It’s the end of week 4.

  • The crystallization of sphere_size2 and cube_size2 are on progress
  • Today I submit the task for all size (sphere and cube size 3,4,5,6) to the HPC for the NPT_T500 with 1 million timesteps.
  • It seems that if I submit task with 16 cpus will have longer queue time compared to the 8 cpus
  • Expected to got the result of NPT on Sat, 9 February and will run for crystallization.

It’s week 4. Not until last week the program for the constant temperature (NPT) was revised. Thank you to the professor, last week I was able to finish the NPT for sphere size 2 and started the Crystallization process. It’s still long way to go. Unfortunately, some of my configuration to incorporate more cpus was not working. Only config with 8 cpus is running. However it runs really slow. Should ask professor how to speed up, otherwise I will need a week to finish the whole simulation of 8 million timesteps. Meanwhile I run the cube size 2 so I can speed up the overall simulation. Three more weeks before the new semester starts.

Monday, January 14th, 2019: I met professor and set up the config together. There should be 4 simulations for OBSTACLE:2 (sphere).

Beginning my capstone, after summer break. Following are the setup conditions from Professor:

Bulk simulation with particle

Shape of the particle, OBSTACLE =1,2, 3 sphere and cube

rcut=1.112, rcut =2.5


Size of the particle with the same molecule (changes volume fraction)

Size ratio of molecule and particle,



Flow induced crystallization


1st simulation
i built start mode with 100,000 steplimit.
and equilibriate with, bodyforce=0.0
but not activate stress_bin
However when i showed the result to Prof. he said it’s good
previous sim, i put it in DUMP

2nd simulation
i rebuilt the start mode with only 5000 steplimit
turns out, out2 file has the same
then i activate stress_bin during equilibrate
I name it TASK2…rev in the PC folder
however keep the name TAG the same in winSCP

[TASK 4]
11 oct
After edited by Prof on the grid number, the START was run well
However i use steplimit 100000

12 oct
I edit the steplimit to 5000 in TASK4_start_rev
and successfuly built initial out2

Use it to Equilibrate
However stop at step mol 23??

seems that because of the StepGridZero 100,000 is too big given Step Equil is 10,000.

simulation 3, change stepgridzero to 5000–>stopped
simulation 4, stepgridzero 10000, StepEquil 20,000 –>stopped

sim 5, without grid activate, stress_bin=0, StepGridzero 100,000
coz I don’t care…LoL

step_webMoll stop at 91

13oct rerun but still stop at 91
change proc file

i use the out2 from 13oct (with step moll 91)
it works for g04, but i can’t open the grid file

change 8 but stopped

14 OKT
Running use more coarse grid

It worked for the g08, achieve stepwebmoll 100
grid file seems ok

I run equilibrium with coarse grid, just to check whether it will achieve moll 100, however i presume it wont be different coz the stress bin is not activated….and it achieve 100

then i use out2_equilibrate_coarse to run
next body force

TASK 2 – 18 Okt
at tutorial, checking progress with professor.
Able to do averaging and SI conversion
next to find steady state, find the velocity, Reynold number and the flow rate

TASK 2 – 19 OKT
Try to run the shorter timesteps
200,000 timesteps
stepGridZero i set to 20,000

I’ll put my logbook of the Computational USYD here. Just to keep it well documented and traceable. Artemis is a Supercomputer facility in USYD by which I do my simulation of Computational Nanotechnology.

Artemis login