[Wien] BoltzTraP2

[Gavin Abo]

Sorry, it is not clear to me what you are asking.

As far as I know, BoltzTrap and the new BoltzTrap2 are 'serial' programs.

By "implement parallel BoltzTrap", are you asking about how to use the 
WIEN2k parallel output from an scf calculation (run[sp]_lapw -p) with 
BoltzTrap?

There is currently a gather_energy.pl script [ 
https://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/msg13418.html 
] for combining the parallel energy files of WIEN2k into a serial one 
that can be used with BoltzTrap.

Note that on the website at

http://blog.levilentz.com/boltztrap-tutorial-for-quantum-espresso/

there is:

/"finally how to run Boltztrap in parallel mode?"-Muthu
/

/I am not sure you can run this is parallel. Unless you have a huge 
number of tangled bands, you really shouldn’t need to. Further, if you 
do have such a system, only consider the mobility of the bands near the 
fermi energy as those will be the only ones participating in electronic 
mobility (in general).-Levi/

Or are asking how to program the BoltzTrap source code to make parts of 
it parallel?

If your trying to rewrite parts of the code to parallelize it, I think 
Intel has some documentation for doing that with their ifort or icc 
compilers. For example, see:

https://software.intel.com/en-us/mpi-developer-guide-linux-compiling-an-mpi-program

Some universities might have courses or hold seminars on programing for 
code parallelization.

There should also be coding school companies that offer online courses.  
If you happen to work for a software company, your company might be 
paying one of those online coding schools to give you access to those 
courses.  Maybe, even some universities do that too.  Your library or 
department could likely tell you what is available to you.  As an 
example, the "High-performance Computing in C++" course website for 
Pluralsight mentions it covers MPI and parallelization topics:

https://www.pluralsight.com/courses/cpp-high-performance-computing

However, I haven't taken the above course, so I cannot vouch for whether 
it would be helpful or not for parallelizing the BolzTrap source code.  
There may be a more appropriate course or tutorial that you could find 
yourself somewhere else on the web.

On 10/29/2017 3:36 PM, Kamil Klier wrote:
> I'd like to join this inquiry, with specifics on how to implement 
> parallel BoltzTrap and whether a link to graphic (ps ?) could be mads 
> available.
>
> With thanks,
>
> Kamil
>
> On Sat, Oct 28, 2017 at 10:15 PM, Gavin Abo <gsabo at crimson.ua.edu 
> <mailto:gsabo at crimson.ua.edu>> wrote:
>
>     All,
>
>     Just curious, has anyone heard if the new BoltzTraP2 package will
>     remain a private code or if it is just not ready yet for public
>     release?
>
>     The BoltzTraP website
>
>     https://www.imc.tuwien.ac.at//forschungsbereich_theoretische_chemie/forschungsgruppen/prof_dr_gkh_madsen_theoretical_materials_chemistry/boltztrap/
>     <https://www.imc.tuwien.ac.at//forschungsbereich_theoretische_chemie/forschungsgruppen/prof_dr_gkh_madsen_theoretical_materials_chemistry/boltztrap/>
>
>     and the blank workshop conclusion slide at
>
>     http://susi.theochem.tuwien.ac.at/events/ws2017/
>     <http://susi.theochem.tuwien.ac.at/events/ws2017/>
>
>     don't seem to provide any hints about that.
>
>     Thanks in advance,
>
>     Gavin
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20171029/b0722a0a/attachment.html>