parallel processor
Recently Published Documents


TOTAL DOCUMENTS

630
(FIVE YEARS 9)

H-INDEX

33
(FIVE YEARS 0)

2021 ◽  
Vol 3 (1) ◽  
pp. 78-84
Author(s):  
S. P. Striamets ◽  
◽  
Opotiak Yu. V. Opotiak Yu. V. ◽  

Descri­bed ele­ment ba­se of a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um, which can be ef­fecti­vely used in the cre­ati­on of sig­nal pro­ces­sing systems using pro­ces­sing al­go­rithms with de­ep pa­ral­le­li­za­ti­on, up to the bit le­vel. The analysis of ad­vanta­ges and di­sad­vanta­ges of the exis­ting ele­ment ba­se is car­ri­ed out and the new impro­ved struc­tu­re of a pro­ces­sor ele­ment of the ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um is of­fe­red. The pos­si­bi­lity of mul­tiple use of each ele­ment of the en­te­red da­ta, which pro­vi­des high com­pu­ta­ti­onal per­forman­ce of a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um, is one of the sig­ni­fi­cant ad­vanta­ges of the systo­lic appro­ach. Ot­her be­ne­fits, such as mat­rix sca­la­bi­lity, simpli­city, and re­gu­la­rity of da­ta flows, are re­ali­zed thro­ugh the use of com­pu­ting cells with the simplest pos­sible archi­tec­tu­re. The archi­tec­tu­re of the ad­vanced pro­ces­sor ele­ment for construc­ti­on of the ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um with the impro­ved cha­rac­te­ris­tics in which me­ans of control of com­pu­ting pro­cess in a com­pu­ting cell, the mec­ha­nism of dup­li­ca­ti­on of the chan­nel of in­put of ad­justing in­forma­ti­on and me­ans of the ac­ce­le­ra­ted chan­ge of the ope­ra­ti­on co­de is of­fe­red. A num­ber of construc­ti­ve pro­po­sals for impro­ving the exis­ting c pro­ces­sor ele­ments and cre­ating a new ele­ment ba­se of a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um are pro­po­sed. The mec­ha­nisms of tes­ting the mat­rix of a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um in or­der to de­tect da­ma­ged and non-functi­oning pro­ces­sor ele­ments in­si­de the mat­rix, the mec­ha­nism of ra­pid chan­ge of the ope­ra­ting co­de of in­di­vid­ual pro­ces­sor ele­ments in­si­de the mat­rix of a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um are descri­bed. The pos­si­bi­lity of per­forming control of the con­fi­gu­ra­ti­on co­de in the pro­cess of wri­ting to the mat­rix and control of the cor­rectness of ope­ra­ti­ons du­ring the ope­ra­ti­on of the mat­rix of a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um is descri­bed. The de­sig­ned pro­ces­sor ele­ment of a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um further compri­ses a di­ag­nostic unit, which gre­atly simpli­fi­es the pro­cess of tes­ting the mat­rix to de­tect ino­pe­rab­le com­pu­ting cells. The mec­ha­nism of incre­asing the pro­duc­ti­vity of the com­pu­ting fi­eld and ways to incre­ase the sur­vi­va­bi­lity of di­gi­tal sig­nal pro­ces­sing systems ba­sed on a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um with re­con­fi­gu­ra­ti­on of the sig­nal pro­ces­sing al­go­rithm ta­king in­to ac­co­unt da­ma­ged, bro­ken or fa­iled pro­ces­sor ele­ments are descri­bed. The mat­rix of a pa­ral­lel spe­ci­ali­zed pro­ces­sor, bu­ilt on the ba­sis of a ho­mo­ge­ne­ous com­pu­ta­ti­onal me­di­um, which con­sists of 720 pro­ces­sor ele­ments, is descri­bed. Its functi­onal units and prin­ciple of ope­ra­ti­on are descri­bed.


