LS-DYNA

LS-DYNA#

Ansys LS-DYNA is the industry-leading explicit simulation software used for applications like drop tests, impact and penetration, smashes and crashes, occupant safety, and more. More information about LS-DYNA can be found on the LS-DYNA website.

Running LS-DYNA#

LS-DYNA can be run in two modes: SMP (Shared Memory Processing) and MPP (Massively Parallel Processing). The choice of mode depends on the size and complexity of your simulation. SMP is suitable for smaller problems that can efficiently utilise up to 8 cores on a single node, while MPP is designed for larger problems that require distributed memory and can scale across multiple nodes.

Guidance provided by Oasys suggests that SMP can scale to ~8 cores, and that the MPP version should be used for larger problems.

Remember to replace <port>@<host> with the appropriate license server details.

SMP single precision (4 cores)#

#!/bin/bash
#SBATCH --job-name=lsdyna_smp_sp
#SBATCH --time=01:00:00
#SBATCH --mem=8G
#SBATCH --nodes=1
#SBATCH --cpus-per-task=4

module add ansys/2023R2
export ANSYSLMD_LICENSE_FILE=<port>@<host>
export LSTC_LICENSE=ANSYS

lsdyna I=example.k ncpu=$SLURM_CPUS_PER_TASK

SMP double precision (4 cores)#

#!/bin/bash
#SBATCH --job-name=lsdyna_smp_dp
#SBATCH --time=01:00:00
#SBATCH --mem=8G
#SBATCH --nodes=1
#SBATCH --cpus-per-task=4

module add ansys/2023R1
export ANSYSLMD_LICENSE_FILE=<port>@<host>
export LSTC_LICENSE=ANSYS

lsdyna -dp I=example.k ncpu=$SLURM_CPUS_PER_TASK

MPP single precision (whole node)#

#!/bin/bash
#SBATCH --job-name=lsdyna_mpp_sp
#SBATCH --time=04:00:00
#SBATCH --mem=256G
#SBATCH --nodes=1
#SBATCH --ntasks=168

module add ansys/2023R1
export ANSYSLMD_LICENSE_FILE=<port>@<host>
export LSTC_LICENSE=ANSYS

lsdyna -dis -np $SLURM_NTASKS -lsdynampp I=example.k

MPP double precision (whole node)#

#!/bin/bash
#SBATCH --job-name=lsdyna_mpp_dp
#SBATCH --time=04:00:00
#SBATCH --mem=256G
#SBATCH --nodes=1
#SBATCH --ntasks=168

module add ansys/2023R1
export ANSYSLMD_LICENSE_FILE=<port>@<host>
export LSTC_LICENSE=ANSYS

lsdyna -dis -np $SLURM_NTASKS -lsdynampp -dp I=example.k

Note

For optimal performance:

  • Use the SMP version for smaller problems that can efficiently utilise up to 8 cores on a single node.

  • Use the MPP version for larger problems that require distributed memory and can scale across multiple cores or nodes.

  • Adjust the number of CPUs (--cpus-per-task), memory (--mem), and tasks (--ntasks) based on your simulation requirements.

Contents