D. Beazley, P. Lomdahl, Niels Grønbech-Jensen, P. Tamayo
{"title":"A high performance communications and memory caching scheme for molecular dynamics on the CM-5","authors":"D. Beazley, P. Lomdahl, Niels Grønbech-Jensen, P. Tamayo","doi":"10.1109/IPPS.1994.288213","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288213","url":null,"abstract":"Presents several techniques that we have used to optimize the performance of a message-passing C code for molecular dynamics on the CM-5. We describe our use of the CM-5 vector units and a parallel memory caching scheme that we have developed to speed up the code by more than 50%. A modification that decreases communication time by 35% is also presented, along with a discussion of how we have been able to take advantage of the CM-5 hardware without significantly compromising code portability. We have been able to speed up our original code by a factor of ten and we feel that our modifications may be useful in optimizing the performance of other message-passing C applications on the CM-5.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"108 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130558453","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}
{"title":"Automatic array alignment as a step in hierarchical program transformation","authors":"B. Cukic, F. Bastani","doi":"10.1109/IPPS.1994.288246","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288246","url":null,"abstract":"Presents an original approach to automatic array alignment, the step in the hierarchical transformation system aimed at the efficient execution of shared memory programs on distributed memory machines. The array alignment algorithm deals with a broad set of intra-dimension and inter-dimension alignment preferences, including offsets, strides, permutations, embeddings, and their combinations. The authors discuss the algorithm and the tests performed on the Connection Machine CM-2.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125666922","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}
{"title":"Dagger: combining benefits of synchronous and asynchronous communication styles","authors":"Attila Gürsoy, L. Kalé","doi":"10.1109/IPPS.1994.288244","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288244","url":null,"abstract":"Communication using blocking receives is the commonly used mechanism in parallel programming today. Message driven execution is an alternate mechanism which does not use receive style statements at all. The message driven execution style promotes the overlap of computation and communication. Programs written in this style exhibit increased latency tolerance: their performance does not degrade significantly with latency. It also induces compositionality: multiple independently developed modules can be combined correctly without loss of efficiency. However, as the flow of control is not explicit in such programs, they are often difficult to develop and debug. The authors present a coordination language called Dagger to alleviate this problem. The language has been implemented in the Charm parallel programming system, and runs programs portably on a variety of parallel machines.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125925746","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}
{"title":"Queue locks on cache coherent multiprocessors","authors":"Peter S. Magnusson, A. Landin, Erik Hagersten","doi":"10.1109/IPPS.1994.288305","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288305","url":null,"abstract":"Large-scale shared-memory multiprocessors typically have long latencies for remote data accesses. A key issue for execution performance of many common applications is the synchronization cost. The communication scalability of synchronization has been improved by the introduction of queue-based spin-locks instead of Test&(Test&Set). For architectures with long access latencies for global data, attention should also be paid to the number of global accesses that are involved in synchronization. We present a method to characterize the performance of proposed queue lock algorithms, and apply it to previously published algorithms. We also present two new queue locks, the LH lock and the M lock. We compare the locks in terms of performance, memory requirements, code size and required hardware support. The LH lock is the simplest of all the locks, yet requires only an atomic swap operation. The M lock is superior in terms of global accesses needed to perform synchronization and still competitive in all other criteria. We conclude that the M lock is the best overall queue lock for the class of architectures studied.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"179 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132354142","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}
{"title":"Deterministic permutation routing on a reconfigurable mesh","authors":"A. Kapoor, Heiko Schröder, B. Beresford-Smith","doi":"10.1109/IPPS.1994.288252","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288252","url":null,"abstract":"Presents an algorithm for Permutation Routing on a reconfigurable mesh based on the divide and conquer paradigm. The algorithm has a worst case running time of 4n+O(log2/sup /n) communication steps and requires a maximum queue size of 5. The O(log/sup 2/ n) term is associated with communication of single words which convey information about the distribution of packets. Hence, if the packet length is much larger than a single word, then the constant associated with this term is negligible.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130135010","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}
{"title":"On evil twin networks and the value of limited randomized routing","authors":"B. Alleyne, I. Scherson","doi":"10.1109/IPPS.1994.288247","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288247","url":null,"abstract":"A dynamic 2-stage Delta network (N inputs and outputs) is introduced and analyzed for permutation routing. The notion of evil twins is introduced and a deterministic procedure is given to route any permutation in no more than 2/spl timessup 4spl radic/(N) network cycles. Two limited randomized routing schemes are then given and yield on average at most N/sub A/+1+1/N (N/sub A/=O(log N/log(log N=1))) and is the greatest integer such that (N/sub A/)!/spl les/N and /spl lsqb/log(log N+1)/spl rsqb/+2+1/N network cycles for any input permutation. The probability of any permutation requiring at least c network cycles more than the average bounds above is at most 1/(c+1)! and 1/N/sup 2exp(c)/.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133480149","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}
