Pub Date : 1994-05-02DOI: 10.1109/MPCS.1994.367079
K. M. George, J. Duan
In this paper we describe a new parallel computer architecture called Tr-machine. Tr-machine instruction set architecture combines reduction with traditional approaches. The architecture uses two types of processing units and a tree structured organized as a balanced tree. The leaf nodes are called C-nodes and they perform the computation. The internal nodes are called S-nodes. The S-nodes facilitate communication and keep track of different parallel computations.<>
{"title":"Tr-machine architecture","authors":"K. M. George, J. Duan","doi":"10.1109/MPCS.1994.367079","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367079","url":null,"abstract":"In this paper we describe a new parallel computer architecture called Tr-machine. Tr-machine instruction set architecture combines reduction with traditional approaches. The architecture uses two types of processing units and a tree structured organized as a balanced tree. The leaf nodes are called C-nodes and they perform the computation. The internal nodes are called S-nodes. The S-nodes facilitate communication and keep track of different parallel computations.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"30 1","pages":"191-195"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"73099459","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367034
H. Guyennet, B. Herrmann, Laurent Philippe, F. Spies
Multicomputers are often dedicated to specialized applications. However, they are increasingly based on general purpose microprocessors, so they can be used for general purpose applications if they provide some support to users, as load balancing. To better understand the behavior of load-balancing algorithms on multicomputers, we have developed a simulation environment using QNAP. The originality of our approach is that we take parallel computers constraints into account and evaluate the ability of load-balancing algorithms to behave as distributed schedulers. The paper gives the first results and comments on this simulation.<>
{"title":"A performance study of dynamic load balancing algorithms for multicomputers","authors":"H. Guyennet, B. Herrmann, Laurent Philippe, F. Spies","doi":"10.1109/MPCS.1994.367034","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367034","url":null,"abstract":"Multicomputers are often dedicated to specialized applications. However, they are increasingly based on general purpose microprocessors, so they can be used for general purpose applications if they provide some support to users, as load balancing. To better understand the behavior of load-balancing algorithms on multicomputers, we have developed a simulation environment using QNAP. The originality of our approach is that we take parallel computers constraints into account and evaluate the ability of load-balancing algorithms to behave as distributed schedulers. The paper gives the first results and comments on this simulation.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"38 1","pages":"538-540"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79462552","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367022
G. R. Johnson
For the purposes of this paper, it is my premise that hardware is not the problem for the acceptance and utilization of massively parallel computing. If hardware is not the problem, what is? The answer has to be software and a deeper understanding of parallel algorithms. However, I believe that software has to be different if massively parallel computing is to be a major factor in computing in the future. And to become so, we must make major shifts in software to facilitate this development. I present my views on appropriate software for massively parallel computing environments. Before outlining possible directions for future development, I review where we have come from.<>
{"title":"Communications is more than I/O","authors":"G. R. Johnson","doi":"10.1109/MPCS.1994.367022","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367022","url":null,"abstract":"For the purposes of this paper, it is my premise that hardware is not the problem for the acceptance and utilization of massively parallel computing. If hardware is not the problem, what is? The answer has to be software and a deeper understanding of parallel algorithms. However, I believe that software has to be different if massively parallel computing is to be a major factor in computing in the future. And to become so, we must make major shifts in software to facilitate this development. I present my views on appropriate software for massively parallel computing environments. Before outlining possible directions for future development, I review where we have come from.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"49 1","pages":"2-5"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80677527","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367082
S. Lor, Hong Shen, P. Maheshwari
This paper proposes a method for partitioning the vertex set of an undirected simple weighted graph into two subsets so as to minimise the difference of vertex-weight sums between the two subsets and the total weight of edges cut (i.e., edges with one end in each subset). The proposed heuristic algorithm works in a divide-and-conquer fashion and is a modification of an algorithm suggested in the literature. The algorithm has the same time complexity as the previous one but is extended to work on weighted graphs.<>
{"title":"Divide-and-conquer minimal-cut bisectioning of task graphs","authors":"S. Lor, Hong Shen, P. Maheshwari","doi":"10.1109/MPCS.1994.367082","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367082","url":null,"abstract":"This paper proposes a method for partitioning the vertex set of an undirected simple weighted graph into two subsets so as to minimise the difference of vertex-weight sums between the two subsets and the total weight of edges cut (i.e., edges with one end in each subset). The proposed heuristic algorithm works in a divide-and-conquer fashion and is a modification of an algorithm suggested in the literature. The algorithm has the same time complexity as the previous one but is extended to work on weighted graphs.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"5 1","pages":"155-162"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88075387","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367042
A. Shafarenko
An applicative paradigm of parallel array processing based on recurrence relations and a data-parallel overloading of constants is presented. It is shown that the suggested principle of anti-currying together with introduction of function-based, eager arrays result in a denotational system superior to array extensions of pragmatic languages in that it can exploit spatial symmetries of arrays to unify the notation. The main novelty here is completely asynchronous treatment of arrays of arrow types (arrays of possibly array-valued functions) which lends itself nicely to a massively parallel data-flow implementation with yet static scheduling due to the imposed strictness of the array constructor. The evolution of data is defined in the tradition form of stream transformation.<>
{"title":"RETRAN: a recurrent paradigm for massively parallel array computing","authors":"A. Shafarenko","doi":"10.1109/MPCS.1994.367042","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367042","url":null,"abstract":"An applicative paradigm of parallel array processing based on recurrence relations and a data-parallel overloading of constants is presented. It is shown that the suggested principle of anti-currying together with introduction of function-based, eager arrays result in a denotational system superior to array extensions of pragmatic languages in that it can exploit spatial symmetries of arrays to unify the notation. The main novelty here is completely asynchronous treatment of arrays of arrow types (arrays of possibly array-valued functions) which lends itself nicely to a massively parallel data-flow implementation with yet static scheduling due to the imposed strictness of the array constructor. The evolution of data is defined in the tradition form of stream transformation.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"262 1-2 1","pages":"478-487"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78442233","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367029
A. Iazzetta, C. Sanges, U. Scafuri
A requirement for new multicomputers is the decoupling of computation and communication functionalities in order to improve the performance of both. The choice of an interconnection network and the routing strategy adopted define the internal architecture of the communication node. The paper presents a new routing strategy for WK-networks, based on two separate networks for the input and output phases, which enables a cost effective communication node to be designed.<>
{"title":"A routing strategy for WK-networks","authors":"A. Iazzetta, C. Sanges, U. Scafuri","doi":"10.1109/MPCS.1994.367029","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367029","url":null,"abstract":"A requirement for new multicomputers is the decoupling of computation and communication functionalities in order to improve the performance of both. The choice of an interconnection network and the routing strategy adopted define the internal architecture of the communication node. The paper presents a new routing strategy for WK-networks, based on two separate networks for the input and output phases, which enables a cost effective communication node to be designed.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"31 1","pages":"576-582"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77325339","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367068
P. K. Robertson, J. Lilleyman
Massively parallel architectures are well suited to a class of SIMD image processing and visualisation algorithms based on separable transformations, for which multi-dimensional operations can be decomposed into independent one-dimensional operations. This allows regularisation of data movements as parallel mappings between computational operations. This paper outlines a generic class of separable visualisation and image processing algorithms, their data movement requirements, and their implementation on a general purpose massively parallel array, the MasPar MP-1. Design and performance issues for a custom ASIC array and an interconnected DSP array are compared against the general purpose implementation. A software framework to support this class of algorithms on parallel platforms, and its extension to include generalisation to tiled data algorithms, are described. Design implications for special purpose hardware are discussed.<>
{"title":"Image based visualisation on massively parallel architectures: design considerations for general purpose and application specific arrays","authors":"P. K. Robertson, J. Lilleyman","doi":"10.1109/MPCS.1994.367068","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367068","url":null,"abstract":"Massively parallel architectures are well suited to a class of SIMD image processing and visualisation algorithms based on separable transformations, for which multi-dimensional operations can be decomposed into independent one-dimensional operations. This allows regularisation of data movements as parallel mappings between computational operations. This paper outlines a generic class of separable visualisation and image processing algorithms, their data movement requirements, and their implementation on a general purpose massively parallel array, the MasPar MP-1. Design and performance issues for a custom ASIC array and an interconnected DSP array are compared against the general purpose implementation. A software framework to support this class of algorithms on parallel platforms, and its extension to include generalisation to tiled data algorithms, are described. Design implications for special purpose hardware are discussed.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"54 3 1","pages":"268-280"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82688482","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367030
G. De Pietro, G. Reitanox, A. Esposito, R. Vaccaro
In heterogeneous environments, very sophisticated modelling techniques and tools are required for the analysis of interactions between hardware and software subsystems. After a detailed discussion about performance evaluation of heterogeneous systems, an online monitoring tool for heterogeneous distributed systems is presented. Portability, modularity and a powerful graphic interface are the main features of the developed tool.<>
{"title":"Supports for performance evaluation in heterogeneous distributed environments","authors":"G. De Pietro, G. Reitanox, A. Esposito, R. Vaccaro","doi":"10.1109/MPCS.1994.367030","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367030","url":null,"abstract":"In heterogeneous environments, very sophisticated modelling techniques and tools are required for the analysis of interactions between hardware and software subsystems. After a detailed discussion about performance evaluation of heterogeneous systems, an online monitoring tool for heterogeneous distributed systems is presented. Portability, modularity and a powerful graphic interface are the main features of the developed tool.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"65 1","pages":"570-575"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85788279","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367076
C. Verdier, A. Lafage, E. Boutillon, A. Demeure
Linear skewing schemes to store arrays of data in a parallel memory provide conflict free access to many linear patterns. In signal processing for SONAR application, linear vectors are accessed which belong to a class called p-ordered vectors. We describe in this article a new bi-dimensional interconnection network to unscramble p-ordered vectors. It is based on the Chinese Remainder Theorem and is easy to implement and control.<>
{"title":"A bidimensional Chinese interconnection network for SIMD architectures","authors":"C. Verdier, A. Lafage, E. Boutillon, A. Demeure","doi":"10.1109/MPCS.1994.367076","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367076","url":null,"abstract":"Linear skewing schemes to store arrays of data in a parallel memory provide conflict free access to many linear patterns. In signal processing for SONAR application, linear vectors are accessed which belong to a class called p-ordered vectors. We describe in this article a new bi-dimensional interconnection network to unscramble p-ordered vectors. It is based on the Chinese Remainder Theorem and is easy to implement and control.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"108 1","pages":"207-210"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75931720","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 : 1994-05-02DOI: 10.1109/MPCS.1994.367017
S. Beaty
List scheduling is a popular method of scheduling. It has the benefits of being a relatively fast technique and producing good results. It has weaknesses when dealing with restricted timing. Two ancillary methods, foresight and lookahead, have been developed to help mitigate this weakness. This paper compares the effectiveness of list scheduling alone, with foresight, and with lookahead. It also shows the benefits of lookahead are accrued with no time overhead.<>
{"title":"List scheduling: alone, with foresight, and with lookahead","authors":"S. Beaty","doi":"10.1109/MPCS.1994.367017","DOIUrl":"https://doi.org/10.1109/MPCS.1994.367017","url":null,"abstract":"List scheduling is a popular method of scheduling. It has the benefits of being a relatively fast technique and producing good results. It has weaknesses when dealing with restricted timing. Two ancillary methods, foresight and lookahead, have been developed to help mitigate this weakness. This paper compares the effectiveness of list scheduling alone, with foresight, and with lookahead. It also shows the benefits of lookahead are accrued with no time overhead.<<ETX>>","PeriodicalId":64175,"journal":{"name":"专用汽车","volume":"88 1","pages":"343-347"},"PeriodicalIF":0.0,"publicationDate":"1994-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72981792","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：