GLOBAL

Section with general information regarding which kind of simulation to perform an parameters for the whole PROGRAM [Edit on GitHub]

Keywords

Keyword descriptions

ALLTOALL_SGL: logical = F

Lone keyword: T

Usage: ALLTOALL_SGL YES

All-to-all communication (FFT) should use single precision [Edit on GitHub]

BLACS_GRID: enum = SQUARE

Usage: BLACS_GRID SQUARE

Valid values:

  • SQUARE Distribution by matrix blocks

  • ROW Distribution by matrix rows

  • COLUMN Distribution by matrix columns

how to distribute the processors on the 2d grid needed by BLACS (and thus SCALAPACK) [Edit on GitHub]

BLACS_REPEATABLE: logical = F

Lone keyword: T

Usage: BLACS_REPEATABLE

Use a topology for BLACS collectives that is guaranteed to be repeatable on homogeneous architectures [Edit on GitHub]

CALLGRAPH: enum = NONE

Lone keyword: MASTER

Usage: CALLGRAPH {NONE|MASTER|ALL}

Valid values:

  • NONE No callgraph gets written

  • MASTER Only the master process writes his callgraph

  • ALL All processes write their callgraph (into a separate files).

At the end of the run write a callgraph to file, which contains detailed timing informations. This callgraph can be viewed e.g. with the open-source program kcachegrind. [Edit on GitHub]

CALLGRAPH_FILE_NAME: string

Usage: CALLGRAPH_FILE_NAME {filename}

Name of the callgraph file, which is written at the end of the run. If not specified the project name will be used as filename. [Edit on GitHub]

DLAF_NEIGVEC_MIN: integer = 1024

Usage: DLAF_NEIGVEC_MIN 512

Minimum number of eigenvectors for the use of the eigensolver from the DLA-Future library. The eigensolver from the ScaLAPACK library is used as fallback for all smaller cases [Edit on GitHub]

ECHO_ALL_HOSTS: logical = F

Lone keyword: T

Usage: ECHO_ALL_HOSTS NO

Echo a list of hostname and pid for all MPI processes. [Edit on GitHub]

ECHO_INPUT: logical = F

Lone keyword: T

Usage: ECHO_INPUT NO

If the input should be echoed to the output with all the defaults made explicit [Edit on GitHub]

ELPA_KERNEL: enum = AUTO

Valid values:

  • AUTO Automatically selected kernel

  • GENERIC Generic kernel

  • GENERIC_SIMPLE Simplified generic kernel

  • BGP Kernel optimized for IBM BGP

  • BGQ Kernel optimized for IBM BGQ

  • SSE Kernel optimized for x86_64/SSE

  • SSE_BLOCK2 Kernel optimized for x86_64/SSE (block=2)

  • SSE_BLOCK4 Kernel optimized for x86_64/SSE (block=4)

  • SSE_BLOCK6 Kernel optimized for x86_64/SSE (block=6)

  • AVX_BLOCK2 Kernel optimized for Intel AVX (block=2)

  • AVX_BLOCK4 Kernel optimized for Intel AVX (block=4)

  • AVX_BLOCK6 Kernel optimized for Intel AVX (block=6)

  • AVX2_BLOCK2 Kernel optimized for Intel AVX2 (block=2)

  • AVX2_BLOCK4 Kernel optimized for Intel AVX2 (block=4)

  • AVX2_BLOCK6 Kernel optimized for Intel AVX2 (block=6)

  • AVX512_BLOCK2 Kernel optimized for Intel AVX-512 (block=2)

  • AVX512_BLOCK4 Kernel optimized for Intel AVX-512 (block=4)

  • AVX512_BLOCK6 Kernel optimized for Intel AVX-512 (block=6)

  • NVIDIA_GPU Kernel targeting Nvidia GPUs

  • AMD_GPU Kernel targeting AMD GPUs

  • INTEL_GPU Kernel targeting Intel GPUs

Specifies the kernel to be used when ELPA is in use [Edit on GitHub]

