The giant zooxanthellae-bearing ciliate Maristentor dinoferus (Heterotrichea) is closely related to folliculinidae

The small subunit rDNA sequence of Maristentor dinoferus (Lobban, Schefter, Simpson, Pochon, Pawlowski, and Foissner, 2002) was determined and compared with sequences from other Heterotrichea and Karyorelictea. Maristentor resembles Stentor in basic morphology and had been provisionally assigned to Stentoridae. However, our phylogenetic analyses show that Maristentor is more closely related to Folliculinidae. Our results support the creation of a separate family for Maristentor, Maristentoridae n. fam., and also confirm the phylogenetic grouping of Folliculindae, Stentoridae, Blepharismidae, and Maristentoridae, which we informally call 'stentorids'. Maristentor, rather than Stentor itself, appears to be most significant in understanding the origins of folliculinids from their aloricate ancestors. Our analyses suggest continued uncertainty in the exact placement of the root of heterotrichs with this phylogenetic marker.

Phylogenetic relationships of the subclass Peritrichia (Oligohymenophorea, Ciliophora) with emphasis on the genus Epistylis, inferred from small subunit rRNA gene sequences

The peritrichs have been recognized as a higher taxon of ciliates since 1968. However, the phylogenetic relationships among them are still unsettled, and their placement within the class Oligohymenophorea has only been supported by the analysis of the small subunit rRNA gene sequence of Opisthonecta henneguyi. DNA was isolated directly from field-sampled species for PCR, and was used to resolve relationships within the genus Epistylis and to confirm the stability of the placement of peritrichs. Small subunit rRNA gene sequences of Epistylis plicatilis, Epistylis urceolata, Epistylis chysemydis, Epistylis hentscheli, Epistylis wenrichi, and Vorticella campanula were sequenced and analyzed using both distance-matrix and maximum-parsimony methods. In phylogenetic trees, the monophyly of both the genus Epistylis and the subclass Peritrichia was strongly supported, while V. campanula clustered with Vorticella microstoma. The topology in which E. plicatilis and E. hentscheli formed a strongly supported sister clade to E. urceolata, E. chrysemydis, and E. wenrichi was consistent with variations in the thickness of the peristomial lip. We concluded that the peristomial area, especially the. peristomial lip, might be the important phylogenetic character within the genus Epistylis.

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。

循环优化序列自动定制方法研究

现代计算机在体系结构和应用场景复杂性的增长使得程序性能的增长、保持程序性能的可移植性以及程序开发效率的提升越来越困难，程序自动调优(auto-tuning) 是解决此问题的一个可行途径。本文研究一种通过为循环嵌套自动定制循环优化序列来进行程序自动调优的方法，提高了被编译程序的性能和可移植性，也有助于提升开发效率。 在综述相关工作的基础上，本文首先分析了循环优化序列自动定制研究需要解决的几个关键问题：循环优化的形式化描述及实施、循环优化序列的性能评估、循环优化序列的自动定制方法以及相关的原型工具。本文围绕这几个问题进行研究。 本文采用了Polyhedron模型来描述循环优化，并基于此实施了循环优化的仿射变换。这一工作基于URUK/WRaP-IT开源软件包实现，并作出了若干重要改进。借助Polyhedron模型，本文还给出了部分常见循环优化之间的使能(enable)关系。循环优化之间的相互影响较为普遍，一个循环优化的执行可能会破坏后续优化的实施条件，也可能不破坏甚至创造相应条件。这一关系可以用于过滤循环优化序列定制过程中产生的无效序列从而加快定制速度。本文分析了两个循环优化以先后次序执行时，循环嵌套内所有依赖距离向量均得以保持的条件，给出了它们之间的使能关系。并且给出了使能关系在Polyhedron模型下的计算方法，以及在序列搜索过程中通过维护依赖距离向量状态计算使能关系的方法。 对于循环优化序列性能评估方法，本文采用了高速缓存失效率简化方程(CMES方程)以动态给出循环嵌套中的高速缓存失效率数据，提出了LoopCost模型以静态分析方式发现循环嵌套中对性能影响最大的维度，这两种评估方法分别适用于不同的序列定制方法。 给出了Cache 失效率简化方程(CMES, Cache Miss Equation Simplified)在Polyhedron 模型下的求解方法。CMES方程可在编译时给出循环嵌套中的Cache失效率数据，不需要实际编译运行目标程序，从而避免相关文献中使用运行时信息来评价优化序列带来的巨大编译开销。给出了CMES方程各个参数从循环嵌套的Polyhedron模型中提取的方法，并给出了具体计算过程。 提出了一种可发现循环嵌套中对性能影响最大维度的循环优化性能评估模型LoopCost。相关文献中所使用的性能评估模型可给出一个循环嵌套的整体性能，但无法给出哪个维度对其性能影响最大；而大多数循环优化实际上是对某一个维度所做的变换，因此这些评估模型不能直接指导优化序列的搜索，只能是在一个迭代过程中提供性能信息。LoopCost模型基于数据依赖关系对一个循环嵌套估算出其每个维度使用的cache行数，作为衡量局部性优劣的指标，适用于紧嵌套循环和非紧嵌套循环，并且其复杂度与循环嵌套个数和引用个数成线性关系。通过对SPEC CPU 2006中5,703个循环嵌套的实验，LoopCost模型对其中98.3%的循环嵌套都能给出正确的性能评估数据。 对于循环优化序列自动定制方法，本文首先提出了一种基于CMES方程的Custom方法，使用迭代方式搜索可行优化序列；为了进一步提升优化效率，又提出了一种基于LoopCost模型的快速定制方法FastCustom。 Custom方法使用CMES方程评估优化序列，使用遗传算法作为序列搜索方法。在遗传算法中的各个算子中利用循环优化的enable关系加速搜索速度。该方法可以避免使用运行时刻信息评估性能所带来的巨大编译开销，又可避免相关工作中利用Polyhedron模型中语句实例执行次序矩阵的性质来搜索带来的局限性，适用的循环优化范围广，编译效率高，同时能取得与相关工作基本相当的加速比。 FastCustom方法优化效率高，并且与Custom方法不同，适用于所有场合。它基于LoopCost模型寻找对性能影响最大的维度，并根据一个启发式策略给出一个循环优化序列。该方法编译速度快，优化效果与Custom方法基本相当。与Custom相比，该方法需要给出启发式策略，因此适用的循环优化范围不如Custom，适用于迭代编译无法进行的情形。 对于循环优化序列自动定制原型工具，本文设计并实现了原型工具CPOLO。该工具改进了URUK/WRaP-IT开源软件包，使其能够支持较为复杂的程序；增强了代码重生成的效率；并且实现了本文所提出的优化序列性能评估模型和自动定制方法。 最后，总结了全文并探讨了进一步的工作方向。

基于序列模式的Servlet容器缓存替换

Servlet缓存能够有效地提高Servlet容器的吞吐量,缩短用户请求的响应时间.然而,Servlet缓存的性能受到缓存替换算法的影响.Servlet容器中的Servlet对应着一定的业务功能,挖掘Servlet之间的业务关联来指导缓存替换算法的设计可以提高Servlet缓存的命中率,进而提高Servlet容器的性能.然而,目前常见的LRU(least recently used),LFU(least frequently used),GDSF(greedy dual size frequency)等缓存替换算法均没有考虑上述问题.将Servlet对应的业务关联定义为Servlet容器序列模式,并提出k步可缓存转移概率图的概念加以表示,给出了序列模式发现算法KCTPG_Discovery.最后,基于Servlet容器序列模式设计了缓存替换算法KP-LRU(k-steps prediction least recently used)和KP-GDSF(k-steps prediction least frequently used).实验结果表明,KP-LRU与KP-GDSF算法比对应的LRU算法和GDSF算法具有更高的缓存命中率,有效地提高了Servlet容器的性能.

