Florencia,
From seeing some of your recent emails with Zheyong, it looks like you may be
using GPUMD v2.4.1. If that is the case, you should not use the GPUMD web page
documentation as it is more for the development version (for now). For GPUMD
v2.4.1, you should use the documentation included with the package (‘gpumd.pdf’
in ‘doc’ folder).
Based on the documentation for v2.4.1, the command looks like the:
compute_shc sample_interval Nc steps_per_correlation group_A group_B
where:
sample_interval – sampling interval between two correlation steps
Nc – total correlation steps
steps_per_correlation – number of steps used to calculate one correlation
function
group_A – label given to group of atoms
group_B – label given to a different group of atoms
Try running with these parameters and let me know how it goes.
Regards,
Alex
From: gpumd-bounce@xxxxxxxxxxxxx <gpumd-bounce@xxxxxxxxxxxxx> On Behalf Of
Florencia Carusela
Sent: Tuesday, April 28, 2020 11:21 AM
To: gpumd@xxxxxxxxxxxxx
Subject: [gpumd] about parameters in compute_shc
Hi Zheyong
I am here again (sorry). I have problems with the input parameters for
compute_shc .
I am using compute_shc with the file xyz.in <http://xyz.in> attached (just
only with 100 lines, last column goes from 0 to 7)
I tried four different options for this command, all with error messages as
follows.
a) compute_shc 10 100 1
Compute SHC.
ERROR: compute_shc should have 5 parameters.
b) compute_shc 4 10 100 1
Compute SHC.
ERROR: compute_shc should have 5 parameters.
c) compute_shc 0 4 10 100 1
Floating point exception (core dumped)
d) compute_shc 1 4 10 100 1
No error, but all the output is zero because there is no "1" grouping_method
I took the options from the "The_compute_shc_keyword" web page, quite different
from your example 4. Am I doing something wrong?
Thanks for your help.
Florencia
