NAMD Hands On

First steps 

1. Copy namd_tutorial to your /scratch directory 
2. Enter the namd_tutorial and check the files. You will see an slm file and also 3 more files into water folder (your simulation inputs). 
3. Open the water_std.slm file and check the SLURM directives( #SBATCH ... ). 
4. Modify the water_std.slm file to request 12 cores, 3900Mb/core on standard nodes 
   1. Solution Expand source

      #SBATCH -n 12 
      #SBATCH --mem-per-cpu=3900Mb 
      #SBATCH –p std 

5. Modify the  water_std.slm file to work on $SCRATCH directory (it requieres to move inputs to $SCRATCH).

   1. Solution Expand source

      cd $SCRATCH 
      cp -r namd_tutorial/water* . 

6. Modify the  water_std.slm file to launch namd/2.13 by mpi (srun command) using the input files provided in namd_tutorial folder .
   
   1. Solution Expand source

      Module load namd/2.13 
      srun `which namd2` water.namd > water_std.out 

7. Modify the  water_std.slm file to move the output to a folder called outputs inside the original namd_tutorial folder. 

   1. Solution Expand source

      mkdir ~/namd_tutorial/water/output 
      cp water_std.out  ~/namd_tutorial/water/output/ 

8.  Launch the job. You can check the progress by the command “squeue”.  
   1. Solution Expand source

      Sbatch water_std.slm  

More specific 

1. Modify the water_std.slm file to run only on one node and 48 cores 
   1. Solution Expand source

      #SBATCH -N 1 
      #SBATCH -n 48 

GPU Version 

1. Now, you would to launch NAMD code on GPU (because you discovered that provides better performance). Copy namd_std.slm to namd_gpu.slm and edit it to run on one GPU. Remember that you need 24 cores on 1 Node, the GPU partition and request a GPU as a generic resource. 
   1. Solution Expand source

      #SBATCH -n 24 
      #SBATCH –N 1 
      #SBATCH -p gpu 
      #SBATCH -gres=gpu:1 

2. Edit your “Executable” and final sections to adapt it to a threaded GPU run (not MPI)

   1. Solution Expand source

      `which namd2` > water_gpu.out 
      mkdir ~/namd_tutorial/water/output 
      cp water_std.out  ~/namd_tutorial/water/output/

3. Compare the results between architectures. Which one is faster?