Compiler-directed data speculation has been implemented on Itanium systems to allow for a compiler to move a load across a store even when the two operations are potentially aliased This not only breaks data dependenc...
详细信息
Compiler-directed data speculation has been implemented on Itanium systems to allow for a compiler to move a load across a store even when the two operations are potentially aliased This not only breaks data dependency to reduce critical path length, but also allows a load to be scheduled far apart from its uses to hide cache miss latencies. However, the effectiveness of data speculation is affected by the sophistication of alias analysis technique as well as the aggressiveness of the instruction scheduler. In general, the more sophisticated is the alias analysis technique, the less performance gain is from data speculation, and the more aggressive is the instruction scheduler, the more opportunity is for data speculation. In this paper we evaluate in various scenarios the performance potentials of data speculation for SPEC2000C benchmarks. For each scenario, we determine the performance contributions of data speculation due to both critical path reduction and cache miss latency reduction. We also show interesting statistics about the effects of scheduling constraints, the percentage of critical dependencies, the impacts of cache miss latencies, and the distances between the load locations before and after data speculation.
Autonomic computing - self-configuring, self-healing, self-optimizing applications, systems and networks - is widely believed to be a promising solution to ever-increasing system complexity and the spiraling costs of ...
详细信息
Autonomic computing - self-configuring, self-healing, self-optimizing applications, systems and networks - is widely believed to be a promising solution to ever-increasing system complexity and the spiraling costs of human system management as systems scale to global proportions. Most results to date, however, suggest ways to architect new software constructed from the ground up as autonomic systems, whereas in the real world organizations continue to use stovepipe legacy systems and/or build ''systems of systems'' that draw from a gamut of new and legacy components involving disparate technologies from numerous vendors. Our goal is to retrofit autonomic computing onto such systems, externally, without any need to understand or modify the code, and in many cases even when it is impossible to recompile. We present a meta-architecture implemented as active middleware infrastructure to explicitly add autonomic services via an attached feedback loop that provides continual monitoring and, as needed, reconfiguration and/or repair. Our lightweight design and separation of concerns enables easy adoption of individual components, as well as the full infrastructure, for use with a large variety of legacy, new systems, and systems of systems. We summarize several experiments spanning multiple domains.
In recent work, we introduced a generalization of ECOC learning under the theory of recursive error correcting codes. We named it RECOC (Recursive ECOC) learning. If long output codewords are allowed, as in the case o...
详细信息
The Constraint Language for Lambda Structures (CLLS)is an expressive language of tree descriptions which combinesdominance constraints with powerful parallelism and bindingconstraints. CLLS was introduced as a uniform...
The Constraint Language for Lambda Structures (CLLS)is an expressive language of tree descriptions which combinesdominance constraints with powerful parallelism and bindingconstraints. CLLS was introduced as a uniform framework for definingunderspecified semantics representations of natural languagesentences, covering scope, ellipsis, and anaphora. This articlepresents saturation-based algorithms for processing the completelanguage of CLLS. It also gives an overview of previous results on questions of processing and complexity.
Hydrographic survey performed by the U.S. Army Corps of Engineers (USACE), on the 350 miles of Mississippi river is discussed. It is observed that the river loses velocity, dropping silt, sand and sediment at each of ...
详细信息
Hydrographic survey performed by the U.S. Army Corps of Engineers (USACE), on the 350 miles of Mississippi river is discussed. It is observed that the river loses velocity, dropping silt, sand and sediment at each of its mouth. USACE uses differential global positioning system (DGPS), computer aided design (CAD), geospatial, wireless and internet technologies to preprocess survey data in the field. The survey data is helpful in dredging operations on the lower Mississippi to ensure navigable waterways.
作者：
Giovanni AdagioDistributed Programming Laboratory
School of Computer Science and Communication Systems Swiss Federal Institute of Technology in Lausanne (EPFL) Institute of Mathematics School of Basic Sciences Swiss Federal Institute of Technology in Lausanne (EPFL) Switzerland
This paper contributes to the characterization of synchronous models of distributed computing using topological techniques. We consider a generic synchronous model with send-omission failures and use a topological str...
This paper contributes to the characterization of synchronous models of distributed computing using topological techniques. We consider a generic synchronous model with send-omission failures and use a topological structure corresponding to a bounded number of rounds of the model. We observe some nice properties of the structure and derive from these properties necessary and sufficient conditions to solve consensus in this model.
Surgical planning as a treatment for vascular diseases re- quires fast blood ow simulations that are eÆcient in handling changing geometry. It is, for example, necessary to try diþerent paths of a planned by...
详细信息
The purpose of Generic Evolutionary Algorithms programming Library (GEA1) system is to provide researchers with an easy-to-use, widely applicable and extendable programming library which solves real-world optimization...
The purpose of Generic Evolutionary Algorithms programming Library (GEA1) system is to provide researchers with an easy-to-use, widely applicable and extendable programming library which solves real-world optimization problems by means of evolutionary algorithms. It contains algorithms for various evolutionary methods, implemented genetic operators for the most common representation forms for individuals, various selection methods, and examples on how to use and expand the library. All these functions assure that GEA can be effectively applied on many problems. GraphGEA is a graphical user interface to GEA written with the GTK API. The numerous parameters of the evolutionary algorithm can be set in appropriate dialog boxes. The program also checks the correctness of the parameters and saving/restoring of parameter sets is also possible. The selected evolutionary algorithm can be executed interactively on the specified optimization problem through the graphical user interface of GraphGEA, and the results and behavior of the EA can be observed on several selected graphs and drawings. While the main purpose of GEA is solving optimization problems, that of GraphGEA is education and analysis. It can be of great help for students understanding the characteristics of evolutionary algorithms and researchers of the area can use it to analyze an EA's behavior on particular problems.
作者：
Fiege, LudgerGärtner, Felix C.Kasten, OliverZeidler, Andreas
Department of Computer Science Databases Distributed System Group D-64283 Darmstadt Germany
School of Computer and Communication Sciences Distributed Programming Laboratory CH-1015 Lausanne Switzerland
Department of Computer Science Distributed Systems Group CH-8092 Zurich Switzerland
Publish/subscribe (pub/sub) is considered a valuable middleware architecture that proliferates loose coupling and leverages reconfigurability and evolution. Up to now, existing pub/sub middleware was optimized for sta...
详细信息
作者：
Gupta, GopalPontelli, EnricoApplied Logic
Programming Languages and Systems Lab. Department of Computer Science University of Texas at Dallas Richardson TX 95083 United States Laboratory for Logic
Databases and Advanced Programming Department of Computer Science New Mexico State University Las Cruces NM 88003 United States
Domain Specific Languages (DSLs) are high level languages designed for solving problems in a particular domain, and have been suggested as means for developing reliable software systems. We present a (constraint) logi...
详细信息
暂无评论