2020 ◽  
Vol 25 (2) ◽  
pp. 63-74
Author(s):  
Mohanad Abdulhamid ◽  
Mwalimu Muthami

AbstractAs data centers have increased in size, there has been a need to create clusters out of cheaper, more affordable commodity parts that can easily be replaced upon failure, and that create more affordable data centers overall. However, such large clusters are still outside of feasibility for individuals and small businesses. It is a worthwhile exercise to see if much smaller clusters could be created for such applications, and to compare their performance / price measure to that of the previous traditional data centers. For this paper, such a cluster is created using Raspberry Pis which are small-sized, single-board computers. A data sharing model is built in Python using message passing interface (MPI) that ran on the cluster of the four Raspberry Pis. So as to evaluate the performance of the system, some greedy algorithms are created. During the implementation process, previously unknown skills, including how to create a cluster, programming the infrastructure are learnt.


2020 ◽  
pp. 22-39
Author(s):  
Leonid Eisymont ◽  
◽  
Anatoly Nikitin ◽  
Dmitriy Bikonov ◽  
Alexey Brazhkin ◽  
...  

2019 ◽  
Vol 34 (6) ◽  
pp. 1723-1726
Author(s):  
Ranko Davidović

Cognitive (intellectual) processes are, on the basis of Plato's psychological trinoma (cognition - affection - connection), the most intensively studied processes in psychology. Intelligence, as the center of these processes, despite serious efforts, remained insufficiently known. The proof of this statement is the existence of numerous conceptions about the types of cognitive functioning.Cognitive abilities were measured in the first lessons,the presence of psychologists, subject teachers and authors of this research. All of the administered tests are essentially speed tests, so respondents do not have time to contact and communicate with other respondents. A battery of KOG-3 tests was used to assess cognitive ability (Wolf, Momirović, and Džamonja 1992).The KOG-3 battery is the minimum battery for evaluating the performance of a perceptual, serial, and parallel processor. The version of that battery was used to achieve the basic goal, ie. to determine the level of general cognitive ability. The battery consists of three tests:Image Comparison Test (IT1), the basic version was designed by Thurstone. The test is designed as a test of the general perceptual factor, which is in fact a synthesis of primary factors of perceptual identification, perceptual analysis and perceptual structuring. The completiontime is limited to 4 minutes.Synonym test (AL4), the basic version was designed by Wels. The test is constructed as a test of verbal comprehension. The completiontime is limited to 2 minutes andVisual Specialization Test(S1), the basic version was designed by Reuchlin and Valin. The test is designed as a classical multiple choice special test. The completiontime is limited to 8 minutes.The aim of this research is to identify and determine the specificity of the cognitive abilities of latent dimensions, as well as their differences, between female and male students in relation to gender.On the basis of the formulated problem, the subject matter and the established general and partial goals, the basic alternative hypothesis was put forward:H2. The structure of female and male students' cognitive abilities will provide unambiguous evidence thatis of hierarchical type, with thegeneral cognitive factor at the top, below which are the three primary cognitive factors, defined as:efficiency of perceptual processor (perceptual reasoning), IT-1; parallel processor efficiency (ability to spot, relate, and correlate), SI-1, and serial processor efficiency (symbolic reasoning). AL-4.H5 - No statistically significant differences in the analyzed cognitive abilities are expected between students of different gender and age.


The efficiency of parallel processors is achieved for the purpose of quick computing is mainly depending on the scheduling of activity. The most important factor for scheduling of activities are depend on the waiting time which is directly influence the computation time of overall activity. Minimizing the Variance of Waiting Time otherwise known as Waiting Time Variance (WTV) is one of the metrics of Quality of Services (QoS) which enhance the efficiency of activity scheduling. Allocate the activity from a set of activity pool and schedule them for each identical parallel processor for execution in a large scale by minimizing WTV is the main focusing area of this paper. In case of large scale computing activities are complex in nature. A prior knowledge of each activity must be known before the preparation of activity scheduling for efficient and rapid computing. A snake walks style of activity distribution among the parallel processor is presented in this paper for minimization problem. The minimization of WTV is measured with the help of three heuristic intend methods named as RSS, VS and BS. The results of the experiment are compared with current conspires and demonstrate the new snake style conspire is presenting the best practices for proven conspires and challenges in a wide range of activity. The algorithm's predictable findings appear as illustrated with graph.


2018 ◽  
Vol 53 (2) ◽  
pp. 1-14 ◽  
Author(s):  
Daichi Fujiki ◽  
Scott Mahlke ◽  
Reetuparna Das

Sign in / Sign up

Export Citation Format

Share Document