We take care of the complete installation of clusters, both the hardware and the set of necessary programs for its management, as well as the middleware that helps to obtain the maximum performance of the applications. Moreover, we make sure that the system is perfectly prepared for this purpose prior entering into production. For this intention, we produce a series of reports based on the results of different benchmarks, configure the batch scheduler according to customer’s needs, and offer customized training to present the new architecture to the end user and local system administrators.
HPCNow! works with the world’s leading suppliers of computer systems and networks to be able to offer customers a solution that suits their needs, both in terms of performance and price. This also allows the user not to be tied to a manufacturer, so that this last one can be changed in successive equipment expansions in the event that market circumstances have changed.
In the section “Solutions”, each specific solution that HPCNow! offers is specified in more detail.
HPCNow! service engineers make their advanced capabilities available to their customers when installing hardware in the Data Processing Center.
We perform the installation of computer clusters paying attention to all the details that make up the solution. We wire in a structured way, label and generate documentation on the installation thinking about its subsequent maintenance, as we know the inherent problems in the support of a poorly installed and wired system. Our experience in this field leads to saving in time and money in the future servicing of the equipment.
We perform the installation of the necessary software for the proper functioning of an HPC system, from the OS to the end user application. Also, we carry out the installation and optimization of scientific and engineering applications in any computing system. For this, we take special care when compiling them with the most efficient tools, math libraries, and compilation options. In this sense, we can improve the performance and scalability of any scientific software in a specific architecture by optimizing the application. In doing so, we achieve the maximum possible performance that a particular system can provide.
|Remote Visualization: TurboVNC, VirtualGL, Websocket, DCV, X2Go|
|HPC Portal: EnginFrame|
|Scientific & Engineering Applications: More than 100 references. Contact us for further information.|
|Parallel File System: BeeGFS, Lustre, GPFS, HDFS, Ceph|
|User Environment: User libraries, Modules, EasyBuild, Spack|
|Development Tools: Compilers: GNU, Intel, PGI, IBM XL compilers; Debuggers and profilers: V-Tune, DDT, GDB|
|Monitoring & Alerts Tools: Ganglia, Nagios, Icinga, Grafana, Elasticsearch|
|Batch Scheduler: Slurm, LSF, PBS, Torque, SGE|
|Cluster Manager: sNow!, Bright Computing, xCat, Rocks, OpenHPC|
|Virtualization: OpenNebula, OpenStack, VMware, Xen-Source|
|Containers: Singularity, Docker, Docker Swarm, LXD|
|Operating System: CentOS, SLES, RHEL, Ubuntu, Debian|
In the “Solutions” section, more details of each solution that HPCNow! is able to offer are specified.
We teach training courses that cover all aspects necessary for a successful administration and a proper use of HPC resources. Our courses are divided in various training modules that can be configured according to the client needs, who could arrange together with us a course a la carte.
The training courses can be online and/or in person, and a complete test virtual environment is provided, which can be used after course completion.