The problem of misallocation of cache lines in processors is studied based on the specific features of memory organization in multi-core and multi-processor systems of distributed systems.Incorrect allocation of cache lines leads to sequential execution of tasks that should be executed in parallel on several cores.It is difficult to determine the problem, and its presence can cause a sharp decrease in the performance of the entire system.In many seismic studies, scientists' efforts are aimed at finding reliable signs of seismic hazard.Abrupt changes in one or another parameter are called abrupt effects, emissions or anomalies, through which the principle of prediction can be implemented -this means predicting the place, strength and time of a future seismic event.The use of distributed systems in the implementation of predictions with the help of machine learning, digital processing of large volumes of digital signals, ensures fast and high-quality determination of results.To increase the speed, it is necessary to develop parallel algorithms.Synchronization of common parallel streams is necessary to solve parallel computing problems and to ignore cache memory in parallel programming.This causes cache lines to be misallocated.Acceleration of parallel algorithms for digital processing of seismic signals and the dependence of the efficiency of using the system's computing resources on the number of parallel flows were analyzed, a parallel algorithm for determining the epicenter using cubic splines and a software tool was created based on this algorithm, and the proposed spline method was used to determine the characteristics of P and S waves.made it possible to determine the detection time 2.5 times faster than usual.
| Jurnal | ТАТУ хабарлари |
|---|---|
| Nashr sanasi | 2024-11-20 |
| DOI | 10.61663/243tuitmct5 |
DOI: 10.61663/243tuitmct5 · Maqolaning asl sahifasi · PDF
The h-index has emerged as a pivotal metric for evaluating the scientific impact and productivity of journals, authors, and institutions.In the context of scientific journals, the h-index is instrumental in determining…
Robustness to geometric distortions and distribution shifts remains a major challenge in deep learning.Standard data augmentation methods, such as random rotations and translations, improve generalization but often fail…
This article studies the use of modeling and algorithms based on 3D fractal geometry for early detection of internal diseases.Fractal models provide an opportunity to clarify the diagnostic process through accurate…
The purpose of the work in this article is to develop models and algorithms to ensure the effectiveness of monitoring the telecommunications network of information transmission in air navigation systems, allowing to…
Algorithm and software for calculating
This study investigates the use of Recursive Feature Elimination (RFE) for feature selection in multi-output classification, particularly with a medical dataset.The classification task requires predicting several…
The purpose of this article is to study the main approaches and algorithms for building recommendation services in online trading systems.Recommendation systems optimize the sales process by identifying interesting and…
The paper presents an overview of methods and technologies used to develop mobile applications using augmented reality (AR).The main focus is on spatial methods that allow for the effective integration of virtual…