Menu Close

Lab equipment

Pythagoras Cluster

The Applied Mathematics and Mathematical Modelling laboratory is supported by a High Performance Computing cluster named Pythagoras.

Pythagoras consists of 11 computing nodes of aggregate computation power of 336 logical cores (168 physical cores/22 processors) with 480GB total physical memory and one node of 24 logical cores (12 physical cores/2 processors) and 48GB ram for managing the computation nodes (front-end node). Each node has a secondary storage while both the computing nodes and the front-end node share a filesystem of 1.5TB through (NFS). This filesystem is expanded to 15TB through an external storage device. However, it could be expanded up to 30TB depending on the the applications that they will be deployed on the cluster.

The operating system environment of Pythagoras is the Rocks Cluster Distribution (version Sidewider) on the CentOS Linux distribution (6.9). The high throughput computing environment is provided by Condor and SGE schedulers which are the standard grid middleware on Rocks Clusters. The HPC is equipped with mathematical libraries (calculus, linear algebra and optimisation for linear and nonlinear problems, multiple precision arithmetic libraries etc). Most of the programming languages (fortran, C/C++, Python, R, Java) and the MPI programming paradigm are supported. In 2020, the installation of the distributed version of MATLAB (MATLAB DCS) for 8 workers (computing nodes) along with the parallel computation toolbox and linear and nonlinear optimization toolboxes is scheduled. Finally, it is to consider the installation of Hadoop programming environment.

Pythagoras wiki pages (Beta version)