RAxML-NG

Official website : https://github.com/amkozlov/raxml-ng

1 Available version
2 Job submission script
- 2.1 OpenMP
- 2.2 OpenMP+MPI
  - 2.2.1 1 node
  - 2.2.2 > 1 node

Available version

Version	Module	OpenMP	OpenMP+MPI

Version	Module	OpenMP	OpenMP+MPI
1.2.0	RAxML-NG/1.2.0-cpeGNU-23.03	raxml-ng	raxml-ng-mpi

Job submission script

OpenMP

#!/bin/bash
#SBATCH -p compute     # specific partition (compute,memory)
#SBATCH -N 1 -c 128    # -N 1 must be 1, -c can be 16, 32, 64 depending on the requirement
#SBATCH -t 24:00:00    # job time limit <hr:min:sec>
#SBATCH -A lt200xxx    # project ID
#SBATCH -J RAxML-NG    # job name

module restore
module load RAxML-NG/1.2.0-cpeGNU-23.03

raxml-ng --all --msa aligned.fasta --prefix output_prefix --model LG+G8+F --bs-trees 200 --thread $SLURM_CPUS_PER_TASK

OpenMP+MPI

1 node

#!/bin/bash
#SBATCH -p compute       # specific partition (compute,memory)
#SBATCH -N 1 -n 8 -c 16  # specify computing resource (1 node, 8 MPI ranks, 16 threads per rank)
#SBATCH -t 24:00:00    # job time limit <hr:min:sec>
#SBATCH -A lt200xxx    # project ID
#SBATCH -J RAxML-NG    # job name

module restore
module load RAxML-NG/1.2.0-cpeGNU-23.03

srun raxml-ng-mpi --all --msa aligned.fasta --prefix output_prefix --model LG+G8+F --bs-trees 200 --thread $SLURM_CPUS_PER_TASK

> 1 node

#!/bin/bash
#SBATCH -p compute        # specific partition (compute,memory)
#SBATCH -N 2 -n 16 -c 16  # specify computing resource (2 nodes, 16 MPI ranks, 16 threads per rank)
#SBATCH -t 24:00:00    # job time limit <hr:min:sec>
#SBATCH -A lt200xxx    # project ID
#SBATCH -J RAxML-NG    # job name

module restore
module load RAxML-NG/1.2.0-cpeGNU-23.03

srun raxml-ng-mpi --all --msa aligned.fasta --prefix output/run128mpi_1 --model LG+G8+F --bs-trees 200 --thread $SLURM_CPUS_PER_TASK