pw_ecut in JTH PAW dataset
Posted: Fri Dec 21, 2018 9:16 am
Hi All,
I use a JTH PAW dataset from my calculation and the file lists the following parameter: <pw_ecut low="12.00" medium="12.00" high="12.00"/>. If I understand correctly this pw_ecut is the recommended value to bet set as "ecut" in the input file. Am I right? If so, how strict is this recommendation, I mean, can we set up the ecut much smaller than the recommended one, like half of the value? The reason I ask this is that my "optimization" calculation is just to expensive to use the recommended one. I first just want to do a rough "optimization" calculation and, perhaps, after the system reach the structure that's closer to the global minimum, I can refine "more cheaply" the calculation by increasing the ecut to get to the global minimum. Any suggestions on this matter would be appreciated.
I use a JTH PAW dataset from my calculation and the file lists the following parameter: <pw_ecut low="12.00" medium="12.00" high="12.00"/>. If I understand correctly this pw_ecut is the recommended value to bet set as "ecut" in the input file. Am I right? If so, how strict is this recommendation, I mean, can we set up the ecut much smaller than the recommended one, like half of the value? The reason I ask this is that my "optimization" calculation is just to expensive to use the recommended one. I first just want to do a rough "optimization" calculation and, perhaps, after the system reach the structure that's closer to the global minimum, I can refine "more cheaply" the calculation by increasing the ecut to get to the global minimum. Any suggestions on this matter would be appreciated.