M3E | Numerics & HPC
page-template-default,page,page-id-618,,qode-title-hidden,qode_grid_1300,qode-theme-ver-10.0,wpb-js-composer js-comp-ver-6.11.0,vc_responsive

Numerics & HPC

With the increasing availability of significant computational resources at accessible costs, the implementation of large numerical models and the management of big amount of data is becoming quite a common effort in the engineering practice. Several High Performance Computing centers offer the opportunity to scientists and engineers of using either parallel computers with thousands of computing cores or small clusters with tens of multi-core processors.

M3E gives you the opportunity to exploit the potential of your machines, or the most recent computational architectures, to save time and money, and solve problems at an unprecedented detail. M3E develops and maintains an extensive library of state-of-art numerical kernels for the accurate and efficient solution of basic operations frequently occurring in engineering applications. In particular, M3E provides customized routines on both sequential and parallel platforms that can easily be included in existing codes for sparse linear algebra operations, managing and processing a huge amount of data, generating computational grids or transferring data between different models or software.

The top level M3E expertise in numerical analysis problems ensures a continuous maintenance and update of the available libraries, including the latest innovations in the field and guaranteeing a highly professional service.

Main activities include:

  • optimization/parallelization of existing codes for specific computing platforms, such as small clusters or supercomputers
  • development and maintenance of linear solvers for sequential and parallel machines, that can be connected to existing scientific software
  • development of ad hoc tools for solving specific numerical problems (such as the solution of complex linear systems or the research for specific eigenvalues) and for managing complex computing grids