基于静态分析的强制访问控制框架的正确性验证

现阶段对操作系统的强制访问控制框架的正确性验证的研究主要集中于对授权钩子放置的验证.文中基于TrustedBSD MAC框架对强制访问控制框架的正确性验证问题进行了研究,在授权钩子放置验证的基础上,提出了安全标记的完全初始化验证和完全销毁验证.为了实现上述验证,文中提出了一个路径敏感的、基于用户自定义检查规则的静态分析方法.该方法通过对集成于编译器的静态分析工具mygcc进行扩展来验证强制访问控制框架的钩子放置的准确性和完备性.该方法具有完全的路径覆盖性,且具有低的误报率和时间开销.

小麦实时控制灌溉的土壤水分含量探头合理埋设深度研究

利用室内控制试验研究了根据不同深度土壤水分传感器灌溉处理对冬小麦生物学性状及水分利用率等的影响。结果显示,以10 cm探头控制灌溉最为省水,同时冬小麦生物学性状及水分利用率等最佳。由于试验冬小麦处于生育前期,随着小麦生育期延伸,当根系超过30 cm深度时,根系吸水的深度增加,探头的埋设深度需要田间试验进行更详细的研究。

Flying lemurs – The 'flying tree shrews'? Molecular cytogenetic evidence for a Scandentia-Dermoptera sister clade

Background: Flying lemurs or Colugos (order Dermoptera) represent an ancient mammalian lineage that contains only two extant species. Although molecular evidence strongly supports that the orders Dermoptera, Scandentia, Lagomorpha, Rodentia and Primates form a superordinal clade called Supraprimates (or Euarchontoglires), the phylogenetic placement of Dermoptera within Supraprimates remains ambiguous. Results: To search for cytogenetic signatures that could help to clarify the evolutionary affinities within this superordinal group, we have established a genome-wide comparative map between human and the Malayan flying lemur (Galeopterus variegatus) by reciprocal chromosome painting using both human and G. variegatus chromosome-specific probes. The 22 human autosomal paints and the X chromosome paint defined 44 homologous segments in the G. variegatus genome. A putative inversion on GVA 11 was revealed by the hybridization patterns of human chromosome probes 16 and 19. Fifteen associations of human chromosome segments (HSA) were detected in the G. variegatus genome: HSA1/3, 1/10, 2/21, 3/ 21, 4/8, 4/18, 7/15, 7/16, 7/19, 10/16, 12/22 (twice), 14/15, 16/19 (twice). Reverse painting of G. variegatus chromosome-specific paints onto human chromosomes confirmed the above results, and defined the origin of the homologous human chromosomal segments in these associations. In total, G. variegatus paints revealed 49 homologous chromosomal segments in the HSA genome. Conclusion: Comparative analysis of our map with published maps from representative species of other placental orders, including Scandentia, Primates, Lagomorpha and Rodentia, suggests a signature rearrangement (HSA2q/21 association) that links Scandentia and Dermoptera to one sister clade. Our results thus provide new evidence for the hypothesis that Scandentia and Dermoptera have a closer phylogenetic relationship to each other than either of them has to Primates.

稀疏矩阵向量乘的存储访问复杂度及自动性能优化技术应用研究

在科学计算中，稀疏矩阵向量乘（SpMV, y=Ax）是一个十分重要的，且经常被大量调用的计算内核，广泛应用在科学计算、信息检索、气象、航天、油藏模拟、天体物理、数据挖掘等科学计算和实际应用中。在实际工程应用中，重复调用稀疏矩阵向量乘内核的次数常常会达到成千上万次。但在现代基于Cache存储层次的计算平台上，稀疏矩阵向量乘的性能很低。如果能够提高稀疏矩阵向量乘的运算速度，整个工程计算的运行效率将会得到很大的改善，计算时间也会大幅度的减少。因此优化稀疏矩阵向量乘的性能成为提高工程效率的关键，在实际应用中有着十分重要的意义。 SpMV的传统算法实现形式运行效率很低，主要原因是浮点计算操作和存储访问操作比率非常低，且稀疏矩阵非零元分布的不规则性使得存储访问模式非常复杂。寄存器分块算法和启发式选择分块算法,通过自适应选择性能最佳的分块大小，然后将稀疏矩阵分成小的稠密分块，所有的非零子块顺序计算，达到重用保存在寄存器中向量x元素的目的，减少存储访问次数和时间，从而提高这一重要内核的性能。我们在Pentium IV、Alpha EV6和AMD Athlon三个平台上，分别测试了十个矩阵下的两种不同算法形式(压缩行存储算法和寄存器分块算法)的性能，平均加速比分别达到1.69、1.90和 1.48。同时也测试了不同次数调用SpMV两种算法所用的时间，发现在实际的迭代算法应用过程中，若想采用启发式-寄存器分块算法达到性能提高的目的，一般情况下，迭代次数需要达到上百次才能有加速效果。 DRAM(h)模型是基于存储层次的并行计算模型，指出算法的复杂性包括计算复杂性和存储访问复杂性，具有近乎相同时间和空间复杂性的同一算法的不同实现形式，会有不同的存储访问复杂度，导致程序实际运行性能的差异；利用DRAM(h)模型进行分析并比较不同算法实现形式的存储访问复杂度，可以判断两种算法形式的优劣，从而为选取性能更高的实现形式提供指导。但利用DRAM(h)模型分析SpMV存储访问复杂度的工作以前没有人做过，并且SpMV的计算性能和存储访问行为跟具体的稀疏矩阵有关，只有到程序运行的时候才能知道。本文中，我们提出模板法和动态统计分析法两种分析SpMV存储访问复杂度的方法。在Pentium IV和Alpha EV6平台上，用RAM(h) 模型分析和计算了稀疏矩阵向量乘两种算法实现形式（即压缩行存储算法和寄存器分块算法）的存储访问复杂度，通过分析和计算在SpMV过程中需要访问的所有数据的存储访问复杂度，可知存储访问行为对整体程序的实际性能有直接影响。我们还在Pentium IV平台上，测试了七个稀疏矩阵的SpMV性能，并统计了两种算法中L1, L2,和TLB的缺失率，实验结果与模型分析的结果一致。

Controlled synthesis of L-lactide-b-epsilon-caprolactone block copolymers using a rare earth complex as catalyst

Well-defined block copolymers of L-lactide-b-epsilon-caprolactone were synthesized by sequential polymerization using a rare earth complex, Y(CF3COO)(3)/Al(iso-Bu)(3), as catalyst system. The compositions of the block copolymers could be adjusted by manipulating the feeding ratio of comonomers. The characterizations by GPC, H-1 NMR, C-13 NMR, and DSC displayed that the block copolymer, poly(epsilon-caprolactone-b-L-lactide) [P(CL-b-LLA)], had a narrow molecular weight distribution and well-controlled sequences without random placement.

COBALT PORPHYRIN NAFION FILM ON CARBON MICROARRAY ELECTRODE TO MONITOR OXYGEN FOR ENZYME ANALYSIS OF GLUCOSE

A microcarbon array electrode was modified by the placement of a Nafion film containing cobalt tetramethylpyridyl phorphyrin on its surface. This electrode was applied to the analysis of solution glucose when it was further modified by the immobilization of glucose oxidase on the outermost surface of the Nafion by the cross-linking of serum albumin with glutaraldehyde. The concomitant decrease in the concentration of oxygen, as it was consumed in the enzymatic reaction of glucose with glucose oxidase, was determined by either cyclic voltammetry or a double potential step method at the porphyrin-Nafion catalytic electrode. Glucose could be determined in the range of 0.01-4 mM rapidly, without interference from substances such as ascorbate or other saccharides.

基于加速度传感器的机器人极点配置控制

本文为工业机器人提出了一种极点配置控制法．这种控制方法的优点有：一是它的积分作用消除了机器人的微小扰动和稳态误差；二是能任意设置系统的极点，因此能保证闭环系统的稳定性和规定状态变量的暂态响应；三是加入了加速度反馈，抑制了由电枢电感所引起的机械手的振动．最后，给出了ＰＵＭＡ５６２机器人的计算机仿真和实验结果验证了此控制法的有效性。

