Upcoming Course: Python Universe for HPC with examples

The head of the HPC Lab, Prof. Ana Proykova and the Lab team invite you to a free intensive HPC training, which will be held ONLINE on May 27th 2022, 10:00 – 18:00 (EEST).

Here is more information about the course: https://hpc-portal.eu/node/1350

  • Title: Python Universe for HPC with examples: Setup, configuration, modules, preparing the code, running jobs, adjusting the code productivity;
  • Event/Course Start: Fri, 27/05/2022 – 10:00 EEST;
  • Event/Course End: Fri, 27/05/2022 – 18:00 EEST;
  • Event/Course Format: Online, Blended (mixture of live and self-paced);
  • Scientific Domain: Computer sciences, Computer Engineering, Electrical Engineering, Telecommunications;
  • Technical Domain: Artificial intelligence (AI), machine and deep learning;
  • Level of Instruction: Intermediate, advanced;
  • Sector of the Target Audience: Research and Academia, Industry;
  • Language of Instruction: English.

In order to participate in the trainings after the course, you must register on email address: metodi.pankov@gmail.com

Upcoming Hartree Centre HPC Courses

1. Practical Guide to High Performance Computing (HPC)

2. How to Scale Up Your Company’s  Code

3. Performance Analysis for Hartree Centre Systems

World record for the longest term reliable trajectory for the Lorenz system achieved on Nestum!

Hristov, I.Hristova, R., et al. achieved a world record for the longest term reliable trajectory for the Lorenz system on Nestum in their work:

Hristov, I., Hristova, R., Dimova, S., Armyanov, P., Shegunov, N., Puzynin, I., Puzynina, T., Sharipov, Z., Tukhliev, Z., On the efficient parallel computing of long term reliable trajectories for the lorenz system (2021) CEUR Workshop Proceedings, 2933, pp. 78-88, SJR (0.177 – 2020), https://www.scopus.com/inward/record.uri?eid=2-s2.0-85113369313&partnerID=40&md5=9a7851e223a8df1da6a75abd2c20fbf9

The previous record was achieved on one of the largest supercomputing clusters in the world – Tianhe. 

The HPC lab team sends our congratulations to the paper authors and we wish them many more similar records in the future.

PRACE Advanced HPC Workshop

Registration for the PRACE Advanced HPC Workshop is opened:
This special online advanced HPC track of courses consists of three separate training events on advanced HPC topics. The advanced track begins with the Message Passing Interface (MPI), the dominating programming model, continues with node level performance optimisation and finishes with hybrid programming. All the advanced courses include effective and personalised, practical hands-on sessions with relevant exercises. The courses include small group breakout rooms.
Target Audience & Prerequisites: These courses are targeted to intermediate and advanced level users, who already have the basic knowledge of HPC and parallel computing skills.
You need to have previous knowledge to participate in all 3 courses of this PRACE Advanced HPC Workshop, in particular you should have:
– Good knowledge of C/C++ or Fortran,
– Good knowledge of threading using OpenMP,
– Basic knowledge of modern CPU architectures,
– Be able to work on the Linux command line.
All trainers are the top specialists in the leading PRACE Training Centres and their partner institutions.
17-20 May 2022 – Parallelization with MPI @ VSC Vienna
7-9 June 2022 – Node Level Performance Optimization @ CSC
22-24 June 2022 – Hybrid programming in HPC – MPI+X @ LRZ
More information and registration:

Successful participation in the BonsAPPs AI challenge

Vasil, an Expert at HPC-lab, participated as an AI developer in BonsAPPs’ AI challenge as a part of the AICAD_SC  team.

During this  challenge a so-called AI Asset was developed that can recognize a driver’s  gestures by using a single, cheap and low-resolution thermal camera. This asset was then optimized and deployed to low-power edge platforms such as the NVidia jetson nano.

Vasil also gave a in-person presentation during a two-day hackathon in Bologna, Italy, presenting the team’s results.

AICAD_SC successfully completed stage one of the hackathon, earning a total score 7.37/10.



The team of the HPC laboratory participate in the project EUMaster4HPC (formerly HERCULES)

The project will be presented in at  The EuroHPC Summit
Week 2022 (22 March 2022 – 24 March 2022, Paris France).
16h30- 18h00: Session: EUMaster4HPC, presentation of the new Master programme.
Moderator: Daniel Opalka, Programme Officer R&I of the EuroHPC JU
Format: short presentations+ panel discussion
  • Pascal Bouvry, Coordinator of the project EUMaster4HPC tp present the project ( 10min);
  • Short presentations of 6 partners Universities of the programme to present their specificity and how they are complementary for the Master.
Programme (5min each)

-Maria Ribera, Universitat Politècnica de Catalunya
-Harald Köstler, Friedrich-Alexander-Universität Erlangen-Nürnberg
-Laura Grigori, Sorbonne Université
-Ana Proykova, Sofia University St. Kliment Ohridski
-Aryan Eftekhari, Università della Svizzera Italiana
-Dirk Pleiter, Kungliga Tekniska Hoegskolan
-Katerine Saleme Ruiz, University of Luxembourg

  • Jean-Philippe Nominé , Vice-President for Research at ETP4HPC,  Presentation of ET4HPC, private partner in the project (5min);
  • Panel Discussion/ Q&A (40min) with Fabrizio Magugliani (ET4HPC) + a representative of ErasmusMundo (Elisa Molinari).

The EuroHPC Summit Week (EHPCSW) 2022 will gather the main European HPC stakeholders from technology suppliers and HPC infrastructures to scientific and industrial HPC users in Europe. As in previous years, PRACE, the Partnership for Advanced Computing in Europe, co-locates its Scientific and Industrial Conference (PRACEdays22) at the EuroHPC Summit Week 2022. Furthermore, the EHPCSW 2022 will provide a great opportunity for the attendees to network.

The main organisers of the EHPCSW 2022 are the Partnership for Advanced Computing in Europe (PRACE), and the European Technology Platform for High-Performance Computing (ETP4HPC). The EuroHPC Joint Undertaking (EuroHPC JU) and the European Commission (EC) are also involved in the organisation of the conference. The logistical organisation is supported by a local host: for the 2022 edition, this will be GENCI (Grand équipement national de calcul intensif).


The HPC laboratory team via the Sofia University is involved in the Project HERCULES

The HPC laboratory team via the Sofia University is involved in the Project HERCULES, Funder by the EuroHPC JU (Horizon 2020).
The project aims at establishment  the first pan-European Master for High Performance Computing (HPC).
As of 15 February 2022, students can apply to EUMaster4HPC, the first pan-European Master for High Performance Computing (HPC). With funding from the EuroHPC Joint Undertaking, the master programme will train the next generation of HPC experts in Europe and provide students with outstanding career prospects in the rapidly expanding field of HPC.From September 2022, eight European universities will offer EUMaster4HPC courses to a first cohort of 50 students. The eight universities are the following:  University of Luxembourg (Luxembourg), Universitat Politècnica de Catalunya (Spain), Politecnico di Milano (Italy), Friedrich-Alexander-Universität Erlangen-Nürnberg (Germany), Sorbonne Université (France), Sofia University St. Kliment Ohridski (Bulgaria)Università della Svizzera Italiana (Switzerland), Kungliga Tekniska Hoegskolan (Sweden).

By bringing together major players in HPC education in Europe, EUMaster4HPC unifies existing programmes into a common pan-European curriculum to educate students in areas such as the design, deployment, operation, and/or the use of current and future generation HPC and HPC-related technologies in Europe.

EUMaster4HPC emphasises collaboration across Europe with innovative teaching paradigms such as co-teaching and strong connections with the industry and the best specialists in HPC education to develop the most relevant and up-to-date teaching contents.

This unique pan-European curriculum will provide a coordinated approach to HPC education and training in Europe that could be adopted by other universities.

The creation of a European network of HPC specialists will strengthen mobility and mutual support between students, teachers and industrial experts. A particular focus on mobility of students and teachers will enable students to rapidly gain experience through internships and exposure to European supercomputing centres.

Students can apply online through the application form until 15 May 2022.

EUwomen4future campaign seminar

On International Women’s day, Commissioner Mariya Gabriel will co-host with President Roberta Metsola a celebratory event, in Strasbourg and online, with four prominent women, featured in the #EUWomen4future social media campaign and a recipient of the EU Prize for Women Innovators. This campaign, which ended in March 2021, put a spotlight women’s professional achievements in culture, education, sport, and science.

The seminar will take the form of a conversation between 6 accomplished women, one of which will be Prof. Ana Proykova –  HPC lab’s head.

The seminar will be livestreamed on the 8th March 2022 here: How to celebrate extraordinary women: #EUwomen4future campaign seminar | European Commission (europa.eu)


Intel OneAPI toolchain now available on Nestum

(За български, четете надолу)

On December 8, 2020 Intel Parallel Studio was superseded by the new Intel toolchain – oneAPI.
Starting today, oneAPI (Basic + HPC toolkit) 2022 is available on the Nestum cluster and can be used by first loading the following module:

module load intel/oneapi-2022

This will load all the necessary binaries and libraries in your $PATH (eg. icc, ifort, intel python 3.9 etc.) and $LD_LIBRARY_PATH (eg. MKL) environmental variables.
Please note, oneAPI is a major upgrade over the old Intel Parallel Studio XE! There are many new tools such as the icx, icpx compilers for Intel’s C++ superset – Data Parallel C++ (DPC++).
All the “classic” tools from the old PSXE are still available under the same names – icc, ifort, icpc etc.
More information about the new compilers can be found here:

To reduce potential issues, please use the “classic” tools, unless you explicitly need the capabilities of the new ones (e.g. you are trying out DPC++). To cite the first of the above mentioned links:

ICX is a new compiler. It has functional and behavioral differences compared to ICC. You can expect some porting will be needed for existing applications using ICC.

A comprehensive webinar on porting your code to ICX/IFX can be found here: Porting Intel Parallel Studio Code to Intel oneAPI – YouTube


На 8.12.2020 г. Intel Parallel Studio XE беше наследено от oneAPI.
От днес, oneAPI (Basic + HPC toolkit) 2022 е налично на Нестум и модулът може да бъде зараден със следната команда:

module load intel/oneapi-2022

Това ще зареди всички необходими изпълними файлове и библиотеки в променливите на средата: $PATH (напр. icc, ifort, intel python 3.9 и т.н.) и $LD_LIBRARY_PATH (напр. MKL).
Внимание, oneAPI е сериозен ъпгрейд над старото Intel Parallel Studio XE! Този пакет включва много нови инструменти като компилаторите icx, icpx за новото “надмножество” на С++ разработелно от Интел – Data Parallel C++ (DPC++).
Разбира се, всички “класически” инструменти като  icc, ifort, icpc са налични в пакета (и модула).

По подробно с новите компилатори и тулчейни може да се запознаете от статиите (на английски):

С цел минимизиране на потенциални проблеми при компилация, препоръчваме да използвате “класическите” иструменти и компилатори, освен ако експлицитно не се нуждаете от способностите на новите такива (напр. пробвате да компилирате DPC++ код).

Подробен уебинар за портване на icc/ifort код към icx/ifx може да бъде намерен тук (на английски):

Porting Intel Parallel Studio Code to Intel oneAPI – YouTube