SMACC-Labs

Computing Cluster

SMACC Computing Cluster is a distributed computing environment for heavy technical computation. The most commonly used analytics and simulation software are installed in the computing cluster, including both commercial and open source code programs. Task- or project-specific software can also be installed to the environment if needed. The computational methods help both large and small companies to streamline product development and develop technically advanced products. This is achieved by using virtual models in the design process, utilizing efficient technical computation, and reducing the number of prototypes required. Computing Cluster offers services related for both parallel computing and the computing technology used in the platform. Computing Cluster can be utilized eg. for challenging optimization tasks as well as extensive computing tasks that require large amounts of computing with different parameters. The services of SMACC Computing Cluster include also modeling and simulation expertise as well as the actual computing execution.

Competitive advantage

Computing Cluster provides a unique entity that combines expertise, computing services and experimental activities (validation and verification of models and calculation results). With the help of SMACC, companies can develop both their own products and their own operating processes to achieve a competitive advantage.

Benefits

Companies can benefit from the cluster environment, for example in the form of computing services to be purchased or by working with SMACC in research projects, whereby computing expertise and technologies can be transferred to companies. SMACC does not directly sell machine time from its Computing Cluster.

Applications

The computational cluster allows customers to efficiently calculate technical analyzes that are virtually impossible to perform on standard personal computers due to the size of models or the long time required for computing. The idea of a parallel computing environment is to divide the computation task into smaller interconnected sub-tasks, which are calculated at the same time on different computers and computing nodes. Typical examples include CFD/FEM-analyzes of very accurate 3D-models, transient simulations, and multi-physics analyzes. Also, the cluster provides good support for result post-processing and visualization. With parametric computing models, it is also possible to run large parameter ranges and optimization problems, and the results can be efficiently achieved with a fast schedule.

In the case of SMEs, for example, the behavior of a particular structure can be calculated with very different initial values. The results can be then used to form simple dependencies that characterize the behavior of a company’s products. This information can be utilized to support design phases or bidding process. The wide range of software available in the computing environment enables the calculation of different technologies very widely.

Areas of expertise

SMACC has a large number of experts who apply computational methods in their research work and can provide research and development services. These topics include:

  • Structural analysis and structural dynamics
  • Sound and vibration
  • Acoustics
  • Dynamics of machines and machine systems
  • Computational fluid dynamics (CFD)
  • Electrical machinery and electromagnetics
  • Coupling of the topics mentioned above, multiphysics
  • Parameter sequencing, sensitivity reviews, optimization tasks

Computing Environment

SMACC Computing Cluster is designed for large-scale parallel computing, that is, computing that uses multiple (up to dozens) computers at the same time. Each cluster computer includes up to dozens of computing cores. Computing Cluster is an ever-expanding computing environment. In early 2017 it consists of 92 computing machines, totaling more than 2,000 computing units. The machines are combined for computational data with a very high-speed InfiniBand network (40 Gbps version) in order to achieve efficient parallel computing. The computers themselves are a variety of M and R series Dell PowerEdge computers with typically 128-512 GB of memory (largest amount of memory in a single PC is PC 1.5 TB). The cluster has a NetApp disk system for recording computational data.