What is high performance computing? What is high performance computing?

Source: Internet
Author: User
Keywords What we run
Tags business computer computer science computing computing clusters cost development high

High-performance Computing (highperformancecomputing) is a branch of computer science, which mainly refers to the research and development of high-performance computers from the aspects of architecture, parallel algorithms and software development. With the rapid development of computer technology, the computational speed of high-performance computers has been increasing, and its standards are also changing.

High-performance computing is simply the completion of certain types of technical workloads on 16 or more servers. It doesn't matter if the number is 8, 12 or 16 servers. Under our definition we assume that each server is running its own stand-alone operating system, and its associated input/output infrastructure is built on the COTS system. In short, we're talking about Linux high-performance computing clusters.

An information center with 20000 servers There is no doubt that molecular dynamics simulations are no problem, like a small engineering company running Computational Fluid dynamics (CFD) simulations in its computer room. The only limit to resolving workloads comes from the technical level. The next thing we want to discuss is what can be applied directly.

Measurement (Metrics)

Today, few people have discussed pure performance issues with Linux's high-performance computing clusters. So far we've come across more words: Performance (configured), performance per watt (Performance/watt), performance per square foot (performance/squarefoot) and performance-price ratio (performance/ Dollar) and so on, for the above mentioned 20000 servers of the dynamic molecular cluster, the reason is obvious. Running such a system is often limited by the server's energy consumption (watts) and volume (square feet). All two elements are included in the total cost of ownership (TCO). It is a great concern to achieve greater economic benefits in total cost of ownership (TCO).

Then, the scope of this issue is defined in terms of performance to help you understand the importance of performance, performance density, and total cost of ownership (TCO) in practice.

Definition of performance

Here we define performance as a calculation rate. For example, the workload completed daily, the speed of floating-point operations per second (FLOPs), and so on. What we're going to think about in the next discussion is the time to finish the work. The two are directly related to the speed =1/(Time/effort). Therefore, the performance is calculated according to the workload of the operation, by calculating its completion time to convert to the required speed.

Quantitative and qualitative

How can I optimize high-performance computing (HPC) performance? The question is easy to answer from a qualitative level, and the answer is a faster processor, more capacity of memory, and a better performance of the network and disk input/output subsystems. But when it comes to deciding whether or not to buy a linu cluster, the answer is not accurate enough.

How to quantitatively analyze the performance of Linux high Performance computing cluster. To this end we are going to introduce some of the quantitative models and techniques that can be very precise in the direction of the business decisions, but also very simple and practical. For example, these business decisions involve areas such as:

Buy---System component sourcing Guide for best performance or most economical performance

Configuring---Identification system and bottlenecks in application software

Plan to develop medium-term business plans---highlighting the relevance and limitations of performance

The Linux high-performance computing cluster model in our paper includes four main hardware components. (1) A compute node or server (2) performing a technical workload is used for cluster management, the main nodes (3) connected to each other, such as work control, and the now highly popular Gigabit Ethernet (GBE) (4) Global storage systems, are as easy to use as NFS files exported by the master node.

At present, the measurement standard of high-performance computer is mainly based on the calculation speed (especially the floating point speed). High-performance computer is the forefront of information technology in the field of security, to promote national defense science and Technology progress, promote the development of cutting-edge weapons has a direct role in the measurement of a country's comprehensive strength is one of the important signs. With the rapid development of information society, human beings have higher demands on the ability to deal with it, not only the high performance computers such as petroleum exploration, meteorological forecast, Aerospace Defense and scientific research, but also the demand for High-performance computing in the fields of finance, government informatization, Education, Enterprise, network game and so on.

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.