EXCELLERAT goes digital: 36th INTERNATIONAL CAE CONFERENCE AND EXHIBITION
In partnership with FocusCoE and MaX, EXCELLERAT took part in the CAE2020 International Conference and Exhibition digital event. In their shared virtual booth “HPC-COES: MAX & EXCELLERAT,” visitors could read the prepared flyers, videos, websites, and EXCELLERAT’s social media channels. Visitors could also video/chat, share their contact details, or request a meeting.
EXCELLERAT’s Second Virtual All-Hands Meeting
Our second virtual project meeting took place from 9 to 11 December 2020. Representatives from all partners, a total of over 30 participants on each of the three days, presented the EXCELLERAT project tasks’ progress and challenges they faced throughout this year, and the need for a continuous evolution and mastering this final project year.
Success story: Towards an increased accuracy in the automotive field simulations: mesh optimising by using an a posteriori adjoint based error estimation
Success Story: Towards an increased accuracy in the automotive field simulations Success story # Highlights: Key words: Mathematical model: Adjoint based morphing Mesh optimising Industry sectors: Automotive, Aerospace Key codes used: FEniCS HPC, Unicorn HPC, Tetgen Fig 1: pressure field distribution over the car surface Organisations & Codes Involved: KTH Royal Institute of Technology in […]
High-performance computing challenge: How can engineers effectively leverage HPC technology?
When developing or improving products, engineers face the challenge of analysing complex simulations to test their designs. High-performance computing (HPC) is used to solve major problems in engineering by performing demanding operations such as simulations relatively quickly. Also, by processing more precise operations, HPC can avoid the errors that might be introduced by humans.
Enabling parallel mesh adaptation with Treeadapt: Enabling parallel mesh adaptation with Treeadapt
Success Story: Enabling parallel mesh adaptation with Treeadapt Success story # Highlights: Keywords: CFD Mesh adaptation Parallel performance Load balancing Industry sector: Aeronautics, High Performance Computing Key codes used: AVBP, Treepart, Treeadapt Mesh adaptation of an inviscid eulerian vortex inital mesh (left), First adapted grid (middle), Final mesh (left) Organisations & Codes Involved: CERFACS is […]
EXCELLERAT collaborates with the HiDALGO Centre of Excellence
HiDALGO – HPC and Big Data Technologies for Global Systems – and EXCELLERAT are successfully collaborating on tasks like improving HiDALGO’s Urban Air Pollution Pilot. After developing the pilot, HiDALGO sought support from EXCELLERAT during its application improvement phase. Together they pushed forward the pilot case in months improving mesh quality, speed-up, and parallel efficiency.
Success Story: Full Airplane Simulations on Heterogeneous Architectures
Success Story: Full Airplane Simulations on Heterogeneous Architectures Success story # Highlights: Keywords: Dynamic Load Balancing/DLB parallel performance airplane simulations GPU computing Co-execution heterogeneous computing Industry sector: Aeronautics Key codes used: Alya Snapshot of Q iso-surfaces of the turbulent flow around an airplane. Organisations & Codes Involved: Barcelona Supercomputing Center-Centro Nacional de Supercomputación (BSC-CNS) is […]
Nek5000 in EXCELLERAT: towards Exascale
One of EXCELLERAT’s tasks is improving Nek5000’s Exascale readiness, particularly for accelerators. We used an OpenACC version of the proxy-app Nekbone and focused on further optimising the small matrix-matrix kernels which constitutes most of the work in the full Nek5000. Tuning the TOpenACC directives improved performance of these kernels by almost a factor of two.
New EXCELLERAT Service Portal is launched
We are pleased to announce EXCELLERAT’s redesigned Service Portal launch. It presents a wide variety of resources in an easily-accessible format and will play an essential role in EXCELLERAT’s mission to create a Centre of Excellence in Engineering. The user-centric Portal offers separate dashboards tailored towards engineers, software developers, and the general engineering community.
Dynamic Load balancing for Airplane simulations on Heterogeneous pre-Exascale architectures
We can affirm with certainty that future Exascale systems will be heterogeneous, including accelerators such as GPUs. We can also expect higher variability on the performance of the various computing devices engaged in a simulation. In this context, dynamic load balancing (DLB) becomes a must for the parallel efficiency of any simulation code.
