Dear lammps-users,
I am trying to regenerate the reax/c and reax benchmarks ( http://lammps.sandia.gov/bench.html#accelerator ) for the new knights landing phi, and I keep getting the following error:
ERROR: Non-existent ReaxFF type (…/pair_reax_c.cpp:320)
I have a smaller system (using Zn and H2O) that runs fine, but I wanted to actually look at the scaling of reax/c on the Cori supercomputer. Any ideas?
-Micholas
Please double check your reaxff force field file. It should match your system composition.
Axel
The ffield.reax file does include the C, H, O & N atoms necessary for the simulation (in the proper order to match the composition in the benchmark). I just find it odd that one of the benchmarks listed on the lammps webpage doesn’t run
-Micholas
The ffield.reax file does include the C, H, O & N atoms necessary for the
simulation (in the proper order to match the composition in the benchmark).
I just find it odd that one of the benchmarks listed on the lammps webpage
doesn't run
these files were created 5 years ago. since LAMMPS is continuously updated
and improved, so every once in a while inputs need to be updated
accordingly. we do not have the manpower and time to check and review old
material. we rather depend on LAMMPS users being capable of comparing input
files with the current documentation and updating those old inputs
correspondingly.
in this case you simply have to replace 1 2 3 4 in the pair_coeff line with
C H O N
axel.