QuantumATK Forum
QuantumATK => General Questions and Answers => Topic started by: fangyongxinxi on December 1, 2010, 03:57
-
Dear Sir,
Question1:
Import the initial calculation is helpful to calculate efficiently, I am familiar with ATK2008.10, but stange to ATK 2010.8,
I wrote a script, part of it is shown blow, I wonder is it right ? and what's the use of the " device_configuration.update() " ?
I also have a question on the manual page 109, as the attach shows. First, v = 0 V, the imported initial_state is just a device_configuration.(just contains the xyz_file, not scf process ). When the 0V calculation is done, but no "nlsave" to store the scf calculation ( I don't know it's right to say like this ), so when v = 0.1, the imported initial_state is the same xyz_file, no scf process. Am I right ? If I am wrong, how to understand the initial_state ?
Question2:
Could you give us some script on how to export the data , especially how to deal with these data. ( for example, how to draw a "v-scan I-V" picture by script. ), just like the previous topics on ATK 2008.10.
Qestion3:
a question about python. How to debug a python script efficiently ? I know how to use " try--except ", but that's not help me too much on how to debug a script. I like the debug in matlab, that's convenient.
Thank you.
#----------------------------------------
# Device Calculator
#----------------------------------------
v=0
calculator = DeviceLCAOCalculator(
electrode_calculators=
[left_electrode_calculator, right_electrode_calculator],
electrode_voltages=( v*Volt, 0.0*Volt)
)
device_configuration.setCalculator(calculator) ## time cost , this is the v=0 scf done.
device_configuration.update()
nlsave("analysis-%.1f.nc" % v, device_configuration)
opt_configuration = OptimizeGeometry(
device_configuration,
maximum_forces=0.05*eV/Ang,
constraints=[ 0, 1, 2, 3, 4, 5, 8, 9, 10, 11, 12,13],
trajectory_filename="analysis-%.1f.nc" % v )
nlsave("analysis-%.1f.nc" % v, opt_configuration)
forces = Forces(device_configuration)
nlsave("analysis-%.1f.nc" % v, forces)
print " ### 0 is done ####"
v= 0.2
calculator = DeviceLCAOCalculator(
electrode_calculators=
[left_electrode_calculator, right_electrode_calculator],
electrode_voltages=( v*Volt, 0.0*Volt)
)
device_configuration.setCalculator(calculator,
initial_state = device_configuration ) # I set the initial_stat here is right ? and after this scipt ,the v=0.2 scf is stored in device_configration, right?
device_configuration.update()
nlsave("analysis-%.1f.nc" % v,device_configuration)
opt_configuration = OptimizeGeometry(
device_configuration,
maximum_forces=0.05*eV/Ang,
constraints=[ 0, 1, 2, 3, 4, 5, 8, 9, 10, 11, 12,13],
trajectory_filename= "analysis-%.1f.nc" % v )
nlsave("analysis-%.1f.nc" % v, opt_configuration)
forces = Forces(device_configuration)
nlsave("analysis-%.1f.nc" % v,forces)
print "########0.2 is done########"
v=0.4....../v=0.6......
-
In my script, I understand it like this:
first I use
"device_configuration.setCalculator(calculator)"
to get the v=0 scf, and this is stored in device_configuration, which will be used by 0.2V (right ?)
then, I use the stored device_configuration to do the optimization and calculate the force.
v = 0.2 V, we set
" device_configuration.setCalculator(calculator, initial_state = device_configuration )"
to get the v=0 scf with the v=0.2 "calculator", at this same time v=0.2 scf is stored in device_configuration, which will be used by 0.4 V .
so that's the loop. right ?
-
Debugging in Python can be done e.g. in Eclipse.
It seems you have indeed understood the voltage loop correctly. Post further questions if there's anything else you wonder about.