ELPA_NEIGVEC_MIN: integer = 16

Usage: ELPA_NEIGVEC_MIN 32

Minimum number of eigenvectors for the use of the eigensolver from the ELPA library. The eigensolver from the ScaLAPACK library is used as fallback for all smaller cases [Edit on GitHub]

ELPA_QR: logical = F

Lone keyword: T

Usage: ELPA_QR

For ELPA, enable a blocked QR step when reducing the input matrix to banded form in preparation for the actual diagonalization step. See implementation paper for more details. Requires ELPA version 201505 or newer, automatically deactivated otherwise. If true, QR is activated only when the the size of the diagonalized matrix is suitable. Print key PRINT_ELPA is useful in determining which matrices are suitable for QR. Might accelerate the diagonalization of suitable matrices. [Edit on GitHub]

ELPA_QR_UNSAFE: logical = F

Lone keyword: T

Usage: ELPA_QR

For ELPA, disable block size limitations when used together with ELPA_QR. Keyword relevant only with ELPA versions 201605 or newer. Use keyword with caution, as it might result in wrong eigenvalues with some matrix orders/block sizes when the number of MPI processes is varied. If the print key PRINT_ELPA is active the validity of the eigenvalues is checked against values calculated without ELPA QR. [Edit on GitHub]

ENABLE_MPI_IO: logical = T

Lone keyword: T

Usage: ENABLE_MPI_IO FALSE

Enable MPI parallelization for all supported I/O routines Currently, only cube file writer/reader routines use MPI I/O. Disabling this flag might speed up calculations dominated by I/O. [Edit on GitHub]

EPS_CHECK_DIAG: real = -1.00000000E+000

Usage: EPS_CHECK_DIAG 1.0E-14

Check that the orthonormality of the eigenvectors after a diagonalization fulfills the specified numerical accuracy. A negative threshold value disables the check. [Edit on GitHub]

EXTENDED_FFT_LENGTHS: logical = F

Lone keyword: T

Usage: EXTENDED_FFT_LENGTHS

Use fft library specific values for the allows number of points in FFTs. The default is to use the internal FFT lengths. For external fft libraries this may create an error at the external library level, because the length provided by cp2k is not supported by the external library. In this case switch on this keyword to obtain, with certain fft libraries, lengths matching the external fft library lengths, or larger allowed grids, or grids that more precisely match a given cutoff. IMPORTANT NOTE: in this case, the actual grids used in CP2K depends on the FFT library. A change of FFT library must therefore be considered equivalent to a change of basis, which implies a change of total energy. [Edit on GitHub]

FFTW_PLAN_TYPE: enum = ESTIMATE

Usage: FFTW_PLAN_TYPE PATIENT

Valid values:

  • ESTIMATE Quick estimate, no runtime measurements.

  • MEASURE Quick measurement, somewhat faster FFTs.

  • PATIENT Measurements trying a wider range of possibilities.

  • EXHAUSTIVE Measurements trying all possibilities - use with caution.

References: Frigo2005

FFTW can have improved performance if it is allowed to plan with explicit measurements which strategy is best for a given FFT. While a plan based on measurements is generally faster, differences in machine load will lead to different plans for the same input file, and thus numerics for the FFTs will be slightly different from run to run. PATIENT planning is recommended for long ab initio MD runs. [Edit on GitHub]

FFTW_WISDOM_FILE_NAME: string = /etc/fftw/wisdom

Usage: FFTW_WISDOM_FILE_NAME wisdom.dat

The name of the file that contains wisdom (pre-planned FFTs) for use with FFTW3. Using wisdom can significantly speed up the FFTs (see the FFTW homepage for details). Note that wisdom is not transferable between different computer (architectures). Wisdom can be generated using the fftw-wisdom tool that is part of the fftw installation. cp2k/tools/cp2k-wisdom is a script that contains some additional info, and can help to generate a useful default for /etc/fftw/wisdom or particular values for a given simulation. [Edit on GitHub]

