A Method for Characterizing High Performance Computing Applications

Abstract

The invention provides a method for characterizing application software operation characteristics of various industries in the field of high performance computing. The method comprehensively examines the load pressure of application programs in five links of inputting, storage, processing, transmission and outputting, and sequentially divides the applications into four large types of compute-intensive type, memory constraint type, I / O intensive type and network intensive type. Through the quantitative representation of the four aspects, the resource demands of the applications on the aspects of CPU occupation, memory capacity, memory throughput, inputting / outputting and network data exchange are fully represented to maximally reflect the operation characteristics of the application software. The method provided by the invention is simple, practical, reliable and effective, and can directly reflect the quantity demand of one application software on the high-performance hardware resource, so that the applications can be operated on a proper high-performance platform in order to maximally exert the performances of the application software. According to the characteristics, the performance bottleneck of the application software can be improved and broken through in purpose, and the application expandability is enhanced.

Description

technical field [0001] The present invention relates to the performance characterization of application software in the field of high-performance computing, and specifically relates to a method for most reasonably reflecting the demand for computing resources of large-scale application software by monitoring and extracting appropriate high-performance parameters. Background technique [0002] With the continuous progress of human society and the development of science and technology, people not only have a wider understanding of the natural world, but also have an increasingly urgent need to explore the unknown world. This has led to a sharp increase in the amount of information and data held by humans, and at the same time, these massive information and data need to be analyzed and processed in a timely manner. For example, a large astronomical radio telescope array can generate more than 100GB of cosmic microwave data in one second, and these data need to be analyzed in ti...

AI Technical Summary

Problems solved by technology

However, after more than 50 years of development, the current supercomputer architecture has undergone earth-shaking changes. It often has hundreds of thousands or even millions of CPU cores, and a considerable number of supercomputers use hybrid heterogeneous (CPU+GPU) cores. / MIC, etc.) architecture, which makes the early physical models and mathematical algorithms incapable and incompetent

This is the main reason for the inefficiency and poor scalability of most application software today.

[0004] To solve these current problems, on the one hand, we should vigorously research and develop new physical models and mathematical algorithms that match the architecture of today's supercomputers. This is the ultimate means to break through existing bottlenecks, but this is an extremely Difficult subjects, it is impossible to see results in a short period of time and achieve large-scale applications; on the other hand, we should start to study the massive application software that has been inherited, reasonably characterize their operating characteristics, and find out their performance Bottleneck, maximize the performance of these applications on the existing platform, in addition to provide a strong basis for the improvement and breakthrough of application performance

Images