An Approach to Building Adaptive Object Accounting Systems Using Artificial Intelligence Methods
Year: 2024
Type: article
Abstract: The needs of practice constantly require improving the performance of computing systems. Forquite a long time, multiprocessor systems have been the main way to build ultra-high performance computingsystems. When creating such systems, many difficult problems arise. They are related to the need toparallelize the computing process in order to efficiently load the system processors, overcome conflictswhen several processors try to use the same system resource, reduce the impact of conflicts on systemperformance, etc. With microelectronics overcoming the milestone of a billion transistors on a siliconchip, a new paradigm of multicore processors has emerged. At the same time, the problem of the ratio ofmulticore and multithreading in modern computers arose. This is due to the dilemma of preference betweenthem. A multicore processor contains two or more electronic computing cores placed on a singlesemiconductor crystal. Each core of a multicore processor is a full-fledged microprocessor. Multicore isan obvious and traditional method of distributed solution of many complex tasks. But this cannot be saidabout multithreading, which relies on the use of very fast cache memory associated with the main memoryand serves to reduce the average access time to the main memory of the processor. The relative novelty ofmodern approaches to the construction of computing systems requires comparative experimental studiesof their capabilities. A promising and convenient mathematical object for these purposes is the distributedmultiplication of matrices of large dimensions. The article presents practical results of distributed multiplicationof square matrices with sizes from 300*300 to 2000*2000 and randomly generated values ofelements in the matrices in the range from -100 to +100. Based on the experimental data presented in thecorresponding tables and graphs, hyperbolic relations are obtained for the dependence of the matrix multiplicationtime on the number of virtual machines (cores) in the laptop used. Similar results were obtainedby multiplying square matrices on single-processor computers connected to a local network. Analyticalexpressions in this case also represent hyperbolic time dependencies. But the numerical values inthem significantly exceed those for the hyperbolic formula obtained for the laptop. Based on the resultsobtained, the conducted research allows us to conclude that the use of a single-processor computer connectedto a local network for multiplying matrices of large dimensions is inferior to the performance of alaptop. This is due to the significant time spent moving data over the local network.
Authors: V. I. Voloshchuk, A. Garyagdiyev, М. А. Kozlovskaya, Y. E. Melnik, A.N. Samoylov
Institution: Southern Federal University
Language: English
Subfield: Hardware and Architecture
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4406397495