Website University of Copenhagen
Job description
The PLTC department of the Department of Computer Science (DIKU) at the University of Copenhagen (UC) invites applicants to apply for a PhD Fellowship in “Massively Locked-Step Parallel Fully Isotropic Space Quantum Chemistry”. The program is part of the research program “Carbon Manifolds”.
The start date is (expected to be) May 1, 2025 or as soon as possible thereafter.
The Project For the past 80 years, the game changer in computational quantum chemistry has been to solve the Schrödinger equation for a given molecule with increasing accuracy in approximation with limited computational resources.
In this project, we subvert this fundamental condition by designing algorithms that can simultaneously analyze the entire isomer space consisting of millions of different molecules. The new theory facilitates extremely fast approximate quantum chemical calculations on surface manifolds derived from bond maps. By exploiting the shared structure of the isomer space, we will design massively lock-step parallel algorithms that maximize the use of multi-GPU (General Purpose Graphics Processor) hardware and scale linearly to millions of computational units. Our goal is to automate the discovery of molecules with desired physical and chemical properties: in this project, we will develop new computational methods, design and build an open-source software platform for performing quantum chemical computational experiments in full isomeric space on workstations and supercomputers, and apply them to socially important problems. The focus of the computational experiments in this project is to find particularly efficient organic photovoltaics (OPVs) in 2.2 billion fullerene molecules (up to C200) to help realize the green transition. In addition, the methods and software will be particularly helpful in discovering polyhedral molecules for drug delivery and nanomedicine. The software will be universally applicable and open source to the community.
Two new technological developments put the ambitions of the project within reach.
First: a combinatorial differential geometry framework (CM) called carbon clusters makes it possible to design extremely fast approximate quantum chemistry algorithms. This formalism makes it possible to derive lock-step parallel algorithms using the structure of the isomer space, in which all computational units perform exactly the same arithmetic operations in lock-step, maximizing the use of the GPUs and achieving peak performance throughput on a GPU cluster with minimal communication overhead.
Second: Cluster Perturbation Theory (CP), an extremely accurate formalism for computing molecular properties developed by two other Ph.D.s. CP improves on the previous “gold standard”, Coupled Cluster Theory (CC), but makes it easier to compute molecular properties via response theory, converges faster (quadratically), and is massively parallelized on GPU clusters with distributed memory.
The approach we plan to develop will 1) find potential top molecules for the fullerene-SubPC OPV device by using lock-step CM algorithms to compute grade-preserving fast rough approximations of the desired molecular properties for the full isomer space containing billions of different molecules, and 2) then perform a large number of high-precision quantum-chemistry molecular property calculations for these molecules using CP, identifying the optimal few candidate molecules for detailed study.
CM and CP will interact in two ways: first, we will perform high-quality CP reference calculations on the 20-100 candidates identified in the CM experimental calculations for 2.16×10^9 fullerenes from C20 to C200. These results will be used to check, identify biases, and guide and improve the development of approximate CM algorithms. The results will be further used to identify approximation biases and to train correction formulas.
In existing quantum chemistry algorithms, the input is a single molecule and the output is the molecular geometry, ground and excited state energies, and molecular properties. The input will be an isomer space and a set of desired molecular properties. The output will be the molecule in the isomer space that best matches the desired target properties.
If successful, this will fundamentally change the way new nanotechnologies and nanomedicines are designed.
We are looking for people in the fields of high performance computer science and computational physics. Applicants can have a background in several of the following areas:
Strong background in computational physics or computational chemistry.
Experience implementing robust, maintainable, high-performance numerical algorithms.
A good understanding of advanced design of parallel algorithms that scale to thousands of nodes and millions of cores.
Experience with high-performance parallel programming, preferably including C++ and underlying GPU parallel frameworks such as SYCL, CUDA, HIP, or OpenCL.
Thorough understanding of the GPGPU architecture and its impact on performance programming.
Experience with quantitative analysis of correctness/accuracy, performance and scalability.
Thorough understanding of molecular electronic structure/computational chemistry.
Our group and research – what can we offer? In the Avery group, we combine programming language research with novel computational methods for massively parallel numerical computation in physics and chemistry, ranging from abstract mathematics to efficient hardware implementations.
The group is part of the Department of Computer Science, Faculty of Science, University of Copenhagen (DIKU). The Programming Languages and Theory of Computation group at the Department of Computer Science, University of Copenhagen, focuses on the theory and practice of programming languages, including formal semantics, language design, program analysis and language-based security. The research also explores parallel programming paradigms to exploit the power of multi-core processors to improve program performance.
We offer creative and stimulating working conditions in a dynamic and international research environment.
The main supervisor is James Emil Avery, Associate Professor, Department of Computer Science, [email protected], direct line: +45 30 22 91 11: +45 30 22 91 11.
