LUMI’s full system architecture revealed
The AMD MI250X GPU is in a class of its own now and for a long time to come. The technical supremacy and performance per watt were the primary reasons why AMD’s MI250X GPUs were selected for LUMI, explains Pekka Manninen, Director of LUMI Leadership Computing Facility.
The full system architecture of LUMI is the following:
The LUMI system is supplied by Hewlett Packard Enterprise (HPE), based on an HPE Cray EX supercomputer.
The GPU partition will consist of 2560 nodes, each node with one 64 core AMD Trento CPU and four AMD MI250X GPUs.
Each GPU node features four 200 Gbit/s network interconnect cards, i.e. has 800 Gbit/s injection bandwidth.
Each MI250X GPU consists of two compute dies, each with 110 compute units each, and each compute unit has 64 stream processors for a total of 14080 stream processors.
The committed Linpack performance of LUMI-G is 375 Pflop/s.
The MI250X GPU comes with a total of 128 GB of HBM2e memory offering over 3.2 TB/s of memory bandwidth.
A single MI250X card is capable of delivering 42.2 TFLOP/s of performance in the HPL benchmarks.
In addition to the GPUs in LUMI there is another partition (LUMI-C) using CPU only nodes, featuring 64-core 3rd-generation AMD EPYC™ CPUs, and between 256 GB and 1024 GB of memory. There are 1,536 dual-socket CPU nodes in total. LUMI-C was #5 on the November 2021 Graph500 list and #76 on the November 2021 Top500 list.
LUMI also has a partition with large memory nodes, with a total of 32 TB of memory in the partition.
When completed LUMI will take over 150m2 of space, which is about the size of a tennis court. The weight of the system is nearly 150 000 kilograms (150 metric tons).
ETAIS & ELIXIR seminar UT HPC Usage 101
During this UT HPC Usage 101 lecture course we will provide the participants basic knowledge required to use to submit, monitor, and control jobs on the compute nodes of UT HPC. We will talk about the principles of computing in a cluster and the difference between computing directly from the command line of a private server. We will introduce the good practices and standards of behavior that good cluster usage practice requires. Also the convenient graphical tools for monitoring the resource of your work will be covered. Finally, we might touch on the basic Slurm commands.
We expect the users to have basic LINUX command line experience, but very little or no cluster compute experience yet.
Learning outcome: Knowing how to submit, monitor and control jobs at UT HPC.
The seminar will take place in Zoom on 18th of November 2021 at 14.00 and will last approximately 90 minutes.
Registration is open at https://forms.gle/pVe86HVqjWB6GpLg8 and will close on 16th of November or when the course gets full.
An OpenMP training organized by the LUMI User Support Team
The course will take place online on November 9. – 10.
This course is part of the first OpenMP training organized by the LUMI User Support Team (LUST) and supported by CSC – IT Center for Science Ltd (Finland) and CSC Autumn of HPC 2021 (five modules program on teaching essential skills in parallel programming for modern GPU-accelerated supercomputers) under PRACE Training Centres activity.
This training aims to help the users to port their code to LUMI, the European pre-exascale supercomputer, that will achieve its high computing power thanks to a large number of nodes with AMD GPUs.
This course address the use of OpenMP for programming co-processors such as GPUs. It focuses on how to get the best out of OpenMP in terms of performance by exploring the implications of possible OpenMP parallelization strategies. Advanced topics such as asynchronous execution, interoperability with CUDA/HIP and the use of multiple GPUs are covered.
Registration is obligatory since the details to access the online course will be provided to the registered and accepted attendees only.
Read more about the course and register now at the PRACE Training website.
ENCCS GPU Hackathon 2021
The EuroCC National Competence Centre Sweden (ENCCS) will be hosting an online NVIDIA GPU hackaton that will take place on December 6th and December 13-15.
GPU Hackathons provide exciting opportunities for scientists to accelerate their AI research or HPC codes under the guidance of expert mentors from National Labs, Universities and Industry leaders in a collaborative environment. Teams of researchers are paired with experienced GPU mentors to learn and apply the accelerated and parallel computing skills needed by the scientific community. Both current or prospective users of large hybrid CPU/GPU HPC clusters who develop HPC and AI applications that could benefit from GPU acceleration are encouraged to participate!
Teams are expected to bring their own code or project to the event and be motivated to make progress during the hackathon. No advanced GPU skills required, but teams are expected to know the basics of GPU programming and profiling at the event.
The application deadline is October 31, 2021. Please click here to find out more about the event.
A free course on Shared Memory Programming with OpenMP
The EuroCC-funded UK National Competence Centre offers an online OpenMP course that will be supported by staff from EuroCC and PRACE to answer questions via chatboards and run regular online tutorials. This course covers OpenMP, the industry standard for shared-memory programming, which enables serial programs to be parallelised easily using compiler directives. Users of desktop machines can use OpenMP on its own to improve program performance by running on multiple cores; users of parallel supercomputers can use OpenMP in conjunction with MPI to better exploit the shared-memory capabilities of the compute nodes.
The course starts on November 1 and ends on December 3. All participants will be offered an account on ARCHER2, the UK National HPC Service, for the duration of the course.
The course is part-funded by the PRACE project and is free to all. Please follow this link for more information and registration.
EuroHPC SUMMIT WEEK TO BE HELD IN PARIS, MARCH 22-24
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 organizers of the EHPCSW 2022 are the Partnership for Advanced Computing in Europe (PRACE), and the European Technology Platform for High-Performance Computing (ETP4HPC). The European Commission (EC) will represent the EuroHPC Joint Undertaking (EuroHPC) in the organization of the conference. The logistical organization is supported by a local host: for the 2022 edition, this will be GENCI (Grand equipement national de calcul intensif).
For more information visit this link and stay tuned for more details.
THE NCC ESTONIA IS CELEBRATING ITS FIRST BIRTHDAY
The HPC Competence Centre of Estonia was established within the scope of the EuroCC project exactly one year ago – on September 1, 2020. Our first year of operation has been full of intensive work, exciting challenges and interesting meetings. We have just completed our annual report, which summarizes the activities and achievements that have taken place during the last year. We are pleased to confirm that the activities of NCC Estonia have proceeded according to initial plan. Most of the key perfomance indicators are fulfilled or even exceeded.
We would also like to congratulate the CASTIEL team, as they are also celebrating its first year of operation. Some essential deliverables of the CASTIEL and EuroCC projects are now available for public.
FF4EuroHPC TO LAUNCH THE SECOND OPEN CALL FOR HPC EXPERIMENTS
FF4EuroHPC is a European initiative that helps facilitate access to HPC related technologies for SMEs. Within the FF4EuroHPC project, European SMEs can develop unique products, innovative business opportunities and become more competitive by using European high-end HPC services. Whether it is running high-resolution simulations, doing large-scale data analyses, or incorporating AI applications into SMEs´ workflows, FF4EuroHPC connects business with cutting-edge technologies.
Two open calls will be offered through the project. The second call for proposals targets highest quality experiments involving innovative, agile SMEs and putting forward work plans built around innovation targets arising from the use of advanced HPC services. Priority will be given to consortia centred on SMEs that are new to the use of advanced HPC services.
Key call details
Submission Deadline: 29th September 2021, at 17:00 Brussels local time.
Expected duration of experiments: maximum 15 months with expected commencement 1st March 2022.
Funding for Call-2: The indicative total funding budget is € 5 M.
Maximum funding request per proposal: € 200,000 (covering all participants).
Find out all the details about the open call and application process: https://www.ff4eurohpc.eu/en/open-calls/open-call/
WORKSHOP ON SARS-COV2 DATA ANALYSIS and monitoring WITH GALAXY
The goal of this upcoming workshop is to build capacity in SARS-CoV-2 data analysis and data management, including data submission to ENA. After the workshop, all participants will be able to upload viral sequencing data, call all variants, create a variety of reports and create consensus alignments.
It will be a 4-day event introducing scalable and reproducible SARS-CoV-2 data analysis with Galaxy. The sessions will be pre-recorded and provided in advance. During the workshop, there will be live support in chat and live Q&A sessions, in which experts will answer questions.
PRACE AUTUMN SCHOOL 2021:
HARNESSING the EuroHPC Flagship supercomputers
The PRACE Autumn School 2021 will bring together undergraduate and graduate students and postdoctoral researchers in different disciplines of scientific computing.
The course consists of lectures and hands-on training on modern, GPU-accelerated high-performance computing: GPU programming and GPU code optimization at scale, as well as understanding and applying machine learning methods. The course includes also scientific case studies about using GPUs in various disciplines. The tutors and lecturers are experts in these fields providing various interesting aspects in the course topics.
The five days of school, organised by CSC – IT Center for Science, are filled with hacking, leisure activities and having good time together in Vuokatti, Finland. Also, the Autumn School includes a visit to the CSC’s Data Center in Kajaani, the home of one of the most powerful supercomputers in the world: LUMI. It is a unique opportunity for the participants to familiarize themselves with the cutting-edge data infrastructure in the field of HPC.
Data Science Seminar:
High-Performance Computing: Big Is Beautiful
High-Performance Computing (HPC) is the application of “supercomputers” to computational problems that are either too large for standard computers or would take too long to solve. HPC is a pervasive tool for both academia and industry. “Today, to Out-Compute is to Out-Compete” best describes the role of HPC.
A recent seminar brought together HPC users, administrators and visionaries. Researchers talked about their experiences with using HPC in their daily work. The administrator of Tartu HPC center introduced the backstage of managing large amounts of data. Finally, we took a closer look at the new supercomputer LUMI, which will launch in 2021, to take computing performance to a new unprecedented level.
Galaxy-ELIXIR webinars series:
Galaxy is supported by a large and active community distributed across the globe, which leads to rapid developments in every release. However, many of this features remain unknown by most of the users.
With this series of four webinars chaired by Hans-Rudolf Hotz, the global Galaxy community will present such advanced features, touching upon workflows, data processing and management, interoperability and interactive tools, among others.
The webinar series takes place between 3 and 24 March 2021.
A new supercomputer LUMI starts operating in summer 2021
A new supercomputer LUMI (Large Unified Modern Infrastructure) is being built in Finland. It is expected to open for users from July 2021.
At the time of installation, LUMI will be one of the world’s fastest computer systems. LUMI’s performance will be more than tenfold compared to one of Europe’s fastest supercomputer today (Piz Daint, Switzerland). LUMI will also be one of the world’s leading platforms for artificial intelligence.
LUMI will be capable of executing 375 petaflops, or more than 375 million billion calculations per second. Theoretically, it’s maximum performance will be more than 550 petaflops per second, equalling that of 1.5 million laptops.
LUMI will be located at CSC’s (IT Center for Science) data center in Kajaani town. LUMI is using 100% hydropowered energy. The waste heat of LUMI will produce 20 percent of the district heat of the area.
The supercomputer will be hosted by the LUMI consortium including ten European countries: Finland, Belgium, the Czech Republic, Denmark, Estonia, Iceland, Norway, Poland, Sweden, and Switzerland.
As a member of the consortium, Estonia can use 2% of LUMI’s supercomputer resources. This opens up unprecedented technological opportunities for Estonian HPC users.
estonia has Joined the eurocc project
Estonia is participating in the EuroCC project, that brings together the necessary expertise to set up a network of more than 30 National Competence Centres (NCCs) in HPC across Europe.
The NCCs will coordinate activities in all HPC-related fields at the national level and serve as a contact point for customers from industry, science, HPC experts, and the general public alike.
The EuroCC project aims to elevate the participating countries to a common high level in the fields of HPC, High Performance Data Analytics (HPDA) and artificial intelligence. Ultimately, the goal is to promote the use of HPC in Europe and to increase the competences of its users.
The EuroCC activities are coordinated by the High-Performance Computing Center Stuttgart (HLRS). The EuroCC project is funded 50 percent through H2020 (EuroHPC Joint Undertaking) and 50 percent through national funding programs within the partner countries.