Rocks+HPC with CUDA Roll lets you build and deploy clusters that leverage NVIDIA GPUs for general purpose computing. By integrating the CUDA Roll from NVIDIA with Rocks+, users gain the benefits of rapid configuration, and reliable, predictable performance from their NVIDIA GPU cluster thanks to the Rocks+ Avalanche installer, database driven library, and central operator’s console.

Key Features of Rocks+HPC with CUDA Roll

  • Fast, easy provisioning and management of General Purpose GPU clusters
  • Access to complete library of commercial Rocks Rolls
  • Updates, patches and upgrade protection for all StackIQ Rolls
  • Toll-free phone and email support including support for installation and configuration of open source Rocks Rolls

     

Performance

  • Clusters built using the CUDA roll and NVIDIA GPUs consistently outperform similarly configured CPU-based clusters.

     

Availability

  • Available now from StackIQ as part of the Rocks+ HPC.  A free download (for up to 16-nodes) is available.

  • Related available Rolls:

    • Open Source: Kernel, Base, HPC, Web Server, Ganglia, etc.

    • StackIQ Rolls: Support Roll, OFED Roll, Intel Roll, PGI Roll, etc.

    • Workload Management Rolls: Moab Roll, PBS Pro Roll, SGE Roll, Torque Roll.

  • StackIQ Rocks+ is also available for Amazon EC2 Elastic Compute Cloud

Useful links: