Stochastic LTMP2 - Revision history

Huebsch at 10:27, 19 July 2022

2022-07-19T10:27:25Z

← Older revision		Revision as of 10:27, 19 July 2022
Line 72:		Line 72:

	----		----
	[[Category:Many-~~Body Perturbation Theory~~]][[Category:MP2]]		[[Category:Many-body perturbation theory]][[Category:MP2]]

Kaltakm at 14:22, 24 February 2020

2020-02-24T14:22:30Z

← Older revision		Revision as of 14:22, 24 February 2020
Line 17:		Line 17:
	{{TAGBL\|PRECFOCK}} = Fast		{{TAGBL\|PRECFOCK}} = Fast
	Make sure that VASP reads the WAVECAR file from the [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]] step. The setting for [[PRECFOCK\|PRECFOCK]] is strongly recommended, since the code heavily relies on real space grid FFTs.		Make sure that VASP reads the WAVECAR file from the [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]] step. The setting for [[PRECFOCK\|PRECFOCK]] is strongly recommended, since the code heavily relies on real space grid FFTs.


	==== NOMEGA flag ====		==== NOMEGA flag ====
Line 28:		Line 27:

	Usually it is sufficient to set NOMEGA to 6. For materials with a small bandgap it is worth checking if the MP2 energy changes with increasing NOMEGA (e.g. 8 or 10). Note, that the MP2 energy diverges with 1/bandgap, independent of NOMEGA.		Usually it is sufficient to set NOMEGA to 6. For materials with a small bandgap it is worth checking if the MP2 energy changes with increasing NOMEGA (e.g. 8 or 10). Note, that the MP2 energy diverges with 1/bandgap, independent of NOMEGA.


	==== ESTOP flag ====		==== ESTOP flag ====

Kaltakm at 14:21, 24 February 2020

2020-02-24T14:21:09Z

← Older revision		Revision as of 14:21, 24 February 2020
Line 67:		Line 67:

	It is also possible to increase the memory per MPI rank using shared memory with OMP. This is a viable option if your available memory per core is too small, decreasing '''KPAR''' does not help or you don't want to set too small '''KPAR''' values. However, in general, it is recommended to solve memory issues with the '''KPAR''' flag first.		It is also possible to increase the memory per MPI rank using shared memory with OMP. This is a viable option if your available memory per core is too small, decreasing '''KPAR''' does not help or you don't want to set too small '''KPAR''' values. However, in general, it is recommended to solve memory issues with the '''KPAR''' flag first.



	~~=== KPAR flag ===~~


	== References ==		== References ==

Kaltakm: /* The INCAR file */

2020-02-24T14:18:19Z

The INCAR file

← Older revision		Revision as of 14:18, 24 February 2020
Line 7:		Line 7:

	The LTMP2 calculation can simply be performed using the following {{TAG\|INCAR}} file		The LTMP2 calculation can simply be performed using the following {{TAG\|INCAR}} file
	ALGO = ACFDTRK		{{TAGBL\|ALGO}} = ACFDTRK
	LSMP2LT = .TRUE.		{{TAGBL\|LSMP2LT}} = .TRUE.
	ESTOP = # accuracy of energy per tau-point		{{TAGBL\|ESTOP}} = # accuracy of energy per tau-point
	NSTORB = # number of stochastic orbitals per cycle		{{TAGBL\|NSTORB}} = # number of stochastic orbitals per cycle
	NOMEGA = # number of tau-points (see below)		{{TAGBL\|NOMEGA}} = # number of tau-points (see below)
	NBANDS = # same valule as in the Hartree-Fock unoccupieds step ( = number of plane-waves)		{{TAGBL\|NBANDS}} = # same valule as in the Hartree-Fock unoccupieds step ( = number of plane-waves)
	ENCUT = # same value as in the Hartree-Fock step		{{TAGBL\|ENCUT}} = # same value as in the Hartree-Fock step
	LORBITALREAL = .TRUE.		{{TAGBL\|LORBITALREAL}} = .TRUE.
	PRECFOCK = Fast		{{TAGBL\|PRECFOCK}} = Fast
	Make sure that VASP reads the WAVECAR file from the [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]] step. The setting for [[PRECFOCK\|PRECFOCK]] is strongly recommended, since the code heavily relies on real space grid FFTs.		Make sure that VASP reads the WAVECAR file from the [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]] step. The setting for [[PRECFOCK\|PRECFOCK]] is strongly recommended, since the code heavily relies on real space grid FFTs.

