heat flux calculation



Axel, since the beginning of my graduate studies(which is about 9 months) I
have learned everything about lammps own my own and the only hand that I can
get is this mail list which I appreciate, I mean there is no mechanic

you are confusing two things here. 1) general proficiency in
computational tools and 2) how lammps is implemented. while you may
not have anybody with a lot of experience in 2) around, there *have*
to be people with knowledge in 1). be it advisers, senior colleagues,
technicians. it is a lame excuse to claim 2) if your problem is
actually 1). if there be nobody with some proficiency in 1) around,
you need to quit immediately, because then any degree you would obtain
will not be worth the paper it is printed on.

helping me. I didn't care about what is going on in the lammps and how to
control it at the beginning as I was interested in modeling a problem and
the results from the engineering point of view; however, I started to
realize that I can't escape from the blues, as you said, in order to deal
with more complicated and innovative problems.


As a result, I can guarantee you that I have a desire to learn more and more
but the problem is my level of experience and I dont know where to start for
example for this compiling issue. I used ubuntu for a while by taking
suggestions like yours(being comfortable) but suprisingly I had several
problems like installing lammps(eventhough I did exactly what is written in
manual) and freezing of command prompt in to the middle of the run etc... So
I dont have plan to drop that desire as there will be even no consequence to
live with for the other option=) May be you can give me a couple of advice
about where to start for gaining this kind of skills like compiling or
related stuff. Thank you very much for your suggestions.

as i mentioned, this has noting to do with LAMMPS, but general
expertise in IT and particularly HPC technology. there are workshops,
courses and classes given in that all over the place. there are online
resources, self-training tutorials and more. you have to look in the
right places and not limit yourself to what pertains to LAMMPS.

you can find my own lecture materials on related topics here:
it has several pointers to other, more detailed resources and
materials from other people.
most notably are perhaps henry neeman's recurring lecture series
"supercomputing in plain english" and the ci-tutor project hosted at


Thank you very much for your attention Mr Kohlmeyer, I will study on what
you gave.

Dear Mr Kohlmeyer and Mr Sirk,

Using an arbitrary compute command for 0 kinetic energy term was a tricky
solution of my question as you remember but I suppose we missed this error:

"Compute heat/flux compute ID does not compute ke/atom" which is mentioned
in manual. When I used for example 'compute improper/local' command(there is
no improper in my data file so it is expected to be 0) and put this
calculation into 'compute heat/flux' instead of kinetic energy, I got that
error. I think 'compute heat/flux' command is forced to use compute ke/atom
command, do you agree with that or do you have additional comments?