• Title/Summary/Keyword: memory latency

Search Result 361, Processing Time 0.028 seconds

Performance Evaluation and Analysis of Multiple Scenarios of Big Data Stream Computing on Storm Platform

  • Sun, Dawei;Yan, Hongbin;Gao, Shang;Zhou, Zhangbing
    • KSII Transactions on Internet and Information Systems (TIIS)
    • /
    • v.12 no.7
    • /
    • pp.2977-2997
    • /
    • 2018
  • In big data era, fresh data grows rapidly every day. More than 30,000 gigabytes of data are created every second and the rate is accelerating. Many organizations rely heavily on real time streaming, while big data stream computing helps them spot opportunities and risks from real time big data. Storm, one of the most common online stream computing platforms, has been used for big data stream computing, with response time ranging from milliseconds to sub-seconds. The performance of Storm plays a crucial role in different application scenarios, however, few studies were conducted to evaluate the performance of Storm. In this paper, we investigate the performance of Storm under different application scenarios. Our experimental results show that throughput and latency of Storm are greatly affected by the number of instances of each vertex in task topology, and the number of available resources in data center. The fault-tolerant mechanism of Storm works well in most big data stream computing environments. As a result, it is suggested that a dynamic topology, an elastic scheduling framework, and a memory based fault-tolerant mechanism are necessary for providing high throughput and low latency services on Storm platform.

Effect of Acupuncture at Oegwan(TE 5) on Auditory P300 (외관(外關) 자침이 Auditory P300에 미치는 영향)

  • Kim, Mi-Sun;Park, Hyun-Cheol;Yoo, Gyung;Kim, Lak-Hyung
    • Journal of Oriental Neuropsychiatry
    • /
    • v.17 no.1
    • /
    • pp.129-136
    • /
    • 2006
  • Objective : Audio-evoked potential is technique used for measuring brain functions such as memory, attention, and concentration. Although a number of studies on acupuncture for pain, there are few reports on the effect of acupuncture cognitive function. Methods : Fifteen healthy volunteers with no neurological condition took part of the study. $(23.27{\pm}2.09)$. In this study, we examined the changes in auditory P300 by acupuncture stimulation to 'Oegwan' on the left hand. Auditory P300 was recorded before, during and after acupuncture. The amplitude and latency of P300 at Fz, Cz, Pz point were calculated. Results : Auditory P300 amplitude decreased during and after acupuncture, but there was no statistical significance.(P< .05) Auditory P300 latency significantly decreased during and after acupuncture.(P< .05) Conclusions : The decrease of Auditory P300 latency time suggests that acupuncture at Oegwan has some effects on the cognitive function.

  • PDF

Effects of Ethylcholine Aziridinium, Scopolamine and Mor-phine on Learning Behaviors in Morris Sater Maze.

  • Lim, Dong-Koo;Wee, Sun-Mee;Ma, Young;Yi, Eunyoung
    • Archives of Pharmacal Research
    • /
    • v.18 no.5
    • /
    • pp.346-350
    • /
    • 1995
  • To assess the behavior of the various chemicals, such as ethylcholine axiridinium (AF54A), scopolamine and morphine, the chemicals were administered into eitheer rat or mice. And water maze tests were performed before and during drug administration. In AF64A-treated groups (3 nmol/each ventricle), the latencies to escape was significantly increased in both of the pretraining-and posttraining groups. In scopolamine-treatment (2mg/kg, sc) to the pretrained group, the latency to escape was significantly shortened after the acute administration of scopolamine. However in subscute treatment group with scopolamine, the latency to escape was significantly increased. In morphine-treated groups (10 mg/kg, ip), the subacute treatmment with morphine, the latency to escape was not ahcnged. The results indicate that each chemical induces the learning impariment. However the chrmical-induced learning impairment may have different characteristics upon the exposed chemical. Also the results suggest that both the motivation and the retrieval of memory might be impaired by AF64A.

  • PDF

Effects of Adenophora triphylla var. japonica Extract on Cognitive Function in Ovariectomized Mice (난소절제 마우스에서 잔대 추출물의 인지능 개선 효과)

  • Jang, Hwan Hee;Kim, Haeng Ran;Lee, Young Min
    • Journal of the East Asian Society of Dietary Life
    • /
    • v.27 no.4
    • /
    • pp.442-449
    • /
    • 2017
  • Adenophora triphylla var. japonica extract has been reported to have estrogenic activity and beneficial effects in postmenopausal women requiring estrogen replacement therapy. This study investigated the effects of A. triphylla extract on cognitive function in ovariectomized mice. Water extracts from A. triphylla were administrated to ovariectomized mice fed a high fat diet for 6 weeks. The step through latency time in the passive avoidance test was shortened in ovariectomized mice as compared with sham-operated mice. However, administration of A. triphylla extracts had a tendency to prolong the latency time in ovariectomized mice. Moreover, A. triphylla supplementation caused a significant decrease in mean escape latency in the water maze test. These results indicate that A. triphylla could be used to improve estrogen deficiency-related learning and memory impairments in postmenopausal women.

Resolving Cycle Extension Overhead Multimedia Data Retrieval

  • Won, Youjip;Cho, Kyungsun
    • Transactions on Control, Automation and Systems Engineering
    • /
    • v.4 no.2
    • /
    • pp.164-168
    • /
    • 2002
  • In this article, we present the novel approach of avoiding temporal insufficiency of data blocks, jitter, which occurs due to the commencement of new session. We propose to make the sufficient amount of data blocks available on memory such that the ongoing session can survive the cycle extension. This technique is called ″pre-buffering″. We examine two different approaches in pre-buffering: (i) loads all required data blocks prior to starting playback and (ii) incrementally accumulates the data blocks in each cycle. We develop an elaborate model to determine the appropriate amount of data blocks necessary to survive the cycle extension and to compute startup latency involved in loading these data blocks. The simulation result shows that limiting the disk bandwidth utilization to 60% can greatly improve the startup latency as well as the buffer requirement for individual streams.