PhD program We offer a standard PhD program only for this position
If you have already been educated to the equivalent of a relevant Danish Master’s degree, you can study full-time for three years within the framework of a regular PhD program (5+3 program).
********************************************************************************
Qualifications required for a regular program To qualify for a regular PhD program, you must have completed a degree program equivalent to a Danish Master’s degree (180 ECTS/3 FTE BSc + 120 ECTS/2 FTE MSc) in a relevant subject area of the program, e.g., computer science, physics, or chemistry. For information on the eligibility of completed programs, please refer to the country-specific General Assessment and Evaluation Database.
Conditions of Employment in Regular Programs PhD researchers are employed on a full-time basis for a maximum period of 3 years.
A condition of employment is that you must be successfully registered as a PhD student at the Faculty of Science, University of Copenhagen. This requires the submission and acceptance of a program-specific application developed by the applicant.
The terms of appointment and remuneration are in accordance with the agreement between the Danish Ministry of Taxation and the Danish National Federation of Professional Associations in Academia. The position is governed by the Work Structure Agreement.
Duties and tasks
Carry out independent research projects under supervision
Completion of a PhD program of approximately 30 ECTS / ½ FTE
Participation in an active research environment, including periods of stay at other research institutions (preferably abroad)
Teaching and knowledge dissemination activities
Writing scientific papers for high-impact journals
Writing and defending a PhD thesis on the basis of the project
We require the following qualifications
Professional qualifications relevant to the PhD program
Relevant publications
Relevant work experience
Other relevant professional activities
Good language skills
***************************************************************************
APPLICATION AND ASSESSMENT PROCEDURE Your application (including all attachments) must be in English and submitted electronically by clicking on “Apply Now” below.
Please specify
Strongly motivated letter of application (maximum of one page)
Resume including information about your education, experience, language skills and other skills relevant to the position
Original B.Sc. or M.Sc. diploma and transcripts in the original language and, if issued in a language other than English or Danish, an authorized English translation. If not completed, a certified/signed copy of a recent transcript or a written statement from the institution or instructor is acceptable.
List of publications (if possible)
Letters of recommendation (if any)
APPLICATION DEADLINE: The application deadline is Thursday, February 27, 2025, +1, 23:59 GMT.
We reserve the right not to consider materials received after the deadline and not to consider applications that do not meet the above requirements.
After the deadline for further procedures, we will select a number of applicants for academic assessment by impartial expert assessors. You will be notified whether or not you have passed the assessment.
The assessors will evaluate the qualifications and experience of the shortlisted applicants against the research areas, technologies, skills and other requirements outlined above. The assessor will make a conclusion as to whether each applicant meets the qualifications and, if so, which of the two modalities. Applicants who are assessed will be given the opportunity to comment on the results. For more information on the recruitment process, please visit http://employment.ku.dk/faculty/recruitment-process/.
Interviews with selected candidates are expected to be held in Week 12.
Questions Specific information about the PhD Fellowship should be directed to the Principal Advisor.
For general information about doctoral study at the Academy, visit the Doctoral School website at https://www.science.ku.dk/phd/.
The University of Copenhagen wishes to reflect the surrounding community and invites all persons to apply for the position, regardless of their personal background. The University of Copenhagen is a member of the International Association of Research Universities (IARU) and one of Europe’s leading universities, committed to promoting research and teaching of the highest international standards. With its rich tradition and modern outlook, the University of Copenhagen offers students and staff the opportunity to develop their talents in an ambitious and eclectic environment. An efficient organization – with good working conditions and a collaborative work culture – creates the ideal framework for a successful academic career.
offer requirements
Strong background in computational physics or computational chemistry.
Experience implementing robust, maintainable, high-performance numerical algorithms.
Full understanding of advanced design of parallel algorithms that scale to thousands of nodes and millions of cores.
Experience with high-performance parallel programming, preferably including C++ and underlying GPU parallel frameworks such as SYCL, CUDA, HIP, or OpenCL.
Thorough understanding of the GPGPU architecture and its impact on performance programming.
Experience with quantitative analysis of correctness/accuracy, performance and scalability.
Thorough understanding of molecular electronic structure/computational chemistry.
Professional qualifications relevant to the PhD program
Relevant publications
Relevant work experience
Other relevant professional activities
Good language skills
offer benefits
We offer creative and stimulating working conditions in a dynamic and international research environment.
PhD program We offer a standard PhD program only for this position
If you have already been educated to the equivalent of a relevant Danish Master’s degree, you can study full-time for three years within the framework of a regular PhD program (5+3 program).
Employment conditions for the regular program PhD researchers are employed full-time for a maximum period of 3 years.
The terms of appointment and remuneration are in accordance with the agreement between the Danish Ministry of Taxation and the Danish National Federation of Professional Associations in Academia. The position is governed by the Work Structure Agreement.
To apply for this job please visit di.ku.dk.