QMD to simulate time-resolved IR

From NWChem

Viewed 381 times, With a total of 4 Posts
Jump to: navigation, search

Forum Regular
Threads 59
Posts 283
I increased the basis set and found that only the last wavenumber might be recognized as a large discrepancy compared with the result obtained with the original input but whether word spherical was used or not caused an negligible difference. When the basis set was increased further, word spherrical did make significant sense, perhaps also the last one could be called mismatch. The next increasement of the baisis set, i.e., the employment of aug-cc-pvtz, showed that word spherical made no sense, but only the one without it might be considered no significant changes for all wavenumbers compared with the previous one. Here, I arbitrarily choose around 100 cm-1 as a difference which cannot be considered not to be accepted. The differences of one of the wavenumbers of all basis sets are smaller than that value throughout.
I understand some kind of correction in both GAMESS and NWCHEM, which can provide ameliorative IR intensities or vibrational modes, perhaps important for exact potential energy surface calculations, not BSSE may account for this.
Edited On 12:07:20 AM PDT - Wed, Apr 3rd 2019 by Xiongyan21

Forum Regular
Threads 59
Posts 283
I just know it is remarkable that NWCHEM can use QMD to simulate time-resolved IR spectrum and the element-wise breakdown of the vibrational DOS at DFT level. Does just the exportation of qmd_tools in contrib or additional path settings during the compilation make live plots through Fourier transforms and autocorrelation functions?
Now I want to try the compilation of NWCHEM6.8.1, perhaps the first exascale, with the tools included.

By the way, the example in the manual illustrating this was tested by me, but the original step_nucl 20000 could not be recognized, and the run was successful after 200 had been chosen arbitrarily.


Thanks a lot !

Very Best Regards!
Edited On 7:23:32 PM PDT - Tue, Apr 16th 2019 by Xiongyan21

Forum Regular
Threads 59
Posts 283
On Ubuntu18.04, the original step_nucl 20000 could be recognized, although I cut off the calculation after 110.059237 femtoseconds elapsed.
I will try it further.
Edited On 7:18:25 PM PDT - Sun, Apr 14th 2019 by Xiongyan21

Forum Regular
Threads 59
Posts 283
On MAC, the original step_nucl 20000 could be recognized. Perhaps previously I used a wrong input. Although something, perhaps the keyboard is hit, causes an error in the terminal, i.e., 0:ssread: error reading from unit:Received an Error in Communication
application called MPI_Abort(comm=0x84000004, 67) - process 0, the calculation actually continues and I cut off it after 276.720367 femtoseconds elapsed.
I will try it further.
Edited On 7:16:42 PM PDT - Sun, Apr 14th 2019 by Xiongyan21

Forum Regular
Threads 59
Posts 283
This work can be successfully finished on Ubuntu 18.04 using NWCHEM6.8 with four threads
QMD Run Information
-------------------
Time elapsed (fs) : ...
...
Dipole (a.u.)  : 20000 ...
and on macOS Mojave it also can finish smoothly using three cores, although slower, perhaps with the identical results because I have only checked the last several steps of the very large number of the data produced
QMD Run Information
           -------------------
Time elapsed (fs) : ...
...
Dipole (a.u.)  : 20000 ...

It seems this example does not exist in the QA tests.
Edited On 4:08:32 AM PST - Fri, Nov 15th 2019 by Xiongyan21


Forum >> NWChem's corner >> Running NWChem



Who's here now Members 0 Guests 1 Bots/Crawler 0


AWC's: 2.5.10 MediaWiki - Stand Alone Forum Extension
Forum theme style by: AWC