CP2K-UK

Lead Research Organisation: King's College London
Department Name: Physics

Abstract

CP2K is a state-of-the-art software package for atomistic simulation using Density Functional Theory and related methods, with a wide range of capabilities and excellent performance that can be used to address problems in the fields of materials science, computational chemistry and biochemistry. However, despite these advantages, the code presents a formidable learning curve for new users, with the result that many groups choose to use other codes, or are simply unable to tackle highly complex and demanding modelling problems.

We propose to create a network 'CP2K-UK' for UK researchers to build increased capability to both use effectively and develop or extend CP2K in their research. The network will be led by a collaboration between EPCC, UCL and KCL. The investigators Bethune (EPCC), Slater and Watkins (UCL) have strong track records in CP2K development - to date, the only UK contributors to the code. They will support a development project in the Kantorovich group (KCL), training a PDRA in the essential software development skills to successfully add new functionality to CP2K.

The development project will implement new algorithms to extend the ability of CP2K to simulate larger and more complex systems which are of key importance to UK research today, for example: understanding the function of large biomolecules, porous systems such as zeolites and Metal-Organic Frameworks (MOFs), design of new materials for fuel cells, nanotechnology including scanning probe microscopies, and next-generation catalysts to create fuels from solar radiation.

Publications

10 25 50
 
Description CP2K is an ab initio electronic structure and molecular dynamics simulation software package. This work concerns mainly on the development of the software in terms of both improvement in functionality as well as efficiency.

1. Implementation of Generalised Langevin Dynamics with thermal region support: The standard ab initio molecular dynamics simulation assumes box of particles with a constant total energy (NVE); to simulate a box that is connected to a heat bath and has a constant temperature (NVT), a thermostat will have to be used during the simulation. This generally involves adding and subtracting fictitious forces to the motion of particles in such a way that overall dynamics will obey the Boltzmann distribution (canonical ensemble). Langevin dynamics is one such approach, with the kinetic energy being added into and taken out of the system via a random distribution. The original version of the CP2K has Langevin dynamics implemented for all of the particles in the simulation cell. The improvement allowed different regions of the simulation cell to be defined, each can perform either Langevin NVT dynamics with a set temperature or NVE dynamics. This allows stochastic boundary conditions (Langevin dynamics only for atoms in the buffer region between the system of interest and the heat bath) calculations (L. Kantorovich, Phys. Rev. B 78, 094304 (2008)) to be carried out.

2. Implementation of multiple atomic lists in fragment definition for Basis Set Superposition Error (BSSE) calculations: The original implementation of CP2K allows only one atomic list in the definition of a structural fragment. Since the CP2K list input format can only allow one continuous list for the atomic indices (e.g. atoms 1..5 meaning 1, 2, 3, 4, 5), it can become a hassle for users to define complex fragments that may involve atoms from several continuous lists with gaps in between---they will have to reorganise the atoms in their input file which may be extremely inconvenient. This a minor improvement which allows the fragment to be defined by several lists if necessary, thus removing the user hassle to reorder atoms in their model.

3. Implemented the filter matrix diagonalisation method (mostly intended for metals and their interfaces with biological systems). The method has been thoroughly tested and a significant time has been spent on its optimisation. It is tested (on bulk and slab metallic of up to 6500 atoms and semiconductor systems as well as water boxes with over 10000 atoms) to give a speed-up of up to 10 times compared with standard diagonalisation methods.

4. Implemented the REPEAT variant of the RESP charge fitting method which allows accurate classical representation of the electrostatic potential by point charges. This allows replacing in part expensive DFT simulations with classical MD.

5. A bug in the dimer method for finding transition states has been found and fixed. A miscellaneous feature introduced for a smooth simulation with minimum user intervention.

6. An automatic installer for the whole CP2K suit (including dependent libraries and tools) has been written (a quite a significant piece of code) which is designed to simplify installation of the code as much as physically possible for an arbitrary platform. It recognises already installed packages and reuses their libraries.

Another objective of this project in collaboration with developers in the UK (EPCC) and in Switzerland (UZH, ETHZ, PSI) to improve user interface and documentation. For this, so far:

- We have written an emacs major mode for the CP2K input file. This has been made available and put into the public domain (GitHub), and is linked to the CP2K website.
- a new simplified access to the cp2k manual pages
- We have written several tutorials and FAQs on how to perform basic and some more advanced calculations using CP2K and published the tutorials onto the main CP2K website, which is available to the public.
- We have contributed to the training courses for CP2K organised by our EPCC collaborators.
- We written an interface between TETR/LEV00 and CHIMERA.
We have also attended the CompMat2014, where we have given a lecture introducing CP2K to a wider audience in the Chemistry and Material Science field.
Exploitation Route The development on the software will be on-going. As well as tutorials and help materials for the users, we are presently in the process of providing help materials for developers such as descriptions of the API functions in the DBCSR sparse matrix algebra library used by CP2K, and the parallel strategy and data structures currently implemented in the code. These types of information are expected to help to speedup future developments of the software and help new developers to become familiar with the code, and are something at the moment that is very much lacking. New functionalities that has been developed in this project, which is still on-going, will be advertised both in the user mailing lists and the training courses to raise awareness when they become available. For this purpose, two CP2K meetings were organised in London on the 6th of February 2015 and 22nd of February 2016.
Sectors Chemicals,Digital/Communication/Information Technologies (including Software),Education,Electronics,Energy,Pharmaceuticals and Medical Biotechnology,Other

