Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544403
A. Jallad, L. B. Mohammad
Efficient software is required in order to make optimum utilization of the full-scale features of the rapidly developing MPSoC hardware. Embedded MPSoCs are characterized by their heterogeneity and resource limitation. In addition, embedded MPSoCs systems would be required to deal with a complex set of tasks whose resource requirements cannot be easily determined statically at design time. Middleware is a particularly important software component of embedded MPSoC that enables the system to overcome heterogeneity and manage resources dynamically at run-time. An important part of the middleware is the run-time system manager responsible for the distributed resource manager of the MPSoC-based system. System management at run-time is important for many modern embedded systems because the tasks performed by the system vary over time. In this paper we explore design trade-offs for middleware systems targeting multi-core embedded systems. The experimental platform implemented for the project is presented and future work is highlighted.
{"title":"Comparative analysis of middleware for multi-processor system-on-chip (MPSoC)","authors":"A. Jallad, L. B. Mohammad","doi":"10.1109/INNOVATIONS.2013.6544403","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544403","url":null,"abstract":"Efficient software is required in order to make optimum utilization of the full-scale features of the rapidly developing MPSoC hardware. Embedded MPSoCs are characterized by their heterogeneity and resource limitation. In addition, embedded MPSoCs systems would be required to deal with a complex set of tasks whose resource requirements cannot be easily determined statically at design time. Middleware is a particularly important software component of embedded MPSoC that enables the system to overcome heterogeneity and manage resources dynamically at run-time. An important part of the middleware is the run-time system manager responsible for the distributed resource manager of the MPSoC-based system. System management at run-time is important for many modern embedded systems because the tasks performed by the system vary over time. In this paper we explore design trade-offs for middleware systems targeting multi-core embedded systems. The experimental platform implemented for the project is presented and future work is highlighted.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133522810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544407
M. Al-khaleel, M. Gander, A. Ruehli
Today, parallel processing is necessary for the solution of large systems of ordinary differential equations (ODEs) as they are obtained from large electronic circuits or from discretizing partial differential equations (PDEs). Using a fine mesh in the discretization of these problems also leads to large compute times and large storage requirements. The waveform relaxation (WR) technique, which is ideally suited for the use of multiple processors for problems with multiple time scales has been used to solve such problems on parallel processors for such large systems of ODEs. However, applying the so-called classical WR techniques to strongly coupled systems leads to non-uniform slow convergence over a window in time for which the equations are integrated. In this paper, we present a so-called optimizedWR algorithm applied to transmission line circuit problems based on the longitudinal partitioning into segments. This greatly improves the convergence for strongly coupled RLCG transmission line (TL) type circuits. The method can be applied to other similar circuits. The method is based on optimal parameters that lead to the optimal convergence of the iterations. Here, we present a practical optimized WR algorithm which is easy to use and is computationally inexpensive.
{"title":"Optimized waveform relaxation solution of RLCG transmission line type circuits","authors":"M. Al-khaleel, M. Gander, A. Ruehli","doi":"10.1109/INNOVATIONS.2013.6544407","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544407","url":null,"abstract":"Today, parallel processing is necessary for the solution of large systems of ordinary differential equations (ODEs) as they are obtained from large electronic circuits or from discretizing partial differential equations (PDEs). Using a fine mesh in the discretization of these problems also leads to large compute times and large storage requirements. The waveform relaxation (WR) technique, which is ideally suited for the use of multiple processors for problems with multiple time scales has been used to solve such problems on parallel processors for such large systems of ODEs. However, applying the so-called classical WR techniques to strongly coupled systems leads to non-uniform slow convergence over a window in time for which the equations are integrated. In this paper, we present a so-called optimizedWR algorithm applied to transmission line circuit problems based on the longitudinal partitioning into segments. This greatly improves the convergence for strongly coupled RLCG transmission line (TL) type circuits. The method can be applied to other similar circuits. The method is based on optimal parameters that lead to the optimal convergence of the iterations. Here, we present a practical optimized WR algorithm which is easy to use and is computationally inexpensive.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114218565","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544396
S. Ahmed, M. Pasquier, G. Qadah
The problem of extracting sentiments from text is a very complex task, in particular due to the significant amount of Natural Language Processing (NLP) required. This task becomes even more difficult when dealing with morphologically rich languages such as Modern Standard Arabic (MSA) and when processing brief, noisy texts such as “tweets” or “Facebook statuses”. This paper highlights key issues researchers are facing and innovative approaches that have been developed when performing subjectivity and sentiment analysis (SSA) on Arabic text in general and Arabic social media text in particular. A preprocessing phase to sentiment analysis is proposed and shown to noticeably improve the results of sentiment extraction from Arabic social media data.
