BayeScan - 2.1

Basic information

Installation

This entry covers the entire process performed in the installation and test of BayeScan on a cluster with the conditions described below.

Usage

In the following example we are going to run a test included in the source code of BayeSacan. Note that we are using one single compute-node, with 16 threads.

slurm.sh
#!/bin/bash
#SBATCH --partition=longjobs
#SBATCH --job-name=bayescan_test
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=16
#SBATCH --time=30:00
#SBATCH --output=bayescan_test%j.slurm.log

# Load module for default version
module load bayescan

bayscan -snp ~/bayescan/BayeScan2.1/input_examples/test_genotype_SNP.txt -all_trace -out_prilot pilot

To run the previous example.

sbatch bayescan_run.sh

Note

In this example we are using the default version of BayeScan module. We recommend to specify the version. To use the version of this entry, pleace load the module as follow:

In Apolo II:

module load bayescan/2.1_intel-2017_update-1

In Cronos:

module load bayescan/2.1_intel-18.0.2

Performance Tests

The following test was performed in Cronos. We were comparing the build-in version and our builded version described in this entry. As input, we use one of the included tests of BayeScan source.

sbatch script for build-in version test:

#!/bin/bash
#SBATCH --partition=longjobs
#SBATCH --job-name=bayescan_test
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=16
#SBATCH --time=30:00
#SBATCH --output=bayescan_test%j.slurm.log

echo -n "Started at:"
date

# Run with build-in binaries 
time ~/BayeScan2.1/binaries/BayeScan2.1_linux64bits -snp ~/bayescan/BayeScan2.1/input_examples/test_genotype_SNP.txt -all_trace -out_prilot pilot

The test for our builded version was executed by the example provided in Usage section of this entry.

Note

As you can see in the sbatch script, the time was calculated using time command.

Results

Compiled version Time
  Real User Sys
Build-in 7m3.124s 46m39.444s 6m46.893s
Intel (icmp) 2m8.054s 33m26.063s 0m42.262s
change 99.970% 28.341% 89.613%

Authors