MCP 1 Chapter 3 Molecular Dynamics
Questions about the lecture 'From Molecular to Continuum Physics 1' of the RWTH Aachen Chapter 3 Molecular Dynamics
Questions about the lecture 'From Molecular to Continuum Physics 1' of the RWTH Aachen Chapter 3 Molecular Dynamics
-
- 1 / 126
-
Lernkarten
What are the options of VMD visualization? [4]
1. Change draw method
2. Remove axis
3. Change visualization mode
4. Add period replicas
What are the initial structures in VMD? [3]
1. Cubic
2. Body centered cubic (bcc) // Cube with central point
3. Face centered cubic (fcc) // Cube with facet central points
What is the usage principle of GROMACS?
Single binary run by commands
What has to be done if we want to use GROMACS with the RWTH compute cluster? [3]
1. module load gromacs/5.1.2dp
2. module load vmd/1.9.2
3. Slow if connection bad
How do we pre-process in GROMACS? [5]
gmx grompp
-f file_name.mdp
-c file_name.pdb
-p file_name.top
-o file_name.tpr
How do we process in GROMACS? [4]
gmx mdrun
-v
-s file_name.tpr
-deffnm dfn // default filename
What are the output files in GROMACS? [7]
1. dfnout.mdp
2. dfn.log
3. dfn.trr
4. dfn.xtc
5. dfn.edr
6. dfn.gro
7. dfn.cpt
What are the characteristics of dfnout.mdp in GROMACS? [2]
1. All simulation parameters
2. By grompp
What are the characteristics of dfn.log in GROMACS? [2]
1. Log of simulation
2. Every nstlog steps
What is the characteristic of dfn.trr in GROMACS?
Complete trajectory information
What are the characteristics of dfn.xtc in GROMACS? [2]
1. Position information in reduced precision
2. Every nstxout-compressed
What are the characteristics of dfn.edr in GROMACS? [2]
1. Energy data
2. Every nstenergy step
What are the characteristics of dfn.gro in GROMACS? [2]
1. Last snapshot of trajectory in gro (GROMACS) format
2. At the end of simulation
What are the characteristics of dfn.cpt in GROMACS? [2]
1. Checkpoint used to restart a simulation
2. Every 15 minutes and at the end of the simulation
How can we scale in GROMACS?
With MPI or OPENMP parallel version of GROMACS using: mpirun -np c gmx_mpi before process command with c\in{1,2,4}
How can we visualize trajectories in VMD? [3]
1. Open .pdb or final .gro file
2. Load data into molecule with .xtc file
3. One frame for description and others for trajectory
How to set the integrator in a simulation parameter file in GROMACS? [3]
1. md is leap-frog
2. md-vv is velocity Verlet
3. md-vv-avek is velocity Verlet with K as average of two half step kinetic energies
How to display dfn.edr in GROMACS? [2]
1. Extract information with gmx energy -f dfn.edr -o Etot.xvg
2. Display extracted information with grace with xmgrace Etot.xvg
How to write a MD program in Fortran? [2]
1. Initialization with t=0
2. MD loop while t<t_max
How to specify the MD loop in Fortran? [4]
1. Call force out of environment // 60%-80% of computational time
2. Call integrate with equations of motion
3. Update t+=dt
4. Call sample averages
What are the challenges of time integration algorithms? [3]
1. Forces change whenever particle position changes
2. Particle motion cause many-body problem
3. No analytical solution
What are the possible errors of time integration algorithms? [2]
1. Truncation and 2. round-off
What is the definition of the truncation error of time integration algorithms?
Intrinsic to the algorithm due to Taylor expansion truncation
What is the definition of the round-off error of time integration algorithms?
Due to particular implementation // Number of digits
How can we react on possible errors of time integration algorithms?
Both errors can be reduced by decreasing dt
What are the effects of decreasing dt for the possible errors of time integration algorithms? [2]
1. For large dt truncation dominates
2. For small dt round-off dominates
What are the characteristic of finite difference methods for time integration algorithms? [2]
1. Discrete time step dt
2. Total force on each particle is constant during dt
What are possible finite difference methods for time integration algorithms? [4]
1. Euler 2. Verlet 3. Leap-frog and 4. velocity Verlet
What are the formulas of the Euler algorithm for time integration algorithms? [2]
1. r_i(t+dt) = r_i(t) + v_i(t)*dt + ½a_i(t)*dt²
2. v_i(t+dt) = v_i(t) + a_i(t)*dt
What is the characteristic of the Euler algorithm for time integration algorithms?
Not stable enough due to missing energy conservation
What are the formulas of the Verlet algorithm for time integration algorithms?
r_i(t+dt) = 2r_i(t) – r_i(t-dt) + a_i(t)*dt² + O(dt⁴)
What is the characteristic of the Verlet algorithm for time integration algorithms? [5]
1. Sum forward and backward r_i(t) of two third-order Taylor
2. No use of velocity for new position
3. Modest requirements of storage
4. Adding small term O(dt²) causes loss of precision
5. v_i(t) is only accurate to the second order in dt
What are the formulas of the Leap-frog algorithm for time integration algorithms? [3]
1. r_i(t+dt) = r_i(t) + v_i(t+dt/2)*dt + O(dt⁴)
2. v_i(t+dt/2) = v_i(t-dt/2) + (f_i(t-dt/2) / m_i)*dt
3. v_i(t) = (v_i(t+dt/2) – v_i(t-dt/2))/2
What is the characteristic of the Leap-frog algorithm for time integration algorithms? [3]
1. Variant of Verlet creating identical trajectories
2. Positions and velocities at interleaved time points
3. Positions and explicitly included velocities not synchronized
What are the formulas of the velocity Verlet algorithm for time integration algorithms? [2]
1. r_i(t+dt) = r_i(t) + v_i(t)*dt + f_i(t)/2m *dt²
2. v_i(t+dt) = v_i(t) + (f_i(t+dt)+f_i(t) / 2m_i)*dt
What is the characteristic of the velocity Verle algorithm for time integration algorithms? [3]
1. Similar to leap-frog but positions, velocities and forces synchronized
2. Best estimate of velocities among Verlet-like algorithms
3. r_new → v_new → f_new
What are the requirements for time integration algorithms? [4]
1. Speed is not essential
2. Accuracy for large time steps
3. Energy conservation // Usually for short OR long-time
4. Time reversible
How many degrees of freedom does a physical system model has?
3*N
What are the degrees of freedom used for?
For total information about configuration in a specific time
How many dimensions does a the phase space has?
6*N