{"title":"Key issues in conducting sentiment analysis on Arabic social media text","authors":"S. Ahmed, M. Pasquier, G. Qadah","doi":"10.1109/INNOVATIONS.2013.6544396","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544396","url":null,"abstract":"The problem of extracting sentiments from text is a very complex task, in particular due to the significant amount of Natural Language Processing (NLP) required. This task becomes even more difficult when dealing with morphologically rich languages such as Modern Standard Arabic (MSA) and when processing brief, noisy texts such as “tweets” or “Facebook statuses”. This paper highlights key issues researchers are facing and innovative approaches that have been developed when performing subjectivity and sentiment analysis (SSA) on Arabic text in general and Arabic social media text in particular. A preprocessing phase to sentiment analysis is proposed and shown to noticeably improve the results of sentiment extraction from Arabic social media data.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121434659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544402
A. Al-Abdullatif, Hadeel Al-Negheimish, Latifah Al-Mofeez, Nuha Al-Khalifa, Lama Al-Andas
This paper describes the research conducted to design the Arabic Brain Communicator (ABC), which is a brain-controlled typing system designed to facilitate communication for people with severe motor disabilities in Arabic. A user centered design was adopted; it included empirical investigations and meetings with Subject-Matter Experts and possible users. Activities conducted in the analysis and design of the system are discussed.
{"title":"Mind-controlled augmentative and alternative communication for people with severe motor disabilities","authors":"A. Al-Abdullatif, Hadeel Al-Negheimish, Latifah Al-Mofeez, Nuha Al-Khalifa, Lama Al-Andas","doi":"10.1109/INNOVATIONS.2013.6544402","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544402","url":null,"abstract":"This paper describes the research conducted to design the Arabic Brain Communicator (ABC), which is a brain-controlled typing system designed to facilitate communication for people with severe motor disabilities in Arabic. A user centered design was adopted; it included empirical investigations and meetings with Subject-Matter Experts and possible users. Activities conducted in the analysis and design of the system are discussed.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127252665","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544387
B. Mokhtar, M. Eltoweissy, H. El-Sayed
The advent of the Internet of things will bring to bear an explosion in the number of interconnected heterogeneous objects as well as the diverse resources and services they may offer. A fundamental goal is to ensure the availability of resources and services to communicating objects ubiquitously, resiliently, on-demand and at low cost while satisfying users' QoS requirements. We hypothesize that to achieve this goal; there is a need to build capabilities for smarter networking to harvest the currently elusive rich semantics that emerge in interactions. In this paper, we propose the concept and primary architecture of a network “memory” (or NetMem) to support smarter data-driven network operations as a foundational component of next generation networks. Guided by the fact that networking activities exhibit spatiotemporal data patterns, we design NetMem to mimic functionalities of the human memory. NetMem provides capabilities for semantics management through integrating data virtualization, cloud-like scalable storage, associative rule learning and predictive analytics. NetMem provides associative access to data patterns and relevant derived semantics to enable enhancements in decision making, QoS guarantees and utilization of resources, early anomaly detection, and more accurate behavior prediction. We evaluate NetMem using simulation. Preliminary results demonstrate the positive impact of NetMem on various network management operations.
{"title":"Network “memory” system for enhanced network services","authors":"B. Mokhtar, M. Eltoweissy, H. El-Sayed","doi":"10.1109/INNOVATIONS.2013.6544387","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544387","url":null,"abstract":"The advent of the Internet of things will bring to bear an explosion in the number of interconnected heterogeneous objects as well as the diverse resources and services they may offer. A fundamental goal is to ensure the availability of resources and services to communicating objects ubiquitously, resiliently, on-demand and at low cost while satisfying users' QoS requirements. We hypothesize that to achieve this goal; there is a need to build capabilities for smarter networking to harvest the currently elusive rich semantics that emerge in interactions. In this paper, we propose the concept and primary architecture of a network “memory” (or NetMem) to support smarter data-driven network operations as a foundational component of next generation networks. Guided by the fact that networking activities exhibit spatiotemporal data patterns, we design NetMem to mimic functionalities of the human memory. NetMem provides capabilities for semantics management through integrating data virtualization, cloud-like scalable storage, associative rule learning and predictive analytics. NetMem provides associative access to data patterns and relevant derived semantics to enable enhancements in decision making, QoS guarantees and utilization of resources, early anomaly detection, and more accurate behavior prediction. We evaluate NetMem using simulation. Preliminary results demonstrate the positive impact of NetMem on various network management operations.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116777832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544422
A. More, S. Tapaswi
Our contribution through this paper is to provide a novel approach for securing hosts inside the home network using predictive attack detection and honeypots. We have given a unique method to secure hosts inside the home network. After securing these hosts we redirect attacker traffic to honeypots for further analysis. Our system also offers techniques to roam an attacker over multiple installed honeypots. This in turn helps to load share the attack traffic between honeypots and to record maximum number of attacks. This method also helps to address key problem of determining honeypot location for maximum attack exposure.
{"title":"A software router based predictive honeypot roaming scheme for network security and attack analysis","authors":"A. More, S. Tapaswi","doi":"10.1109/INNOVATIONS.2013.6544422","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544422","url":null,"abstract":"Our contribution through this paper is to provide a novel approach for securing hosts inside the home network using predictive attack detection and honeypots. We have given a unique method to secure hosts inside the home network. After securing these hosts we redirect attacker traffic to honeypots for further analysis. Our system also offers techniques to roam an attacker over multiple installed honeypots. This in turn helps to load share the attack traffic between honeypots and to record maximum number of attacks. This method also helps to address key problem of determining honeypot location for maximum attack exposure.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"119 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127265468","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544418
Amany M. Abu Safia, Z. Al Aghbari, I. Kamel
An important application of Wireless Sensor Networks (WSN) is to detect phenomena from the data gathered from the sensors. Traditional phenomena detection techniques are not directly applicable to WSN due to the nature of sensor data and specific requirements and limitations of the WSNs. The most critical requirement in WSN applications is extending the networks' life time by reducing the energy consumption while maintaining connectivity and accuracy. To meet this requirement, many approaches has been developed taking into consideration a static and dynamic WSNs. In this paper, we propose an algorithm to elect group heads that are used to detect phenomena in data gathered from mobile sensor nodes while minimizing the energy expenditure. The algorithm elects group heads based on the collected information of the global reported phenomena. The algorithm makes use of the phenomena location information to effectively select group heads. The paper shows the feasibility of the proposed algorithm by comparing it with a naïve approach for electing group heads in terms of energy consumption.
{"title":"A study on group head election for detecting phenomena in mobile WSN","authors":"Amany M. Abu Safia, Z. Al Aghbari, I. Kamel","doi":"10.1109/INNOVATIONS.2013.6544418","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544418","url":null,"abstract":"An important application of Wireless Sensor Networks (WSN) is to detect phenomena from the data gathered from the sensors. Traditional phenomena detection techniques are not directly applicable to WSN due to the nature of sensor data and specific requirements and limitations of the WSNs. The most critical requirement in WSN applications is extending the networks' life time by reducing the energy consumption while maintaining connectivity and accuracy. To meet this requirement, many approaches has been developed taking into consideration a static and dynamic WSNs. In this paper, we propose an algorithm to elect group heads that are used to detect phenomena in data gathered from mobile sensor nodes while minimizing the energy expenditure. The algorithm elects group heads based on the collected information of the global reported phenomena. The algorithm makes use of the phenomena location information to effectively select group heads. The paper shows the feasibility of the proposed algorithm by comparing it with a naïve approach for electing group heads in terms of energy consumption.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127442332","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544401
Mazen Ezzeddine, Hussein Karaki, Haitham Akkary
We propose a ubiquitous computing platform that supports multiple industrial instruction set architecture (ISA) standards via hardware assisted ISA virtualization. The proposed architecture features the following innovations: 1) It employs an efficient hardware assisted dynamic binary translation approach, 2) It implements a native semantically rich instruction set architecture that achieves when translating user level code one-to-one mappings for RISC ISAs such as ARM and a minimal one-to-multiple mappings for CISC ISAs such as Intel X86, and 3) It tackles, without compromising performance or energy efficiency and without requiring recompilation of software or rewriting of compilers, the issue of software portability across different ISA platforms through unhosted hardware assisted ISA virtualization. After introducing a metric to compare the semantic difference between an emulated virtual ISA and the native platform ISA, we present performance results as well as measurements of the effectiveness of our proposed platform when running ARM and X86 binaries.
{"title":"Ubiquitous computing platform via hardware assisted ISA virtualization","authors":"Mazen Ezzeddine, Hussein Karaki, Haitham Akkary","doi":"10.1109/INNOVATIONS.2013.6544401","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544401","url":null,"abstract":"We propose a ubiquitous computing platform that supports multiple industrial instruction set architecture (ISA) standards via hardware assisted ISA virtualization. The proposed architecture features the following innovations: 1) It employs an efficient hardware assisted dynamic binary translation approach, 2) It implements a native semantically rich instruction set architecture that achieves when translating user level code one-to-one mappings for RISC ISAs such as ARM and a minimal one-to-multiple mappings for CISC ISAs such as Intel X86, and 3) It tackles, without compromising performance or energy efficiency and without requiring recompilation of software or rewriting of compilers, the issue of software portability across different ISA platforms through unhosted hardware assisted ISA virtualization. After introducing a metric to compare the semantic difference between an emulated virtual ISA and the native platform ISA, we present performance results as well as measurements of the effectiveness of our proposed platform when running ARM and X86 binaries.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114242444","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544392
Xiaohu Chen, Qilin Fan, H. Yin
One of the defining characters of Information-Centric Networking(ICN) is in-network caching which enables content retrieval shift its emphasis from the exact content providers to the content that customers mostly care about. Undoubtedly, how to cache effectively is of primary concern in ICN. However, caching strategies such as Least Recently Used(LRU) are not tailored for the characteristics of content retrieval in ICN, thus compromising the capability of in-network storage. In this paper, we propose a new caching strategy called LUV-Path, in which all the routers along the delivery path will implicitly cooperate in deciding whether to cache the content and each content cached will be assigned a value that combines Least Unified Value(LUV) with router's distance from provider to reflect its relative importance. We evaluate LUV-Path with cache algorithms including LRU, FIFO under various network topology such as highly structured string and tree topologies and irregularly structured Abilene network. Our results suggest that LUV-Path significantly outperforms other caching strategies in reducing customer delay and network traffic as well as alleviating provider pressure across various network topologies with different structural property consistently.
{"title":"Caching in Information-Centric Networking: From a content delivery path perspective","authors":"Xiaohu Chen, Qilin Fan, H. Yin","doi":"10.1109/INNOVATIONS.2013.6544392","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544392","url":null,"abstract":"One of the defining characters of Information-Centric Networking(ICN) is in-network caching which enables content retrieval shift its emphasis from the exact content providers to the content that customers mostly care about. Undoubtedly, how to cache effectively is of primary concern in ICN. However, caching strategies such as Least Recently Used(LRU) are not tailored for the characteristics of content retrieval in ICN, thus compromising the capability of in-network storage. In this paper, we propose a new caching strategy called LUV-Path, in which all the routers along the delivery path will implicitly cooperate in deciding whether to cache the content and each content cached will be assigned a value that combines Least Unified Value(LUV) with router's distance from provider to reflect its relative importance. We evaluate LUV-Path with cache algorithms including LRU, FIFO under various network topology such as highly structured string and tree topologies and irregularly structured Abilene network. Our results suggest that LUV-Path significantly outperforms other caching strategies in reducing customer delay and network traffic as well as alleviating provider pressure across various network topologies with different structural property consistently.","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131797765","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2013-03-17DOI: 10.1109/INNOVATIONS.2013.6544384
A. Asaduzzaman, Fadi N. Sibai, H. El-Sayed
The advancement of multicore systems demands applications with more threads. In order to facilitate this demand, parallel programming models such as message passing interface (MPI) are developed. By using such models, the execution time and the power consumption can be reduced significantly. However, the performance of MPI programming depends on the total number of threads and the number of processing cores in the system. In this work, we experimentally study the impact of Open MPI and POSIX Thread (Pthread) implementations on performance and power consumption of multicore systems. Data dependent (like heat conduction on 2D surface) and data independent (like matrix multiplication) applications are used with high performance hardware in the experiments. Simulation results suggest that both implementations of more threads running in a system with more cores have potential to reduce the execution time with negligible or little increase in total power consumption. It is observed that the performance of MPI implementation varies (due to the dynamic communication overhead among the processing cores).
{"title":"Performance and power comparisons of MPI Vs Pthread implementations on multicore systems","authors":"A. Asaduzzaman, Fadi N. Sibai, H. El-Sayed","doi":"10.1109/INNOVATIONS.2013.6544384","DOIUrl":"https://doi.org/10.1109/INNOVATIONS.2013.6544384","url":null,"abstract":"The advancement of multicore systems demands applications with more threads. In order to facilitate this demand, parallel programming models such as message passing interface (MPI) are developed. By using such models, the execution time and the power consumption can be reduced significantly. However, the performance of MPI programming depends on the total number of threads and the number of processing cores in the system. In this work, we experimentally study the impact of Open MPI and POSIX Thread (Pthread) implementations on performance and power consumption of multicore systems. Data dependent (like heat conduction on 2D surface) and data independent (like matrix multiplication) applications are used with high performance hardware in the experiments. Simulation results suggest that both implementations of more threads running in a system with more cores have potential to reduce the execution time with negligible or little increase in total power consumption. It is observed that the performance of MPI implementation varies (due to the dynamic communication overhead among the processing cores).","PeriodicalId":438270,"journal":{"name":"2013 9th International Conference on Innovations in Information Technology (IIT)","volume":"195 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133355738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：