/ufrc/data/training/SLURM/prime/prime
 
date


Another example, setting OMP_NUM_THREADS:

Download the multi_processor_job2.sh script

#!/bin/bash
#SBATCH --job-name=parallel_job_test # Job name
#SBATCH --mail-type=END,FAIL         # Mail events (NONE, BEGIN, END, FAIL, ALL)
#SBATCH --mail-user=email@ufl.edu    # Where to send mail	
#SBATCH --ntasks=1                   # Run a single task	
#SBATCH --cpus-per-task=4            # Number of CPU cores per task
#SBATCH --mem=600mb                  # Total memory limit
#SBATCH --time=00:05:00              # Time limit hrs:min:sec
#SBATCH --output=parallel_%j.log     # Standard output and error log
date;hostname;pwd
 
export OMP_NUM_THREADS=4
 
module load intel
 
./YOURPROGRAM INPUT
 
date

MPI job

Note
If you are using OpenMPI 3.0.0 or 3.1.0 use srun --mpi=pmix_v1. For OpenMPI-3.1.2, use srun --mpi=pmix_v2.

This script can serve as a template for MPI, or message passing interface, applications. These are applications that can use multiple processors that may, or may not, be on multiple compute nodes.

Our testing has found that it is best to be very specific about how you want your MPI ranks laid out across nodes and even sockets (multi-core CPUs). SLURM and OpenMPI have some conflicting behavior if you leave too much to chance. Please refer to the full SLURM sbatch documentation, but the following directives are the main directives to pay attention to:

  • -c, --cpus-per-task=<ncpus>
    • Advise the Slurm controller that ensuing job steps will require ncpus number of processors per task.
  • -m, --distribution=arbitrary|<block|cyclic|plane=<options>[:block|cyclic|fcyclic]>
    • Specify alternate distribution methods for remote processes.
    • We recommend -m cyclic:cyclic, which tells SLURM to distribute tasks cyclically over nodes and sockets.
  • -N, --nodes=<minnodes[-maxnodes]>
    • Request that a minimum of minnodes nodes be allocated to this job.
  • -n, --ntasks=<number>
    • Number of tasks (MPI ranks)
  • --ntasks-per-node=<ntasks>
    • Request that ntasks be invoked on each node
  • --ntasks-per-socket=<ntasks>
    • Request the maximum ntasks be invoked on each socket
    • Notes on socket layout:
      • hpg2-compute nodes have 2 sockets, each with 16 cores.
      • hpg1-compute nodes have 4 sockets, each with 16 cores.


The following example requests 24 tasks, each with one core. It further specifies that these should be split evenly on 2 nodes, and within the nodes, the 12 tasks should be evenly split on the two sockets. So each CPU on the two nodes will have 6 tasks, each with its own dedicated core. The distribution option will ensure that MPI ranks are distributed cyclically on nodes and sockets.

SLURM is very flexible and allows users to be very specific about their resource requests. Thinking about your application and doing some testing will be important to determine the best request for your specific use.

Download the mpi_job.sh script

#!/bin/bash
#SBATCH --job-name=mpi_job_test      # Job name
#SBATCH --mail-type=END,FAIL         # Mail events (NONE, BEGIN, END, FAIL, ALL)
#SBATCH --mail-user=email@ufl.edu    # Where to send mail	
#SBATCH --ntasks=24                  # Number of MPI ranks
#SBATCH --cpus-per-task=1            # Number of cores per MPI rank 
#SBATCH --nodes=2                    # Number of nodes
#SBATCH --ntasks-per-node=12         # How many tasks on each node
#SBATCH --ntasks-per-socket=6        # How many tasks on each CPU or socket
#SBATCH --distribution=cyclic:cyclic # Distribute tasks cyclically on nodes and sockets
#SBATCH --mem-per-cpu=600mb          # Memory per processor
#SBATCH --time=00:05:00              # Time limit hrs:min:sec
#SBATCH --output=mpi_test_%j.log     # Standard output and error log
pwd; hostname; date
 
echo "Running prime number generator program on $SLURM_JOB_NUM_NODES nodes with $SLURM_NTASKS tasks, each with $SLURM_CPUS_PER_TASK cores."
 
module load intel/2018.1.163 openmpi/3.0.0
 
srun --mpi=pmix_v1 /ufrc/data/training/SLURM/prime/prime_mpi
 
date

Hybrid MPI/Threaded job

This script can serve as a template for hybrid MPI/Threaded applications. These are MPI applications where each MPI rank is threaded and can use multiple processors.

Our testing has found that it is best to be very specific about how you want your MPI ranks laid out across nodes and even sockets (multi-core CPUs). SLURM and OpenMPI have some conflicting behavior if you leave too much to chance. Please refer to the full SLURM sbatch documentation, as well as the information in the MPI example above.

The following example requests 8 tasks, each with 4 cores. It further specifies that these should be split evenly on 2 nodes, and within the nodes, the 4 tasks should be evenly split on the two sockets. So each CPU on the two nodes will have 2 tasks, each with 4 cores. The distribution option will ensure that MPI ranks are distributed cyclically on nodes and sockets.

Download the hybrid_pthreads_job.sh script

#!/bin/bash
#SBATCH --job-name=hybrid_job_test      # Job name
#SBATCH --mail-type=END,FAIL            # Mail events (NONE, BEGIN, END, FAIL, ALL)
#SBATCH --mail-user=email@ufl.edu       # Where to send mail	
#SBATCH --ntasks=8                      # Number of MPI ranks
#SBATCH --cpus-per-task=4               # Number of cores per MPI rank 
#SBATCH --nodes=2                       # Number of nodes
#SBATCH --ntasks-per-node=4             # How many tasks on each node
#SBATCH --ntasks-per-socket=2           # How many tasks on each CPU or socket
#SBATCH --mem-per-cpu=100mb             # Memory per core
#SBATCH --time=00:05:00                 # Time limit hrs:min:sec
#SBATCH --output=hybrid_test_%j.log     # Standard output and error log
pwd; hostname; date
 
module load  intel/2018.1.163  openmpi/3.0.0 raxml/8.2.12
 
srun --mpi=pmix_v1 raxmlHPC-HYBRID-SSE3 -T $SLURM_CPUS_PER_TASK \
      -f a -m GTRGAMMA -s /ufrc/data/training/SLURM/dna.phy -p $RANDOM \
      -x $RANDOM -N 500 -n dna
 
date

The following example requests 8 tasks, each with 8 cores. It further specifies that these should be split evenly on 4 nodes, and within the nodes, the 2 tasks should be split, one on each of the two sockets. So each CPU on the two nodes will have 1 task, each with 8 cores. The distribution option will ensure that MPI ranks are distributed cyclically on nodes and sockets.

Also note setting OMP_NUM_THREADS so that OpenMP knows how many threads to use per task.

Download the hybrid_OpenMP_job.sh script

#!/bin/bash
#SBATCH --job-name=LAMMPS
#SBATCH --output=LAMMPS_%j.out
#SBATCH --mail-type=END,FAIL
#SBATCH --mail-user=<email_address>
#SBATCH --nodes=4              # Number of nodes
#SBATCH --ntasks=8             # Number of MPI ranks
#SBATCH --ntasks-per-node=2    # Number of MPI ranks per node
#SBATCH --ntasks-per-socket=1  # Number of tasks per processor socket on the node
#SBATCH --cpus-per-task=8      # Number of OpenMP threads for each MPI process/rank
#SBATCH --mem-per-cpu=2000mb   # Per processor memory request
#SBATCH --time=4-00:00:00      # Walltime in hh:mm:ss or d-hh:mm:ss
date;hostname;pwd
 
module load intel/2018 openmpi/3.1.0
 
export OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK
 
srun --mpi=pmi_v1 /path/to/app/lmp_gator2 < in.Cu.v.24nm.eq_xrd
 
date
  • Note that MPI gets -np from SLURM automatically.
  • Note there are many directives available to control processor layout.
    • Some to pay particular attention to are:
      • --nodes if you care exactly how many nodes are used
      • --ntasks-per-node to limit number of tasks on a node
      • --distribution one of several directives (see also --contiguous, --cores-per-socket, --mem_bind, --ntasks-per-socket, --sockets-per-node) to control how tasks, cores and memory are distributed among nodes, sockets and cores. While SLURM will generally make appropriate decisions for setting up jobs, careful use of these directives can significantly enhance job performance and users are encouraged to profile application performance under different conditions.

Array job

Please see the SLURM_Job_Arrays page for information on job arrays. Note that we use the simplest 'single-threaded' process example from above and extending it to an array of jobs. Modify the following script using the parallel, mpi, or hybrid job layout as needed.

Download the array_job.sh script

#!/bin/bash
#SBATCH --job-name=array_job_test   # Job name
#SBATCH --mail-type=FAIL            # Mail events (NONE, BEGIN, END, FAIL, ALL)
#SBATCH --mail-user=email@ufl.edu   # Where to send mail	
#SBATCH --ntasks=1                  # Run a single task
#SBATCH --mem=1gb                   # Job Memory
#SBATCH --time=00:05:00             # Time limit hrs:min:sec
#SBATCH --output=array_%A-%a.log    # Standard output and error log
#SBATCH --array=1-5                 # Array range
pwd; hostname; date
 
echo This is task $SLURM_ARRAY_TASK_ID
 
date

Note the use of %A for the master job ID of the array, and the %a for the task ID in the output filename.

GPU job

Please see GPU_Access for more information regarding the use of HiPerGator GPUs.

#!/bin/bash
#SBATCH --job-name=vasptest
#SBATCH --output=vasp.out
#SBATCH --error=vasp.err
#SBATCH --mail-type=ALL
#SBATCH --mail-user=email@ufl.edu
#SBATCH --nodes=1
#SBATCH --ntasks=8
#SBATCH --cpus-per-task=1
#SBATCH --ntasks-per-node=8
#SBATCH --distribution=cyclic:cyclic
#SBATCH --mem-per-cpu=7000mb
#SBATCH --partition=gpu
#SBATCH --gres=gpu:tesla:4
#SBATCH --time=00:30:00
 
module purge
module load cuda/10.0.130  intel/2018  openmpi/4.0.0 vasp/5.4.4
 
srun --mpi=pmix_v3 vasp_gpu