FFT_POOL_SCRATCH_LIMIT: integer = 15

Usage: FFT_POOL_SCRATCH_LIMIT {INTEGER}

Limits the memory usage of the FFT scratch pool, potentially reducing efficiency a bit [Edit on GitHub]

FLUSH_SHOULD_FLUSH: logical = T

Lone keyword: T

Usage: FLUSH_SHOULD_FLUSH

Flush output regularly, enabling this option might degrade performance significantly on certain machines. [Edit on GitHub]

OUTPUT_FILE_NAME: string

Usage: OUTPUT_FILE_NAME {filename}

Name of the output file. Relevant only if automatically started (through farming for example). If empty uses the project name as basis for it. [Edit on GitHub]

PREFERRED_DGEMM_LIBRARY: enum = BLAS

Usage: PREFERRED_DGEMM_LIBRARY SPLA

Valid values:

  • SPLA SPLA library

  • BLAS BLAS library

Specifies the DGEMM library to be used. If not available, the BLAS routine is used. This keyword affects some DGEMM calls in the WFC code and turns on their acceleration with SpLA. This keyword affects only local DGEMM calls, not the calls to PDGEMM (see keyword FM%TYPE_OF_MATRIX_MULTIPLICATION). [Edit on GitHub]

PREFERRED_DIAG_LIBRARY: enum = ELPA

Usage: PREFERRED_DIAG_LIBRARY ELPA

Valid values:

  • ELPA ELPA library

  • SCALAPACK ScaLAPACK library

  • SL ScaLAPACK library (shorthand)

  • CUSOLVER cuSOLVER (CUDA GPU library)

  • DLAF DLA-Future (CUDA/HIP GPU library)

Specifies the diagonalization library to be used. If not available, the ScaLAPACK library is used [Edit on GitHub]

PREFERRED_FFT_LIBRARY: enum = FFTW3

Usage: PREFERRED_FFT_LIBRARY FFTW3

Valid values:

  • FFTSG Stefan Goedecker’s FFT (FFTSG), always available, will be used in case a FFT library is specified and not available.

  • FFTW3 a fast portable FFT library. Recommended. See also the FFTW_PLAN_TYPE, and FFTW_WISDOM_FILE_NAME keywords.

  • FFTW Same as FFTW3 (for compatibility with CP2K 2.3)

References: Frigo2005

Specifies the FFT library which should be preferred. If it is not available, use FFTW3 if this is linked in, if FFTW3 is not available use FFTSG. Improved performance with FFTW3 can be obtained specifying a proper value for FFTW_PLAN_TYPE. Contrary to earlier CP2K versions, all libraries will result in the same grids, i.e. the subset of grids which all FFT libraries can transform. See EXTENDED_FFT_LENGTHS if larger FFTs or grids that more precisely match a given cutoff are needed, or older results need to be reproduced. FFTW3 is often (close to) optimal, and well tested with CP2K. [Edit on GitHub]

PRINT_LEVEL: enum = MEDIUM

Aliases: IOLEVEL

Usage: PRINT_LEVEL HIGH

Valid values:

  • SILENT Almost no output

  • LOW Little output

  • MEDIUM Quite some output

  • HIGH Lots of output

  • DEBUG Everything is written out, useful for debugging purposes only

Mentions:How to Converge the CUTOFF and REL_CUTOFF

How much output is written out. [Edit on GitHub]

PROGRAM_NAME: enum = CP2K

Aliases: PROGRAM

Usage: PROGRAM_NAME {STRING}

Valid values:

  • ATOM Runs single atom calculations

  • FARMING Runs N independent jobs in a single run

  • TEST Do some benchmarking and testing

  • CP2K Runs one of the CP2K package

  • OPTIMIZE_INPUT A tool to optimize parameters in a CP2K input

  • OPTIMIZE_BASIS A tool to create a MOLOPT or ADMM basis for a given set of training structures

  • TMC Runs Tree Monte Carlo algorithm using additional input file(s)

  • MC_ANALYSIS Runs (Tree) Monte Carlo trajectory file analysis

  • SWARM Runs swarm based calculation

Which program should be run [Edit on GitHub]

PROJECT_NAME: string = PROJECT

Aliases: PROJECT

Usage: PROJECT_NAME {STRING}

Mentions:Molecular Dynamics

Name of the project (used to build the name of the trajectory, and other files generated by the program) [Edit on GitHub]

RUN_TYPE: enum = ENERGY_FORCE

Usage: RUN_TYPE MD

Valid values:

  • NONE Perform no tasks

  • ENERGY Computes energy

  • ENERGY_FORCE Computes energy and forces

  • MD Molecular Dynamics

  • GEO_OPT Geometry Optimization

  • MC Monte Carlo

  • SPECTRA Computes absorption Spectra

  • DEBUG Performs a Debug analysis

  • BSSE Basis set superposition error

  • LR Linear Response

  • PINT Path integral

  • VIBRATIONAL_ANALYSIS Vibrational analysis

  • BAND Band methods

  • CELL_OPT Cell optimization. Both cell vectors and atomic positions are optimised.

  • WFN_OPT Alias for ENERGY

  • WAVEFUNCTION_OPTIMIZATION Alias for ENERGY

  • MOLECULAR_DYNAMICS Alias for MD

  • GEOMETRY_OPTIMIZATION Alias for GEO_OPT

  • MONTECARLO Alias for MC

  • ELECTRONIC_SPECTRA Alias for SPECTRA

  • LINEAR_RESPONSE Alias for LR

  • NORMAL_MODES Alias for VIBRATIONAL_ANALYSIS

  • RT_PROPAGATION Real Time propagation run (fixed ionic positions)

  • EHRENFEST_DYN Ehrenfest dynamics (using real time propagation of the wavefunction)

  • TAMC Temperature Accelerated Monte Carlo (TAMC)

  • TMC Tree Monte Carlo (TMC), a pre-sampling MC algorithm

  • DRIVER i-PI driver mode

  • NEGF Non-equilibrium Green’s function method

References: Ceriotti2014, Schonherr2014

Mentions:How to Converge the CUTOFF and REL_CUTOFF, ⭐Electronic band structure from GW, ⭐GW + Bethe-Salpeter equation, ⭐Time-Dependent DFT, ⭐X-Ray Absorption from TDDFT, ⭐Real-Time Propagation and Ehrenfest MD, ⭐Monte Carlo

Type of run that you want to perform Geometry optimization, md, montecarlo,… [Edit on GitHub]

SAVE_MEM: logical = F

Lone keyword: T

Usage: SAVE_MEM

Some sections of the input structure are deallocated when not needed, and reallocated only when used. This reduces the required maximum memory [Edit on GitHub]

SEED: integer = 2000

Usage: SEED {INTEGER} .. {INTEGER}

Initial seed for the global (pseudo)random number generator to create a stream of normally Gaussian distributed random numbers. Exactly 1 or 6 positive integer values are expected. A single value is replicated to fill up the full seed array with 6 numbers. [Edit on GitHub]

TRACE: logical = F

Lone keyword: T

Usage: TRACE

If a debug trace of the execution of the program should be written [Edit on GitHub]

TRACE_MASTER: logical = T

Lone keyword: T

Usage: TRACE_MASTER

For parallel TRACEd runs: only the master node writes output. [Edit on GitHub]

TRACE_MAX: integer = 2147483647

Usage: TRACE_MAX 100

Limit the total number a given subroutine is printed in the trace. Accounting is not influenced. [Edit on GitHub]

TRACE_ROUTINES: string

Usage: TRACE_ROUTINES {routine_name1} {routine_name2} …

A list of routines to trace. If left empty all routines are traced. Accounting is not influenced. [Edit on GitHub]

WALLTIME: string

Aliases: WALLTI

Usage: WALLTIME {real} or {HH:MM:SS}

Maximum execution time for this run. Time in seconds or in HH:MM:SS. [Edit on GitHub]