Print Page - geometry optimization for graphene

i am getting segmentation fault when trying to optimize geometry for a device ... is there a basic example of optimizing the geometry of a 2 probe graphene system?

If your are getting a segmentation fault, then is something wrong. If your system is confidential, please send it by email to support.

Please note you cannot use Huckel for optimizations.

I know .. I was able to try in VNL on windows, I got memory error .. then I tried on a very powerful linux machine with tons of memory ... segmentation fault ... I made the script first using VNL, following a tutorial ... my code looks like this:

Code

device_configuration = nlread(<filename>, object_id="cfg")[0]
calculator2 = my_DeviceLCAOCalculator(oList, bias, 0.0*Volt)
device_configuration.setCalculator(calculator2)
device_configuration.update()
device_configuration = OptimizeGeometry(
	device_configuration, max_forces=0.05*eV/Ang, max_steps=100, max_step_length=0.5*Ang, 
	trajectory_filename=None, disable_stress=True, optimizer_method=QuasiNewton())

where my_DeviceLCAOCalculator is:

Code

def my_DeviceLCAOCalculator(oList, leftBias, rightBias):

	#----------------------------------------
	# Basis Set
	#----------------------------------------
	basis_set = [
		LDABasis.Carbon_SingleZeta,
		LDABasis.Boron_SingleZeta,
		LDABasis.Hydrogen_SingleZeta,
		LDABasis.Nitrogen_SingleZeta,
		]


	# -------------------------------------------------------------
	# Calculator
	# -------------------------------------------------------------

	if oList["SCF"] == 1:

		#----------------------------------------
		# Electrode Calculators
		#----------------------------------------
		left_electrode_calculator = LCAOCalculator(
			basis_set=basis_set,
			numerical_accuracy_parameters=oList["numerical_accuracy_parameters"],
			# set these for SCF
			iteration_control_parameters=oList["iteration_control_parameters"],
		)


		right_electrode_calculator = LCAOCalculator(
			basis_set=basis_set,
			numerical_accuracy_parameters=oList["numerical_accuracy_parameters"],
			# set these for SCF
			iteration_control_parameters=oList["iteration_control_parameters"],
		)

		#----------------------------------------
		# Device Calculator
		#----------------------------------------
		calculator = DeviceLCAOCalculator(
			basis_set=basis_set,
			# set these for SCF
			numerical_accuracy_parameters=oList["numerical_accuracy_parameters"],
			iteration_control_parameters=oList["iteration_control_parameters"],
			electrode_calculators=
				[left_electrode_calculator, right_electrode_calculator],
			electrode_voltages=( leftBias, rightBias )
		)

	else:
		#----------------------------------------
		# Electrode Calculators
		#----------------------------------------
		left_electrode_calculator = LCAOCalculator(
			basis_set=basis_set,
			iteration_control_parameters=NonSelfconsistent,
			numerical_accuracy_parameters=oList["numerical_accuracy_parameters"],
		)

		right_electrode_calculator = LCAOCalculator(
			basis_set=basis_set,
			iteration_control_parameters=NonSelfconsistent,
			numerical_accuracy_parameters=oList["numerical_accuracy_parameters"],
		)

		#----------------------------------------
		# Device Calculator
		#----------------------------------------
		device_algorithm_parameters = DeviceAlgorithmParameters(
				initial_density_type=NeutralAtom(electrode_constraint_length=0.0*Ang),
		)
		calculator = DeviceLCAOCalculator(
			basis_set=basis_set,
			numerical_accuracy_parameters=oList["numerical_accuracy_parameters"],
			iteration_control_parameters=NonSelfconsistent,
			device_algorithm_parameters = device_algorithm_parameters,
			electrode_calculators= [left_electrode_calculator, right_electrode_calculator],
			electrode_voltages=( leftBias, rightBias)
		)

	return calculator

The only unusual setting I see is the electrode_constrain_length, try not setting it to zero. Where is the segfault? Insert some print statements to check when it actually occurs.

it is on the OptimizeGeometry call

actually hat portion is not being run because I have it using the self-consistent portion ..

This is too complex to troubleshoot without exact script.

well that is the very basic script, and i think the only thing needed is the nc file, but i will send by email thank you

QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: esp on August 30, 2012, 22:00