广义预测极点配置前馈控制的一种算法

本文提出了广义预测极点配置前馈自校正控制算法,计算机仿真结果表明,该算法控制质量好,能够消除系统可测扰动对输出的影响。

湘西地区上震旦统上部－下寒武统底部黑色岩系沉积学及地球化学

Widespread black chert-shales occur in the Ediacaran-Cambrian(E-C) boundary successions along the flank of Yangtze Platform, South China, remarkable changes in sedimentology, geochemistry and biology were recorded. Although extensive studies were carried out upon this boundary succession, the origin of black chert-shales still remain controversial. This paper focuses on the E-C black chert-shales in western Hunan, South China, upon which detailed depositional and geochemical changes are documented, accordingly a depositional model for black chert-shales is proposed. Stratigraphic anatomy across the depositional strike demonstrates that the shallow-water Dengying dolostone along the platform margin sharply pass basinward into the Liuchapo chert successions, which indicate syndepositional extensional faulting at depth could have occurred along the platform margin. The deep-water Niutitang phosphorite-rich black shales are either underlain by the Dengying dolostones on the platform margin toward platform interior or directly by the Liuchaopo chert successions farther basinwards. By detailed investigation, silica chimneys are firsly identified approximately in the chert along platform margin; two types of silica chimneys, including mounded and splayed/funnelized chert(generally brecciated) bodies are further sorted out. The mounded chert are exitbited by domed or hummocky surfaces on the top and irregular spongy to digitiform internal fabrics; within the silica mounds, abundant original vesicles/voids and/or channels were mostly plugged by initial chalcedony, quartze crystals with minor dolomite and bladed barite crystals. Splayed/funnelized brecciated chert “intrusion” cross-cut the uppermost dolostones capping to the horizon underneath, and are directly overlain by the Niutitang phosphorite-rich black shales. Their similarities to the silica chimneys reported from the oceanic spreading centres suggest a similar origin responsible for these unique silica bodies which is also supported by the microthermonmetric data and element geochemistry. High P, Ba, Fe contents and positive correlation between Fe and TOC concentrations in the Niutitang black shales indicate a high palaeo-productivity in the Early Cambrian ocean. The low Th/U and the high V/Cr, V/Sc, V/(V+Ni) ratios in the black shales suggest an anoxic water condition during this interval. Furthermore, Positive Eu anomalies and high Ba contents in the sediments also imply a hydrothermal influence on the formation of Niutitang black shales. To better constrain the placement of deep-water successions straddling the E-C boundary and the timing of hydrothermal silica chimneys, sensitive high-resoluton ion microprobe(SHRIMP) U-Pb dating of zircon grains from tuffs within the chert succession of Liuchapo Formation at Ganziping was conducted and yields a weighted-mean 206Pb/238Pb age at 536.6±5.5Ma, younger than E-C boundary age(542.0±0.3Ma). This age combined with carbon isotopic data is then proposed to correspond to the U-Pb age of zircons(538.2±1.5Ma) from the Zhongyicun member of Meishucun Formation at Meishucun in eastern Yunna, thus, the E-C boundary in Gazngziping was placed between the Dengying formations and Liuchapo formatioms. therefore, the silica chimneys took place at the beginning of the Cambrian period. The temporal coincidence of silica chimneys and negative excursions of δ13C and δ34Spy pairs suggest hydrothermal activities were likely responsible for the isotopic changes. Under such a circumstance, vast amounts of greenhouse gases(CO2, CH4, H2S), with highly 13C-depleted carbon and 34S-depleted sulfur would be released into the ocean and atmosphere. A positive shift in δ34Scas and Δ34S values from the late Ediacaran to the Early Cambrian could be a reflection of enhanced bacterial sulfate reduction(BSR), strengthened by the intensified oceanic anoxia stimulated by hydrothermal activities. Based on the analyses of sedimentology and geochemistry, a model- “oceanic anoxia induced by hydrothermal–volcanic activies” was proposed to responsible for the formation of black chert-shales during this E-C transition. Under this case, hydrothermal-volcanic activies could release large large amount of greenhouse into atmosphere and metal micronutrients into the ocean, which may lead to global warming, stratified ocean, thereby a high palaeoproductivity; on the other hand, the massive releasing of reduced hydrothermal fluids with abundant H2S, could have in turn enhanced the ocean anoxia. All of these were favourable the for preservation of organic matter, and subsequent extensive deposition of black silica-shales.