高速缓存优化的并行连接算法

由于嵌套循环连接操作过程中存在较大的高速缓存缺失,严重影响了连接查询的性能。提出了一种基于缓冲的高速缓存参数无关的嵌套循环并行连接算法。通过高速缓存参数无关和缓冲技术,提高了连接算法的空间局部性和时间局部性。理论分析和实验结果表明,高速缓存优化后的串行连接算法的性能是原来的2倍,其并行算法效果近似线性加速比。

模型检测MESIF Cache一致性协议

在处理器从单核向多核演进的过程中,为了获得更好的性能和可扩展性,适用于多核处理器系统的Cache一致性协议变得越来越复杂。Cache一致性协议的验证一直是模型检测在工业界主要应用之一,被工业界和学术界关注。相对传统方法而言,微结构级的模型检测能够描述和验证更多的协议细节。利用NuSMV工具对Intel公司的MESIF Cache一致性协议进行模型检测在微结构层次上进行了建模,并对该协议进行模型检测,试验结果证明了此方法的有效性。

The piston model of coronal transients

In the present paper, the piston model of the coronal transient (see Hu. 1983a, b is discssed in detail, and the quantitative results of unsteady gasdynamics are applied to the coronal transient processes. The piston model explains the major features of the transient observations, such as the density profile, the geometric configuration, the kinetic process and the classifications of the coronal transient. Based on the idea of piston model, the bright feature and the dark feature of the transient are the gasdynamical response of the dense plasma ejecting into the corona, and associate with the compressed and rarefied flows, respectively. The quantitative results show that the density increment in the compressed region and the density decrement in the rarefied region are one order of magnitude larger and smaller, respectively, to the density in the quiet corona, it agrees quantitatively with the observations, and both the bright feature and dark feature are explained at the same time.

A combined statistical model for multiple motifs search

Transcription factor binding sites (TFBS) play key roles in genebior 6.8 wavelet expression and regulation. They are short sequence segments with de¯nite structure and can be recognized by the corresponding transcription factors correctly. From the viewpoint of statistics, the candidates of TFBS should be quite di®erent from the segments that are randomly combined together by nucleotide. This paper proposes a combined statistical model for ¯nding over- represented short sequence segments in di®erent kinds of data set. While the over-represented short sequence segment is described by position weight matrix, the nucleotide distribution at most sites of the segment should be far from the background nucleotide distribution. The central idea of this approach is to search for such kind of signals. This algorithm is tested on 3 data sets, including binding sites data set of cyclic AMP receptor protein in E.coli, PlantProm DB which is a non-redundant collection of proximal promoter sequences from di®erent species, collection of the intergenic sequences of the whole genome of E.Coli. Even though the complexity of these three data sets is quite di®erent, the results show that this model is rather general and sensible.

Maze model to study spatial learning and memory in freely moving monkeys

Many types of mazes have been used in cognitive brain research and data obtained from those experiments, especially those from rodents' studies, support the idea that the hippocampus is related to spatial learning and memory. But the results from non-huma

DEFECT MODEL OF PHOTOCONVERSION BY OXYGEN IN SEMIINSULATING GAAS

The basic idea of a defect model of photoconversion by an oxygen impurity in semi-insulating GaAs, proposed in an earlier paper, is described in a systematic way. All experiments related to this defect, including high-resolution spectroscopic measurements, piezospectroscopic study, and recent measurements on electronic energy levels, are explained on the basis of this defect model. The predictions of the model are in good agreement with the experiments. A special negative-U mechanism in this defect is discussed in detail with an emphasis on the stability of the charge states. The theoretical basis of using a self-consistent bond-orbital model in the calculation is also given.

Biomimetic (topological) pattern recognition - A new model of pattern recognition theory and its application

A new theoretical model of Pattern Recognition principles was proposed, which is based on "matter cognition" instead of "matter classification" in traditional statistical Pattern Recognition. This new model is closer to the function of human being, rather than traditional statistical Pattern Recognition using "optimal separating" as its main principle. So the new model of Pattern Recognition is called the Biomimetic Pattern Recognition (BPR)(1). Its mathematical basis is placed on topological analysis of the sample set in the high dimensional feature space. Therefore, it is also called the Topological Pattern Recognition (TPR). The fundamental idea of this model is based on the fact of the continuity in the feature space of any one of the certain kinds of samples. We experimented with the Biomimetic Pattern Recognition (BPR) by using artificial neural networks, which act through covering the high dimensional geometrical distribution of the sample set in the feature space. Onmidirectionally cognitive tests were done on various kinds of animal and vehicle models of rather similar shapes. For the total 8800 tests, the correct recognition rate is 99.87%. The rejection rate is 0.13% and on the condition of zero error rates, the correct rate of BPR was much better than that of RBF-SVM.