Expertise

RunTime Computing assists customers with High Performance Computing (HPC) across a wide range of environments. We facilitate HPC on real-time and embedded platforms and systems, traditional cluster computing, and cloud computing environments.

RunTime provides shrink-wrap products for two key industry standards: the MPI Standard for Message Passing and the VSIPL standard for signal and image processing. Our MPI/Pro and VSI/Pro products offer high-performance, portable interfaces for embedded computing applications in long-life defense and mil/aero programs.

Our expertise in developing VSIPL and MPI applications enables us to advise customers on how to use these application programming interfaces for optimal results. We have developed and optimized applications for customers utilizing MPI, VSIPL, and other reusable libraries such as PETSc used in specific kinds of numerical simulations.

Underlying our ability to deliver software, advice, optimized code, and improved performance for existing customer code is a wide variety of deep, broad HPC expertise.

VSIPL

  • Vector Signal and Image Processing Standard (Object Management Group Standard)
  • RunTime involvement as co-chair of the VSIPL standards body
  • Fielded high quality implementation (VSI/Pro) used in many defense programs over 13 years
  • Leading expert in the use of the VSIPL programming model for real applications
  • Ability to train, customize, and extend our shrink-wrap products for specific defense purposes
  • Support key processor families used in defense programs: x86, PowerPC, and GPU, as well as emerging platforms
  • Support industry standard Linux and VxWorks for deploying customer applications that utilize VSI/Pro

MPI

  • Message Passing Interface Standard
  • Embedded MPI/Pro product supports the Version 2.0 standard, which is mature and aimed directly at high performance computing in embedded applications
  • Support key operating environments such as Linux, with ability to port to OSes and RTOSes
  • Current network support includes TCP/IP, InfiniBand, and RapidIO
  • Offer customized porting of MPI/Pro to new network interconnects on a consulting basis for customers developing new platforms

Cluster Computing

  • Assist with architecting and deploying small, medium, and large clusters
  • Assist in utilizing clusters, and configuring and managing interconnect and storage resources
  • Knowledge of how and where to use open source and commercial add-ons in a cluster environment
  • Understand issues associated with cluster computing in a private or public cloud and the types of applications that perform well in the cloud
  • Experienced with StarCluster

High Speed Networking

  • Experts in high speed networking including InfiniBand, RapidIO, and high-speed Ethernet (10Gb and 40Gb Ethernet)
  • Expertise in porting and tuning the OFED software stack for InfiniBand and other Remote Direct Memory Access (RDMA) transports
  • Expertise in Sandia Labs Portals standard porting and programming in the Portals APIs
  • Experience and expertise with InfiniBand Verbs programming

Cloud Computing

  • Help customers deploy Amazon EC2 and OpenStack/Crowbar clouds
  • Train customers on how to utilize cloud infrastructure and integrate it with their own in-house systems, for split cloud/in-house computing
  • Our expertise on HPC allows us to guide customers on how to provision instances in the cloud as a function of their intended uses
  • Show customers how and when to use public clouds vs. how and when to build private clouds
  • Show customers how to analyze economics of moving to a public cloud or building a private cloud