06-640: Molecular Simulations Homework 3 Due Date: Thursday 2/22

06-640: Molecular Simulations Homework 3 Due Date: Thursday 2/22
06-640: Molecular Simulations
Homework 3
Due Date: Thursday 2/22
This aim of this homework is for you to run a Molecular Dynamics code. We will use the
Case Study codes provided with our textbook, so before doing the homework, you must
work through the steps outlined below to install the codes and learn how to run them.
Commands that you need to type and screen output are shown in Courier font.
Installation of the Case Study Source Codes
1. Log in to your andrew account on one of the unix.andrew.cmu.edu machines. While
in your home directory, execute the following commands
cp /afs/andrew.cmu.edu/course/06/640/www/Casestudies.tar.gz .
gunzip Casestudies.tar.gz
tar xvf Casestudies.tar
rm Casestudies.tar
These commands copy a complete set of the source codes for Frenkel & Smit’s case
studies into your account under a newly created directory called Frenkel_Smit. Have
a look at the directory structure that has been set up by moving into that directory,
cd Frenkel_Smit
and listing the contents of the directories by typing ls .
2. Change directories into the subdirectory for Case Study 4 (the first one we will use)
and list the contents of the directory:
cd CaseStudy_4
The structure of this directory is typical for all the case studies. The subdirectory
Source contains the source codes and a script for compiling them. The subdirectory
Run contains a script for running the executable code and various input and output files
for the code.
3. As an example of how to compile the case study codes, we will compile Case Study
4, a code that performs MD simulations of a Lennard Jones system. First, move into
the directory containing the source codes and clean up the existing object files:
cd ~/Frenkel_Smit/CaseStudy_4/Source
rm *.o
The latter step is necessary because the codes come precompiled for a specific Linux
operating system and we need to make sure the executable is compiled for the specific
operating system we are using. To compile the code, simply type
This will sequentially compile all the necessary codes and link them together. It may also
generate a few warnings from the FORTRAN compiler. In the end, you should get a very
brief message telling you the compilation was complete (typically it will say done).
4. To successfully run Case Study 4, we also need to compile the codes used for block
averaging of the data. To do this, execute the following series of commands:
cd ../../Appendix
rm *.o
Introduction to Case Study 4
Frenkel & Smit’s case studies are not set up in a particularly user friendly format. Before
we can learn about MD simulations by running some simulations, we need to learn about
the input and output files and how to run the code.
1. First we will simply run the code in the form that was provided. Move into the
appropriate directory:
cd ~/Frenkel_Smit/CaseStudy_4/Run
and then type ./run . This command executes a script that runs the main code and
moves around some of the output files. A line of output should appear on your screen that
contains information about how long it took to run the code, such as
1.6u 0.0s 0:03 54% 0+0k 0+1io 0pf+ 0w
The main useful information from all of this is the first number, how many total seconds
of CPU time were used (1.6 s in this case), and the third number, the total real time
elapsed while the code ran (3 s in this case, since the computer was also busy with other
2. The run you performed above has created two output files, out and lj.gr. The first
is a general purpose output file, the second is a data file containing the radial
distribution function, g(r), computed by the simulation. Note that if you run the code
again, the older version of lj.gr is replaced by a new one, while the new output
file is appended to the end of the older one.
3. For future reference, it is a good idea to make copies of the original input and output
files so you can always recreate the state of the original distribution. To do this:
cp out out.original
cp run run.original
4. Read through the supporting documentation that has been prepared to describe the
input script, run. This information will allow you to change the input conditions for
the code and run the MD simulation for various conditions. This documentation is
available at www.andrew.cmu.edu/course/06-640/CaseStudy_4.input.pdf
5. Read through the supporting documentation that has been prepared to describe the
output file, out. This information will allow you to interpret the results of your
simulations. This documentation is available at available at
Exercises for Case Study 4
The objective of this exercise is to familiarize you with some of the practical details of
running MD simulations. It would be a good idea to start by rereading the description of
the case study in the text, p. 98-100. You must hand in a report describing your results.
1. Using the same temperature, density, time step and system size as the original case
study, plot the radial distribution function, g(r), for the following run lengths: (a) t =
0.5 with no equilibration, (b) t = 2.5 with no equilibration, (c) t = 2.5 after 0.5 time
units of equilibration, (d) t = 10 after 0.5 time units of equilibration. Explain your
results. Specifically, list the differences between the four results, and give physical
reasons for these differences. Also briefly explain the physical significance of the
result that you believe is the most accurate. Note that the code is written so that the
velocities are rescaled every 20 time steps during the equilibration period.
2. Again using the same temperature, density, time step and system size as the original
case study, run the simulation for (a) t = 5.0 with no equilibration and (b) t = 5.0 after
equilibrating for 1.0 unit of time. In each case note the average temperature and
potential energy for the run and examine the run’s energy conservation. Describe and
explain your results.
3. The initial velocities in the simulation are assigned randomly using a random number
generator. The sequence of random numbers that is generated by the code is defined
by the random number seed (defined in the input file). If you run the code twice with
the same seed, you will get identical results. Run the code 5 times using the
conditions from part 2(b) and different random number seeds. Note the reported T, P,
and U from each run. Are the variations in these values from run to run consistent
with the error estimates provided by the code?
4. If an MD algorithm integrates Newton’s equations perfectly, the total energy of the
system should be exactly conserved. Run the code for the same conditions as part
2(b) using time steps of 0.001, 0.005, 0.01, 0.02, and 0.025. For each run, record the
change in total energy between time 1.2 and time 6.0 as listed in the output file. Note
that the total energy is not conserved between time 0.6 and 1.2 because the velocities
are rescaled every 20 time steps during the equilibration period. Explain the results
you observe.
5. For a system containing a single chemical species in a single phase, the equation of
state is an expression relating the pressure, temperature, and density. One well known
example is the ideal gas law, PV = nRT. Compute a partial equation of state for the
Lennard-Jones system by computing P at ρ = 0.5, 0.6, 0.7, 0.8, and 0.9 at T = 1.2.
You must choose an appropriate way to run the Case Study code to obtain these
results. You should report how you ran the code (i.e., run length, equilibration, time
step etc.) and why you believe these conditions give accurate results. For the purposes
of this exercise, any run resulting in an average temperature between 1.19 and 1.21
can be considered to have an acceptable temperature. Compare your results with the
predictions of the ideal gas law for methane, which has Lennard-Jones parameters ε =
148 K and σ = 0.373 nm. Is the ideal gas law a reasonable approximation under the
conditions you have simulated?
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF