Integrex - touch the future

Cluster Software Environments

Platform HPC Suite Demo

platform computing logo

Integrex HPC Solutions are based on world-class HPC software from Platform Computing - the global leader in Cluster, Grid and Cloud management software. Platform products are used by leading global organisations in industries such as Electronics, Industrial manufacturing, Oil and Gas, Financial Services, Government and Scientific Research, for their superior performance, stability and scalability. Platform software is the best solution for your HPC Cluster, and is backed by global 24x7 support as standard.

Platform Computing's flagship product - LSF - is widely regarded as the best Scheduling and Workload Management solution available, and is used on many systems in excess of 40,000 CPU Cores, proving its performance and scalability. Now, Platform Computing offers LSF as part of a suite of software for both Workgroup and Enterprise class HPC systems. These solutions bundles provide the most comprehensive HPC Software stack available, and incorporate full Cluster Management and Provisioning, Commercial MPI stack (Formerly HP MPI), enterprise class workload management and scheduling, and adaptive multi-boot functionality.

Solutions for Workgroup Clusters (Up to 32 Nodes)

Platform HPC Workgroup Edition supports workgroup level clusters up to 32 nodes. It includes Platform Cluster Manager, that provides provisioning and management for Linux® clusters. It also includes Platform LSF Workgroup, a sophisticated workload scheduler for handling job submissions and managing the workload on the cluster. Platform MPI (formerly HP MPI) is the market's fastest and most scalable MPI implementation, and supports complex parallel workloads. A unified management web interface provides an easy to use portal for users to manage the cluster, and Platform Adaptive Cluster gives users the ability to make their clusters dynamic, allowing both Linux and Windows operating Systems to co-exist on the same hardware, and for the workload on the system to determine the profile of the cluster. So for example, if many more Linux jobs are being submitted compared to Windows based jobs, the cluster will provision more nodes to Linux in order to execute the workload.

Solutions for Enterprise Clusters (Above 32 Nodes)

Platform HPC Enterprise Edition uses the same components as the HPC Workgroup Edition, along with additional capabilities and functionality appropriate for Enterprise Class HPC environments. So, users still have leading cluster provisioning and management with platform Cluster Manager, class-leading MPI, full LSF for workload management, and Adaptive Cluster for multi-boot, multi-OS installations. It also includes the Platform Applications Centre Web Portal, which gives users a very simple interface to the cluster for their job submission and management, shielding them from the complexity of the HPC Cluster. HPC Enterprise Edition also features the Platform Real Time Monitoring tools, giving the administrator a powerful dashboard showing the health and performance of the cluster hardware, and the queues and jobs running on the system.

World Class Support

All Platform solutions are backed by class-leading support from the Platform Global support organization, which has locations around the globe, providing support on a 24 x 7 basis for any critical software issues.

As used by…

Platform software is running in many of the largest HPC datacenters in industry. Platform customers include the following organisations:

Hewlett Packard, Procter & Gamble, Fiat, Shell, Audi, Red Bull Technology, JPMorgan Chase, British Gas, Société Générale, MIT, Pratt & Whitney, Nvidia, Northrop Grumman, Hitachi, Nissan, Bombardier, Lockheed Martin, Petrobras, Ford, Rambus, Motorola, AstraZeneca, StatoilHydro, Samsung, BAE Systems, Honda, Fujitsu, Kimberly Clark, DuPont, SAS, Airbus, CERN, Sanger Centre, Canon, Toyota, Toshiba, Eli Lilly, Prudential, Sony and many other leading organisations.

Compilers

Compilers can be provided from INTEL, PGI, or Pathscale. Open source GCC compilers will be installed on the system as standard.

HPC Libraries

The following HPC libraries will be installed.

  • FFTW
  • ScaLAPACK / LAPACK
  • Blas
  • Atlas
  • MKL
  • ACML
  • Goto Blas

Power Management

All servers are IPMI enabled to support remote power control. In addition intelligent power supplies can be supplied to enable power control of nodes over the Ethernet network.

chart ui graphs white dwarf metalloid