{"title":"Sorting strings and constructing digital search trees in parallel","authors":"J. JáJá, K. Ryu, U. Vishkin","doi":"10.1109/IPPS.1994.288278","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288278","url":null,"abstract":"We describe two simple optimal-work parallel algorithms for sorting a list /spl Lscr/=(X/sub 1/,X/sub 2/,...,X/sub m/) of m strings over an arbitrary alphabet /spl Sigma/, where /spl Sigmasub i=1sup mspl verbar/X/sub ispl verbar/=n. The first algorithm is a deterministic algorithm that runs in O((log/sup 2/ m)/(log log m)) time and the second is a randomized algorithm that runs in O(log m) time. Both algorithms use O(m log(m)+n) operations. Compared to the best known parallel algorithms for sorting strings, the algorithms offer the following improvements: the total number of operations used by the algorithms is optimal while all previous parallel algorithms use a non-optimal number of operations; we make no assumption about the alphabet while the previous algorithms assume that the alphabet is restricted to /spl lcub/1,2,..., n/sup O(1/)/spl rcub/; the computation model assumed by the algorithms is the Common CRCW PRAM unlike the known algorithms that assume the Arbitrary CRCW PRAM; and the presented algorithms use O(m log m+n) space, while previous parallel algorithms use O(n/sup 1+/spl epsiv) space, where /spl epsiv/ is a positive constant. We also present optimal-work parallel algorithms to construct a digital search tree for a given set of strings and to search for a string in a sorted list of strings. We use the parallel sorting algorithms to solve the problem of determining a minimal starting point of a circular string with respect to lexicographic ordering.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"109 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133267781","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}
N. Abu-Ghazaleh, Xianzhi Fan, P. Wilsey, D. Hensgen
{"title":"Variable instruction issue for efficient MIMD interpretation on SIMD machines","authors":"N. Abu-Ghazaleh, Xianzhi Fan, P. Wilsey, D. Hensgen","doi":"10.1109/IPPS.1994.288285","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288285","url":null,"abstract":"Programming SIMD hardware to interpret (in parallel) programs and data resident in each PE is a technique for obtaining a cost effective, massively parallel MIMD processing environment. Although heavily dependent on each application that is interpreted, the performance of the synthesized MIMD environment is greatly influenced by the organization of the instruction interpreter. For example, it is possible to delay the interpretation of infrequent operations to improve the overall performance of the MIMD processing environment. Interpreters that attempt improved MIMD performance by deferring infrequent operations are called variable issue control loops. This paper examines the construction of optimized variable issue control loops. In particular, we study the problem of building control loops that optimize either (i) instruction throughput or (ii) PE utilization and present three heuristic algorithms for deriving near optimal variable issue control loops.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"289 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132566628","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}
{"title":"Solving the all-pair shortest path problem on interval and circular-arc graphs","authors":"D. Chen, D. T. Lee","doi":"10.1109/IPPS.1994.288297","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288297","url":null,"abstract":"We show that, given the interval model of an unweighted n-vertex interval graph, the information for the all-pair shortest paths can be made available very efficiently, both in parallel and sequentially. After sorting the input intervals by their endpoints, an O(n) space data structure can be constructed optimally in parallel, in O(log n) time using O(n/log n) CREW PRAM processors. Using the data structure, a query on the length of the shortest path between any two input intervals can be answered in O(1) time using one processor, and a query on the actual shortest path can be answered in O(1) time using k processors, where k is the number of intervals on that path. Our parallel algorithm immediately implies a new sequential result: After an O(n) time preprocess, shortest paths can be reported optimally. Our techniques can be extended to solve the problem on circular-arc graphs, both in parallel and sequentially, in the same complexity bounds. The previously best known sequential algorithm for computing the all-pair shortest paths in interval graphs lakes O(n/sup 2/) time and uses O(n/sup 2/) space to store the lengths of the all-pair shortest paths.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124687379","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}
{"title":"Fuzzy communication for guided loop scheduling in multicomputers","authors":"Chien-Ming Cheng, K. Hwang","doi":"10.1109/IPPS.1994.288265","DOIUrl":"https://doi.org/10.1109/IPPS.1994.288265","url":null,"abstract":"We propose the use of guided loop scheduling and fuzzy communications to map shared-variable communications into message passing operations among multicomputers. The mapping mechanism converts scalar message passing operations into multiple broadcast or multiple multicast operations. The proposed method is evaluated by both simulation experiments and theoretical analysis. The performance results, considering both communication and computation, are reported for mapping two distributed matrix algorithms. Simulated benchmark results demonstrate improved performance and minimized execution time. The proposed method offers the potential to take advantage of the programmability of shared-memory systems and the scalability of distributed-memory systems.<<ETX>>","PeriodicalId":258702,"journal":{"name":"Proceedings of 8th International Parallel Processing Symposium","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129471939","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}