vSMP ClusterONE

Consolidate servers into a single system: Turn an HPC cluster of up to 32 servers into a single, easy-to-manage system, eliminating all cluster complexity and reducing management overhead.

vSMP ClusterONE turns an HPC cluster into a unified system running a single operating system. Instead of trying to ease cluster management, it eliminates the cluster altogether.

The software effectively creates a super-sized workstation and enables the leap from desktop computing to HPC-scale deployment without any special knowledge requirements.

With vSMP ClusterONE, a physical cluster of servers is transformed into a single large system with all the computing-power, memory and I/O capabilities of the underlying hardware, yet running a single instance of an operating system.

This effectively eliminates any human-resource overhead for cluster management, and any Linux system admin can manage the system effectively without knowing anything about system provisioning, cluster management, fabric management or parallel file systems. All aspects are available as standard Linux OS functions on the aggregated system.


vSMP ClusterONE simplifies cluster management into a standard Linux administration task by turning a cluster of physically separate server systems into a single logical computer, much like a super-workstation. If you can manage a Linux workstation, you can manage a system using vSMP ClusterONE.

Once deployed, users can work with the system interactively, just like with a workstation. Standard Linux operating system management tools work as-is; Linux volume management will serve a high-performance file system for scratch; and it’s not necessary to manage subnets or fabrics, or worry about multiple operating system instances.

Enhanced Capabilities
vSMP ClusterONE also provides enhanced capabilities over standard cluster systems by offering:

  • Ability to work in both batch and interactive modes.
  • A single OS — no need to ensure software is the same across many nodes.
  • Built-in, shared I/O, enabling local volumes like high-performance scratch, direct-attached, etc.
  • Flexibility to use high-performing InfiniBand fabric without any learning curve and without any drivers.


HPC provides an unfair advantage to organizations using it: They design better products while spending less. Organizations can simulate more cases and choose the solution that is more likely to work.

Traditionally, moving up from workstation-class systems to HPC environments has been a huge leap, requiring expensive, specially trained professionals for even small cluster deployments. With vSMP ClusterONE, there are no barriers to entry. Regardless of the number of server nodes, vSMP ClusterONE behaves just like a big workstation with many cores, lots of RAM and massive I/O — all shared.


  • Up to 32 dual-socket nodes, up to 6GB/core
  • See www.scalemp.com/spec for certified hardware and operating systems
  • Local storage on nodes recommended (for fast software-raid scratch across all drives)
  • Redhat, CentOS, or SUSE operating systems
  • vSMP Productivity Pack for auto-tuned and hassle-free cluster operation