An Effective Pre-refresh Mechanism for Embedded Web Browser of Mobile Handheld Devices

  • Li Huaqiang;Kim Young-Hak;Kim Tae-Hyung
    • Journal of Korea Multimedia Society
    • /
    • v.7 no.12
    • /
    • pp.1754-1764
    • /
    • 2004
  • Lately mobile handheld devices such as Personal Digital Assistant (PDA) and cellular phones are getting more popular for personal web surfing. However, today most mobile handheld devices have relatively poor web browsing capability due to their low performance so their users have to suffer longer communication latency than those of desktop Personal Computers (PCs). In this paper, we propose an effective pre-refresh mechanism for embedded web browser of mobile handheld devices to reduce this problem. The proposed mechanism uses the idle time to pre-refresh the expired web objects in an embedded web browser's cache memory. It increases the utilization of Central Processing Unit (CPU) power and network bandwidth during the idle time and consequently reduces the client's latency and web browsing cost. An experiment was done using a simulator designed by us to evaluate the efficacy of the proposed mechanism. The experiment result demonstrates that it has a good performance to make web surfing faster.

  • PDF

Enhanced Prediction Algorithm for Near-lossless Image Compression with Low Complexity and Low Latency

  • Son, Ji Deok;Song, Byung Cheol
    • IEIE Transactions on Smart Processing and Computing
    • /
    • v.5 no.2
    • /
    • pp.143-151
    • /
    • 2016
  • This paper presents new prediction methods to improve compression performance of the so-called near-lossless RGB-domain image coder, which is designed to effectively decrease the memory bandwidth of a system-on-chip (SoC) for image processing. First, variable block size (VBS)-based intra prediction is employed to eliminate spatial redundancy for the green (G) component of an input image on a pixel-line basis. Second, inter-color prediction (ICP) using spectral correlation is performed to predict the R and B components from the previously reconstructed G-component image. Experimental results show that the proposed algorithm improves coding efficiency by up to 30% compared with an existing algorithm for natural images, and improves coding efficiency with low computational cost by about 50% for computer graphics (CG) images.

Task Scheduling in Fog Computing - Classification, Review, Challenges and Future Directions

  • Alsadie, Deafallah
    • International Journal of Computer Science & Network Security
    • /
    • v.22 no.4
    • /
    • pp.89-100
    • /
    • 2022
  • With the advancement in the Internet of things Technology (IoT) cloud computing, billions of physical devices have been interconnected for sharing and collecting data in different applications. Despite many advancements, some latency - specific application in the real world is not feasible due to existing constraints of IoT devices and distance between cloud and IoT devices. In order to address issues of latency sensitive applications, fog computing has been developed that involves the availability of computing and storage resources at the edge of the network near the IoT devices. However, fog computing suffers from many limitations such as heterogeneity, storage capabilities, processing capability, memory limitations etc. Therefore, it requires an adequate task scheduling method for utilizing computing resources optimally at the fog layer. This work presents a comprehensive review of different task scheduling methods in fog computing. It analyses different task scheduling methods developed for a fog computing environment in multiple dimensions and compares them to highlight the advantages and disadvantages of methods. Finally, it presents promising research directions for fellow researchers in the fog computing environment.

Concurrent blockchain architecture with small node network (소규모 노드로 구성된 고속 병렬 블록체인 아키텍처)

  • Joi, YongJoon;Shin, DongMyung
    • Journal of Software Assessment and Valuation
    • /
    • v.17 no.2
    • /
    • pp.19-29
    • /
    • 2021
  • Blockchain technology fulfills the reliance requirement and is now entering a new stage of performance. However, the current blockchain technology has significant disadvantages in scalability and latency because of its architecture. Therefore, to adopt blockchain technology to real industry, we must overcome the performance issue by redesigning blockchain architecture. This paper introduces several element technologies and a novel blockchain architecture TPAC, that preserves blockchain's technical advantage but shows more stable and faster transaction processing performance and low latency.

A $CST^+$ Tree Index Structure for Range Search (범위 검색을 위한 $CST^+$ 트리 인덱스 구조)

  • Lee, Jae-Won;Kang, Dae-Hee;Lee, Sang-Goo
    • Journal of KIISE:Databases
    • /
    • v.35 no.1
    • /
    • pp.17-28
    • /
    • 2008
  • Recently, main memory access is a performance bottleneck for many computer applications. Cache memory is introduced in order to reduce memory access latency. However, it is possible for cache memory to reduce memory access latency, when desired data are located on cache. EST tree is proposed to solve this problem by improving T tree. However, when doing a range search, EST tree has to search unnecessary nodes. Therefore, this paper proposes $CST^+$ tree which has the merit of CST tree and is possible to do a range search by linking data nodes with linked lists. By experiments, we show that $CST^+$ is $4{\sim}10$ times as fast as CST and $CSB^+$. In addition, rebuilding an index Is an essential step for the database recovery from system failure. In this paper, we propose a fast tree index rebuilding algorithm called MaxPL. MaxPL has no node-split overhead and employs a parallelism for reading the data records and inserting the keys into the index. We show that MaxPL is $2{\sim}11$ times as fast as sequential insert and batch insert.