Creation of Supercomputer

High Performance Computing Systems, HPC or supercomputers is optimum solution to solve certain tasks.

Among them:

• engineering structure designing, building designing;
• designing of cars, vessels and aircrafts;
• designing of electronic and semiconductive devices;
• computational fluid dynamics (CFD);
• modeling of atmosphere and World ocean, weather forecasts;
• geological surveys;
• various mathematical physics tasks;
• astrophysics and space research;
• molecular sciences, genetics;
• modeling in medicine, including ASPDC;
• development of drugs;
• parallel DBMS, electronic commerce and information systems;
• virtual reality and visualization, image processing;
• nuclear power energetics;
• military tasks;
• training complexes;
• on-board, integrated systems, real-time systems;
• financial analysis, stock machines, etc.

Open Technologies has been working on construction of supercomputers from 2000 and offers both classic solutions based on standard architecture servers and more effective hybrid supercomputers.

Historically the term "High Performance Computing" (supercomputer) was connected to something very expensive, energy-intensive, occupying a lot of space. Development of the industry leads to changes of approach to construction of supercomputers. They become efficient in terms of energy, space occupied in technological premise and financially.

Hybrid supercomputer based on graphic processing units is characterized by lower power consumption, significantly lower cost and compactness as compared to conventional ones. At this it allows reaching performance up to tens and hundreds of teraflops. This is due to the fact that task solution is ensured by processors initially intended for graphic computations.

To save design resources of potential customers Open Technologies specialists created a range of standard compact hybrid supercomputer configurations - cheap, highly reliable, balanced, and produced in the form of finished product. As distinct from unique solutions, using of standard configurations allows reducing costs of solution design and commissioning, preventing component incompatibility, minimizing costs of document preparation, reducing terms of supercomputer preparation for delivery to the customer and solving issues at the stage of technical support more quickly due to pre-developed procedures.

Standard hybrid supercomputer is easy to introduce (when choosing the appropriate product it is enough to follow the requirements on performance and certain task peculiarities). The solution of tasks of integration into infrastructure is minimized (the system is supplied as ready-to-use platform containing all components necessary for operation and control).

If the high performance computing complex is already used but its resources are insufficient to solve current tasks there is a need to upgrade the system. In this case conventional extensive approach to extension of computing capacities is possible. The most optimal way to upgrade the supercomputer is reorganization by adjustment to hybrid infrastructure. Application of specialized constructs for deployment of graphic processing units allows maximum transparency of transition from supercomputer of classic scheme to hybrid supercomputer preserving all its advantages. This is a mechanism to protect customer's investments and allow significant reduction of total cost of ownership of the high performance computing hardware.

Supercomputer operation specifics require special approach to organization of systems in charge of its long and correct operation. The main systems are the engineering infrastructure systems. Open Technologies implements complex projects on construction of engineering infrastructure complying with requirements of hybrid supercomputers operation and its introduction into the existing customer's infrastructure.

Successful construction and operation of supercomputer is ensured by Open Technologies advantages, such as:

• customized designing;
• special attention to information protection of the solution;
• possibility to construct solution of any difficulty and performance;
• execution of the whole range of works (service development and feasibility study of its introduction, designing of technological solution, hardware and software supplies, installation and construction works, commissioning, organization of required engineering infrastructure at the supercomputer installation site);
• in-house service center: advanced support of installed solution (including outsourcing);
• high competence in solutions of the world leading manufacturers of hardware and software confirmed by corresponding partner statuses of our company;
• offices in many Russian regions;
• execution of works on the basis of industry regulatory documents, GOST, all necessary licenses and certificates available;
• Competency Center and laboratory to simulate solutions where specialists receive trainings on programming and program code optimization for GPU computations and solve issues on adaptation of existing software to work with hybrid high performance platform.