 
Description The implementation of Generalised Langevin Dynamics with thermal regions have been used by a research group at King's College London. The emacs tool has been introduced during the CP2K training courses, and received positive user feedbacks. It is now included as part of the CP2K package. Although we have not yet made an effort to collect data on the impact enabled by our project, the online tutorials will be helpful for beginner users to use the code and perform realistic and scientifically useful calculations. The installer has been used during our tutorials and workshops, and at least on one occasion a user was able to finally install the code on his machine without any help. The corrected dimer method is being used in our group at KCL, and the filter matrix implementation is currently being prepared for intensive simulations with metal water interfaces in collaboration with the group of Prof. F. Tielens. The REPEAT method has been used by a UCL group and a paper has been submitted. The interface between TETR/LEV00 and CHIMERA has been used in L. Kantorovich's group and in Lincoln.
First Year Of Impact 2017
Sector Chemicals,Digital/Communication/Information Technologies (including Software),Education,Electronics,Energy,Pharmaceuticals and Medical Biotechnology,Other
Impact Types Cultural

 
Description EPCC 
Organisation University of Edinburgh
Country United Kingdom 
Sector Academic/University 
PI Contribution We contributed to the code, to the tutorials and hosting user meetings.
Collaborator Contribution Help in code development and courses on code development.
Impact New features added to the code, website of the code improved.
Start Year 2013
 
Description UCL CP2K 
Organisation University College London
Country United Kingdom 
Sector Academic/University 
PI Contribution A code is being developed which our partners will use as well.
Collaborator Contribution Our partners helped us a lot with advise and their experience in using and developing the code.
Impact Code website and code development.
Start Year 2013
 
Description UCL CP2K 
Organisation University College London
Country United Kingdom 
Sector Academic/University 
PI Contribution A code is being developed which our partners will use as well.
Collaborator Contribution Our partners helped us a lot with advise and their experience in using and developing the code.
Impact Code website and code development.
Start Year 2013
 
Description Zurich 
Organisation ETH Zurich
Country Switzerland 
Sector Academic/University 
PI Contribution We are currently developing a software and the website which would contribute to this partner as well as to the whole community.
Collaborator Contribution They have provided explanations and help to us when working on the software and the website.
Impact We have made significant changes to the code website by simplifying it, putting tutorials and adding new functionality to the code.
Start Year 2013
 
Title CP2K 
Description CP2K is a program to perform atomistic and molecular simulations of solid state, liquid, molecular, and biological systems. It provides a general framework for different methods such as e.g., density functional theory (DFT) using a mixed Gaussian and plane waves approach (GPW) and classical pair and many-body potentials. 
Type Of Technology Software 
Year Produced 2014 
Open Source License? Yes  
Impact New functionality added and the website improved. 
URL http://cp2k.org
 
Title cp2k graphical interface tool 
Description The software enables one to interface KCL software TETR (needed to create an input to CP2K, VASP or SIESTA density functional codes) and LEV00 (to work out their output) with CHIMERA graphical molecular visualisation software. 
Type Of Technology Software 
Year Produced 2017 
Open Source License? Yes  
Impact The developed software allows one to improve enormously the experience with TETR and LEV00, simplifies working with these codes and hence working with mentioned DFT codes. 
URL https://nms.kcl.ac.uk/lev.kantorovitch/codes/lev00/index.html
 
Description CP2K workshop 2014-2017 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Type Of Presentation keynote/invited speaker
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact Annual workshops [in London (2014), ??? (2015), ??? (2016), Lincoln (2017)] were organised in which people from all over the UK came. We gave a number of talks with an introduction to the CP2K code as well as a number of how to topics. We also listed examples of real applications. Developers of the code made overview presentations. We invited people to share their experiences with the code. Participants were also given a chance to express their views on the current version of the code and its possible future developments.

People expressed interest in using the code and existing users have suggested a number of directions in which the code should be developed.
Year(s) Of Engagement Activity 2014,2015,2016,2017
URL https://www.epcc.ed.ac.uk/content/cp2k-uk-workshop-2014
 
Description SSI Workshop: Impact of international research software collaboration 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Supporters
Results and Impact Workshop hosted by the Software Sustainability Institute to report back on the impact of software development projects on international research collaboration.
Year(s) Of Engagement Activity 2018
URL https://ibethune.github.io/files/CP2K-impact-talk.pdf