met-jobs@lists.reading.ac.uk
August 2016
Message 22

[Periods|Index by:DateThreadSubjectAuthor|Date:PreviousNext|Thread:(Previous)(Next)|List Information]

[Met-jobs] PhD student/Early career Postdoc - Scientific programming

From Zbigniew Piotrowski <zbigniew.piotrowski@imgw.pl>
To <met-jobs@lists.reading.ac.uk>
Date Tue, 9 Aug 2016 12:38:08 +0200

PhD student/Early career Postdoc - Scientific programming

Project: Numerical Weather Prediction for sustainable Europe

Salary: 4000-5000/6000-9000 PLN net of tax (approx. 1000/2000 EUR) and state health
insurance. Benefits e.g. private health insurance scheme are available. Final salary will be
negotiated commensurate with experience. Early career status denotes that (the first) PhD
degree was granted no earlier than in 2011.

Duration: three years (PhD student), at least one year and up to three years
(Postdoc), with the possibility of extension to the fourth year subject to resources.

Project leader: dr Zbigniew Piotrowski

Location: IMGW-PIB headquarter in Warsaw, Poland

-----------------------------------------------

PDF version of this ad: http://imgw.pl/attachments/2098_firstteamoffer2.pdf

-----------------------------------------------

IMGW-PIB and the Project

The Institute of Meteorology and Water Management - National Research Institute is a research-
development unit and the Polish national meteorological and hydrological service. IMGW-
PIB carries research within the European Consortium for Small-scale Modeling (COSMO)
www.cosmo-model.org, a coalition of weather services on the development of operational nu-
merical weather prediction models.
The Numerical Weather Prediction for Sustainable Europe project is financed by the
European fund within the Smart Growth Operational Programme 2014–2020, Measure
4.4. “Boosting human potential in R&D sector“, through the First Team 1/2016 award of
Foundation for Polish Science. The project aims at accelerating the operationalization of
the EULAG dynamical core of COSMO model, complementary to the CELO priority project
of COSMO Consortium.
The research and implementation tasks of this Project include integration of EULAG dy-
namical core within the COSMO suite of MeteoSwiss, the first operational GPU-based weather
forecast. Running redesigned weather forecast suite on GPU results in energy savings by a
factor of 5. Therefore, it is highly desirable to exploit modern supercomputing architectures
to reduce the CO2 footprint of weather forecasting. Furthermore, application of the world-
leading EUlerian-LAGrangian EULAG research model(www.eulag.org) is expected to benefit
the forecasting scores and to augment Alpine weather prediction at the sub-km resolutions.

Summary of the position

The focus of the position is on the implementation of the EULAG dynamical core into the
modern software engineering framework using domain specific language STELLA and GridTools
library. This will be achieved by performing analytical work on the legacy Fortran 90 code of
EULAG dynamical core, identifying the possibilities for optimal implementation of the stencil
operations with corresponding boundary conditions. Furthermore, the Fortran code will be
translated into C/C++ using STELLA/Gridtools. Backends for the optimal integration on the
newest generation of accelerators (GPU/Xeon Phi) will be created. In parallel, the reference
Fortran + OpenACC formulation will be developed. Ultimately, the COSMO-EULAG suite
will be tested on accelerators aiming at the operational use.

Main duties and key responsibilities

• Complete implementation of the EULAG dynamical core in STELLA DSL and/or Grid-
tools library and integration with COSMO GPU weather forecasting framework, in par-
allel to the existing Runge-Kutta dynamical core.
• Implementation of Fortran/OpenACC version of EULAG dynamical core within the
COSMO-GPU weather forecasting framework.
• Participation in the development of the reference Fortran version of EULAG dynamical
core, in particular with regards to the optimal implementation of the three-dimensional
generalized Laplace operator in curvilinear coordinates in both: Cartesian and spherical
geometries.
• Preparation of testing frameworks and documentation.
• Preparation of scientific publications for the leading journals.
• Communicating the project results by public dissemination activities and presentations
at the workshops and conferences.

Personal attributes


• Enthusiasm to work independently and in a small team.
• Positive attitude towards working in an international environment.
• Ability to communicate with professionals and the general public.
• Dedication to advance the cutting-edge technologies of scientific programming in the area
of the computational fluid dynamics.
• Willingness to accept the operational priorities of the national weather service environ-
ment.
• Commitment to advance the scientific and software engineering standards at the IMGW-
PIB.

Qualification and experience required


Commensurate with the applicant’s career stage:
• Excellent knowledge of modern C++ and Unix working environment.
• Experience in the development of the scientific software and/or open-source software.
• Experience with high-performance computing and parallel programming.
Desired qualifications:
• Experience with CUDA/OpenACC GPU and hybrid MPI/OpenMP programming.
• Experience with C++ 11 and STL/BOOST libraries would be a significant advantage.
• Experience with software quality assurance, version control tools and testing frameworks.• Experience in the development of meteorological/geophysical software, including EULAG and COSMO models.
• Scientific achievements and peer-review publication record.
• Experience with L A TEX.
The working language is English. Candidate must be able to communicate, read and write
fluently in English. The successful candidate is expected to be able to travel several times a
year and collaborate with partners at MeteoSwiss in Zurich. She/He will also participate in the
training sessions at the leading institutions and/or spend a number of weeks at the fellowships
with the foreign partners. Candidates at the PhD level are expected to be enrolled in the
recognized programme of PhD studies or to enrol at the earliest possible opportunity, e.g. at
the University of Warsaw. Candidates advanced in the PhD processes but who do not possess
the PhD student status (e.g. expecting to defend the degree within one year) will be reviewed
on a case-by-case.

Our offer


We offer an unique position which provides the opportunity to work on leading edge software and
hardware technologies applied to the next-generation numerical weather prediction for Europe,
to collaborate with internationally renowned research groups and major weather services, and
to develop production quality weather prediction software, within a motivated team. We offer
opportunities for training in the field of modern technologies, international internships and
cooperation, and a competitive salary (based upon qualifications and experience).

Documentation


Please send the following information to the propoze@imgw.pl with the keyword DSL in the
subject:
• CV,
• cover letter explaining the interest and skills of the applicant,
• email addresses (and, if possible, phone numbers) of the two academic/professional refer-
ees who may be contacted by the recruiting committee,
• proof of the recent degree certificate and the transcript of records or similar document
(only at the entry PhD level),
• list of scientific and software development achievements,
• other proofs of qualifications as desired, up to five journal publications,
• passport-size photo.

Deadline and policies

Initial consideration will be given to applications received by the August, 28th. After this
date, the applications will be considered as long as the position is filled. Employment will start
on October 1 st , 2016 or as soon as possible after this date. Submission of the application implies
consent to the processing of the personal data for the purpose of the selection process. IMGW-
PIB is committed to the equal opportunity policy and all suitable candidates are encouraged
to apply.

Attachment: firstteamoffer2.pdf
Description: Adobe PDF document



Go to: Periods · List Information · Index by: Date (or Reverse Date), Thread, Subject or Author.