Karsai at 13:34, 7 June 2019

2019-06-07T13:34:14Z

← Older revision		Revision as of 13:34, 7 June 2019
Line 79:		Line 79:

	----		----
	[[Category:Many-Body Perturbation Theory]][[Category:MP2~~]][[Category:Tutorials~~]]		[[Category:Many-Body Perturbation Theory]][[Category:MP2]]

Karsai at 13:29, 7 June 2019

2019-06-07T13:29:23Z

← Older revision		Revision as of 13:29, 7 June 2019
Line 1:		Line 1:
	{{Template:MP2}}		{{Template:MP2 - Tutorial}}
	On this page, we explain how to perform a calculation using the '''stochastic LTMP2'''<ref name="schaefer2018"/> algorithm. Make sure you successfully completed the preparation steps [[MP2 ground state calculation#Preparation:_the_Hartree-Fock_ground_state\|Hartree-Fock ground state]] and [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]].		On this page, we explain how to perform a calculation using the '''stochastic LTMP2'''<ref name="schaefer2018"/> algorithm. Make sure you successfully completed the preparation steps [[MP2 ground state calculation#Preparation:_the_Hartree-Fock_ground_state\|Hartree-Fock ground state]] and [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]].

Tobias at 18:15, 26 March 2019

2019-03-26T18:15:07Z

← Older revision		Revision as of 18:15, 26 March 2019
Line 77:		Line 77:
	<ref name="schaefer2018">[https://doi.org/10.1063/1.5016100 T. Schäfer, B. Ramberger, and G. Kresse, J. Chem. Phys. 148, 064103 (2018).]</ref>		<ref name="schaefer2018">[https://doi.org/10.1063/1.5016100 T. Schäfer, B. Ramberger, and G. Kresse, J. Chem. Phys. 148, 064103 (2018).]</ref>
	</references>		</references>

			----
			[[Category:Many-Body Perturbation Theory]][[Category:MP2]][[Category:Tutorials]]

Tobias at 16:40, 28 January 2019

2019-01-28T16:40:55Z

← Older revision		Revision as of 16:40, 28 January 2019
Line 2:		Line 2:
	On this page, we explain how to perform a calculation using the '''stochastic LTMP2'''<ref name="schaefer2018"/> algorithm. Make sure you successfully completed the preparation steps [[MP2 ground state calculation#Preparation:_the_Hartree-Fock_ground_state\|Hartree-Fock ground state]] and [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]].		On this page, we explain how to perform a calculation using the '''stochastic LTMP2'''<ref name="schaefer2018"/> algorithm. Make sure you successfully completed the preparation steps [[MP2 ground state calculation#Preparation:_the_Hartree-Fock_ground_state\|Hartree-Fock ground state]] and [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]].

	NOTE: If you use this algorithm, please cite reference <ref name="schaefer2018"/> in your publication in addition to the standard VASP reference.		'''NOTE:''' ''If you use this algorithm, please cite reference <ref name="schaefer2018"/> in your publication in addition to the standard VASP reference.''

