Differences

This shows you the differences between two versions of the page.

--- production_run [2016/09/20 12:15] – external edit 127.0.0.1
+++ production_run [2022/09/28 17:23] (current) – admin
@@ Line 1: / Line 1: @@
 [[bilayer_simulations]]
-Here are production run instructions for DT1/DT2.
+==== Production run instructions for Zaratan ====
-Listed below are some instructions for the production run simulation. As we discussed, this will automatically run your simulation in chunks with each being a certain length.  The files to copy (not the sub-directories) are in the following location on DT1/DT2 to your directory with step6.1 to step6.6 namd files:
-for DT1, **/homes/jbklauda/sim-inputs/namd/npt-membrane**
+Listed below are some instructions for the production run simulation on Zaratan. As we discussed, this will automatically run your simulation in chunks with each being a certain length.  The files to copy (not the sub-directories) are in the following location on Zaratan to your directory with step6.1 to step6.6 namd files:
-for DT2, **/homes/jbklauda/sim-inputs/namd**
+**/home/jbklauda/sim-inputs/namd**
-There are key things to change in the dyn.csh file
+There are key things to change in the dyn.csh file:
-. # number of tasks and nodes to run your job (talk to Dr. Klauda)
+  - # number of nodes to run your job (talk to Dr. Klauda)
+  - You should adjust the job name to something for your system.
+  - Adjust the SBATCH and namd run lines per the number of nodes (see examples in the the following directory: ~jbklauda/sim-inputs/namd)
+  - Also there is a line with set  simt = 1000000. This is to run 1 Million steps of 2ns per block (with 2fs timesteps).  Discuss with Dr. Klauda and/or a graduate student on what to change this to and also change in the dyn-2.inp file.
-. You should adjust the job name to something for your system.
+You then will need to adjust dyn-1.inp and dyn-2.inp to match with what step7_production.inp has. These files will be combined into a dyn.inp file with the firsttimestep defined (see dyn.csh file). The easiest way to do this is to copy and paste the psf/pdb lines in step7*inp to replace that in dyn-1.inp. Then, copy most of the step7*inp to replace the material in dyn-2.inp. One key important change will likely be the values for the "run" and "numsteps" lines in dyn-2.inp based on how many ns per dyn file. These should be set to the simt value.
-. mpirun -np command with the same number of tasks
+Note: dyn-1.inp inputname should stay as dyn
-. Also there is a line with set  simt = 1000000. This is to run 1 Million steps of 2ns per block.  Discuss with Dr. Klauda and/or a graduate student on what to change this to and also change in the dyn-2.inp file.
-You then will need to adjust dyn-1.inp and dyn-2.inp to match with what step7_production.inp has.  This file will be combined into a dyn.inp file with the firsttimestep defined (see dyn.csh file).  One key important change could be the final line in dyn-2.inp. This should be set to the simt value.
 Once the changes are made you will also need to copy the step*6.6.xsc/.vel/.coor to dyn.xsc/vel/coor so the job will use these to restart from your equilibration runs. Then to run the production run, just type: