Path-oriented scheduling methods, such as trace scheduling and hyperblock scheduling, use speculation to extract instruction-level parallelism from control-intensive programs. These methods predict important execution...
ISBN:
(纸本)9780818676413
Path-oriented scheduling methods, such as trace scheduling and hyperblock scheduling, use speculation to extract instruction-level parallelism from control-intensive programs. These methods predict important execution paths in the current scheduling scope using execution profiling or frequency estimation. Aggressive speculation is then applied to the important execution paths, possibly at the cost of degraded performance along other paths. Therefore, the speed of the output code can be sensitive to the compiler's ability to accurately predict the important execution paths. Prior work in this area has utilized the speculative yield function by Fisher, coupled with dependence height, to distribute instruction priority among execution paths in the scheduling scope. While this technique provides more stability of performance by paying attention to the needs of all paths, it does not directly address the problem of mismatch between compile-time prediction and run-time behavior. The work presented in this paper extends the speculative yield and dependence height heuristic to explicitly minimize the penalty suffered by other paths when instructions are speculated along a path. Since the execution time of a path is determined by the number of cycles spent between a path's entrance and exit in the scheduling scope, the heuristic attempts to eliminate unnecessary speculation that delays any path's exit. Such control of speculation makes the performance much less sensitive to the actual path taken at run time. The proposed method has a strong emphasis on achieving minimal delay to all exits. Thus the name, speculative hedge, is used. This paper presents the speculative hedge heuristic, and shows how it controls over-speculation in a superblock/hyperblock scheduler. The stability of output code performance in the presence of execution variation is demonstrated with six programs from the SPEC CINT92 benchmark suite.
The proceedings contains 114 papers from the 1995 Association for computing Machinery (acm) symposium on appliedcomputing held at the Opryland Hotel, in Nashville on February 26-28, 1995. Topics discussed include app...
详细信息
The proceedings contains 114 papers from the 1995 Association for computing Machinery (acm) symposium on appliedcomputing held at the Opryland Hotel, in Nashville on February 26-28, 1995. Topics discussed include appliedcomputing in education applications, appliedcomputing database systems, scientific computing, biomedical computing, programming languages, parallel and distributed algorithms, and artificial intelligence.
This paper discusses the use of an individual computing environment (PC, workstation, etc.) to design and analyze fractals. An algorithm that facilitates the analysis and design of many fractals is given as an example.
ISBN:
(纸本)9780897916585
This paper discusses the use of an individual computing environment (PC, workstation, etc.) to design and analyze fractals. An algorithm that facilitates the analysis and design of many fractals is given as an example.
A new method for routing messages to the most available server by considering the change of information in the past and in the future is introduced. The use of fuzzy logic for a router to make routing decisions in an ...
详细信息
ISBN:
(纸本)9780897916585
A new method for routing messages to the most available server by considering the change of information in the past and in the future is introduced. The use of fuzzy logic for a router to make routing decisions in an individual computing environment is discussed. The method of routing will make the system more efficient.
Medical computer based applications always require very careful work to attain correct and patient-safe results. A good procedure to gain them is to apply learning methods like evolutionary algorithms. The use of such...
详细信息
ISBN:
(纸本)9780897916585
Medical computer based applications always require very careful work to attain correct and patient-safe results. A good procedure to gain them is to apply learning methods like evolutionary algorithms. The use of such kind of optimizing methods depends on the construction of a so called objective function which offers the direction and base of the learning process. But the development of those function is limited by the power of today's computer architectures. This paper will show this for the case of medical applications.
A learner-controlled system of instruction is described in which one or more key instructional decisions are delegated to the learner with the aid of the Internet for distance education. Two distance courses have been...
详细信息
ISBN:
(纸本)9780897916585
A learner-controlled system of instruction is described in which one or more key instructional decisions are delegated to the learner with the aid of the Internet for distance education. Two distance courses have been developed: Information technology and Systems analysis and design. An environment of just-in-time help was also developed and these courses are summarized in the paper.
Microcomputers running data acquisition, analysis, and communication software have helped scientists automate their laboratories, improve the quality of research, enhance productivity, and facilitate collaboration. De...
详细信息
ISBN:
(纸本)9780897916585
Microcomputers running data acquisition, analysis, and communication software have helped scientists automate their laboratories, improve the quality of research, enhance productivity, and facilitate collaboration. Despite the growing use of these and other software tools, reliance on the traditional pen and laboratory notebook for recording and organizing experimental data remains commonplace. This paper presents a data model of scientific experimentation for designing database systems that chronicle the empirical study of natural processes.
This paper describes the application of ray tracing, rendering, and morphing technology to the study and analysis of the spread of dwarf mistletoe. The practical problems involved in examining the spread and speed of ...
详细信息
ISBN:
(纸本)9780897916585
This paper describes the application of ray tracing, rendering, and morphing technology to the study and analysis of the spread of dwarf mistletoe. The practical problems involved in examining the spread and speed of such a slow pathogen, with such serious economic consequences, is illustrated. The utility of applying this standard graphics technology to similar forest diseases of this type is also established. Practical examples of this technology transfer are given on a database of 6000 trees with information collected over 15 years.
In using deterministic models for weather forecasting, the quality of the forecast depends critically on the initial and/or boundary conditions and values of certain key parameters of interest to the model. Since the ...
详细信息
In using deterministic models for weather forecasting, the quality of the forecast depends critically on the initial and/or boundary conditions and values of certain key parameters of interest to the model. Since the initial and/or boundary conditions and the key parameters control the quality of the forecast, they are often known as control variables. In this paper we analyze the effectiveness of the so-called adjoint method for the optimal determination of control variables for a class of non-linear dynamical model, called the mixed-layer model. We illustrate the use of this method by assimilating the actual data collected in a cold air outflow over the Gulf of Mexico.
Finding a minimum node cutset in a delay network is a very important operation in the tinting optimization phase of a number of synthesis systems. The simplest version of this problem can be described as: Given a netw...
详细信息
Finding a minimum node cutset in a delay network is a very important operation in the tinting optimization phase of a number of synthesis systems. The simplest version of this problem can be described as: Given a network (directed acyclic graph), N = (V, E), with two distinct vertices s and t (s and t are not connected by any edge), find a minimum cardinality subset of V, V′, such that any path froth s to t uses at least one member of V′. V′ is called a minimum node cutset. In this paper, an efficient algorithm for finding a minimum node cutset from a vertex-disjoint path set is proposed. The time complexity of the algorithm is bounded by O (n+e).
暂无评论