[PDF][PDF] Enabling task parallelism for many-core architectures

PR Atkinson - 2021 - research-information.bris.ac.uk
The requirements placed on computer architectures from modern computational workloads
have driven constant performance improvements. In the last 15 years, the largest source of …

NUMA-Awareness as a Plug-In for an Eventify-Based Fast Multipole Method

L Morgenstern, D Haensel, A Beckmann… - International Conference …, 2020 - Springer
Following the trend towards Exascale, today's supercomputers consist of increasingly
complex and heterogeneous compute nodes. To exploit the performance of these systems …

[PDF][PDF] Research on Parallel Hierarchical Matrix Construction

Z Bai - 2023 - repository.kulib.kyoto-u.ac.jp
Abstract Hierarchical matrix (H-matrix), which is widely used in boundary element method
(BEM) and N-body simulations, is an approximated form to represent N× N correlations of N …

Performance and scalability study of FMM kernels on novel multi-and many-core architectures

A Rey, FD Igual, M Prieto-Matías, JF Prins - Procedia Computer Science, 2017 - Elsevier
We provide efficient implementations of common Fast Multipole Method (FMM) tasks for
modern multi-core (Intel Xeon Haswell), many-core (Intel Xeon Phi Knights Landing) and …

[PDF][PDF] Adaptation de codes industriels de simulation en Calcul Haute Performance aux architectures modernes de supercalculateurs

N Möller - 2019 - researchgate.net
Computational science is at the core of many scientific domains as much in academic
research than in industrial research and development. Computer simulations are used in a …

Industrial Code Modernization of High Performance Computing simulations on modern supercomputer architectures

N Möller - 2019 - theses.hal.science
For many years, the stability of the architecture paradigm has facilitated the performance
portability of large HPC codes from one generation of supercomputers to another. The …