	== The INCAR file ==		== The INCAR file ==

Tobias at 10:15, 26 January 2019

2019-01-26T10:15:53Z

← Older revision		Revision as of 10:15, 26 January 2019
Line 32:		Line 32:
	==== ESTOP flag ====		==== ESTOP flag ====

	This flag defines the energy accuracy in units of eV for each individual tau-point of the two individual MP2 energy contributions (direct MP2 term + exchange MP2 term). Since the statistical errors of each contribution is independent, the standard deviation of the MP2 energy can be estimated as		This flag defines the stop condition for the stochastic algorithm. It defines the energy accuracy in units of eV for each individual tau-point of the two individual MP2 energy contributions (direct MP2 term + exchange MP2 term). Since the statistical errors of each contribution is independent, the standard deviation of the MP2 energy can be estimated as

	<math>		<math>

Tobias at 17:07, 25 January 2019

2019-01-25T17:07:11Z

← Older revision		Revision as of 17:07, 25 January 2019
Line 17:		Line 17:
	PRECFOCK = Fast		PRECFOCK = Fast
	Make sure that VASP reads the WAVECAR file from the [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]] step. The setting for [[PRECFOCK\|PRECFOCK]] is strongly recommended, since the code heavily relies on real space grid FFTs.		Make sure that VASP reads the WAVECAR file from the [[MP2 ground state calculation#Calculating_the_unoccupied_Hartree-Fock_orbitals\|Hartree-Fock virtuals]] step. The setting for [[PRECFOCK\|PRECFOCK]] is strongly recommended, since the code heavily relies on real space grid FFTs.


			==== NOMEGA flag ====

			The number of <math>\tau</math>-points is controlled by the NOMEGA flag. This is necessary to calculate the Laplace transformed energy denominator (see Ref <ref name="schaefer2018"/> for details),

			<math>
			\frac{1}{\varepsilon_i + \varepsilon_j - \varepsilon_a -\varepsilon_b} = - \int_0^\infty \textrm e^{-(\varepsilon_i + \varepsilon_j - \varepsilon_a -\varepsilon_b)\tau} \; \textrm d \tau \;.
			</math>

			Usually it is sufficient to set NOMEGA to 6. For materials with a small bandgap it is worth checking if the MP2 energy changes with increasing NOMEGA (e.g. 8 or 10). Note, that the MP2 energy diverges with 1/bandgap, independent of NOMEGA.


	==== ESTOP flag ====		==== ESTOP flag ====
Line 28:		Line 40:
	According to our experience, the error of the resulting MP2 energy can then be safely estimated by <math>\pm 2 \sigma</math>.		According to our experience, the error of the resulting MP2 energy can then be safely estimated by <math>\pm 2 \sigma</math>.

	So if you require an MP2 energy with a maximum error of <math>\Delta</math>, you should set		Thus, if you require an MP2 energy with a maximum error of <math>\Delta</math>, you should set

	<math>		<math>
Line 34:		Line 46:
	</math>		</math>

			==== NSTORB flag ====

	~~==== NOMEGA~~ flag ~~====~~		This flag defines the number of stochastic orbitals per cycle, i.e. the number of stochastic orbitals that define one stochastic sample. If the sample is not large enough, the calculations is repeated until the accuracy, defined by ESTOP, is reached.

	~~The number~~ of ~~<math>\tau</math>-points is controlled by the NOMEGA flag. This is necessary~~ to ~~calculate the Laplace transformed energy denominator (see Ref <ref name="schaefer2018"/> for details),~~		As a rule of thumb we recommend to set

	<math>		<math>
	\~~frac~~{~~1}{\varepsilon_i + \varepsilon_j - \varepsilon_a -\varepsilon_b~~} = - \~~int_0^\infty~~ \~~textrm e^~~{~~-(\varepsilon_i + \varepsilon_j - \varepsilon_a -\varepsilon_b)\tau~~} ~~\; \textrm d \tau~~ \;.		\texttt{NSTORB} = \sqrt{\texttt{NBANDS}} \;.
	</math>		</math>

	Usually it is sufficient to set NOMEGA to 6. For materials with a small bandgap it is worth checking if the MP2 energy changes with increasing NOMEGA (e.g. 8 or 10). Note, that the MP2 energy diverges with 1/bandgap, independent of NOMEGA.

	== Parallelization ==		== Parallelization ==