申威处理器硬件数据预取技术的实现

硬件数据预取技术可以有效提升处理器的访存性能,是申威处理器性能优化过程中亟需突破的一项技术。硬件开销和处理器架构的制约是硬件预取技术实现中的主要难点。借鉴学术界对硬件预取技术的研究成果和工业界的应用现状,紧密结合申威处理器的结构特点,研究了申威处理器硬件预取技术的实现方法。以流预取为例,在处理器核心面积增加0.97%的情况下,硬件预取技术的应用可以将目前申威处理器的整数性能平均提升5.17%,最高提升28.88%;浮点性能平均提升6.39%,最高提升30.11%。     

巨型浮吊船电力系统动态响应的联想记忆识别

巨型浮吊海洋工程船的作业依赖于船舶电力系统的稳定运行。这类船上大功率变频器、大功率交流电机等设备的大量应用,使得控制系统受干扰影响严重,检测到的动态响应常常含有干扰成分,给电力自动化设备的反馈控制带来了困难。联想记忆具有按信息内容检索的功能,可以模仿人脑进行联想,处理缺损和不确定的信息,具有容错功能。根据联想记忆原理,运用Hopfield网络结构,对Hebb学习规则进行了改进,构建了自联想离散Hopfield网络。该网络通过对动态响应标准样本进行学习,得到了稳定模式,在受干扰污染、信息缺损的巨型浮吊船电力系统动态响应的识别中表现出了良好的滤除干扰的能力。     

Phase change memory alloys: GST cell array characterization using picosecond ultrasonics

Phase change random access memory alloys (PRAM or PCM) are a class of non-volatile memory that is thought as viable alternatives to flash memory technology or to supplement other memory technologies depending on the end applications and its key performance requirements. Ge₂Sb₂Te₅ alloy (GST) is the most widely used chalcogenide material for PCM application, and has many unique properties, including strong temperature-dependent film properties, low thermal conductivity, and high electrical resistivity. Picosecond ultrasonics was used to make non-contact, non-destructive measurements of GST films on blanket wafers and directly on product wafers. On-product wafer measurements were made on various via array (0.5 μm and 1 μm between cell edges with CD size from 250 to 800 nm). Measurements have shown excellent correlation to cross-section SEM and were consistent with CMP polish times for both blanket and pattern wafer measurement. Excellent repeatability based on extensive measurements demonstrates the capability and reliability of picosecond ultrasonic technology. Picosecond ultrasonic measurements also provide rapid characterization across the whole wafer at production-worthy throughputs.     

The thermochemical behavior of some binary shape memory alloys by high temperature direct synthesis calorimetry

The standard enthalpies of formation of some shape memory alloys have been measured by high temperature direct synthesis calorimetry at 1373 K. The following results (in kJ/mol of atoms) are reported: CoCr (−0.3 ± 2.9); CuMn (−3.7 ± 3.2); Cu₃Sn (−10.4 ± 3.1); Fe₂Tb (−5.5 ± 2.4); Fe₂Dy (−1.6 ± 2.9); Fe₁₇Tb₂ (−2.1 ± 3.1); Fe₁₇Dy₂ (−5.3 ± 1.7); FePd₃ (−16.0 ± 2.7); FePt (−23.0 ± 1.9); FePt₃ (−20.7 ± 2.3); NiMn (−24.9 ± 2.6); TiNi (−32.7 ± 1.0); TiPd (−60.3 ± 2.5). The results are compared with some earlier experimental values obtained by calorimetry and by EMF technique. They are also compared with predicted values on the basis of the semi empirical model of Miedema and co-workers and with ab initio calculations when available. We will also assess the available information regarding the structures of these alloys.     

Martensitic transformation of Ti50Ni30Cu20 alloy prepared by powder metallurgy

Phase transformation behavior of Ti50Ni30Cu20 shape memory alloys prepared by powder metallurgy is analyzed with respect to the duration of mechanical alloying. The processed blends were studied by differential scanning calorimetry and room temperature X-ray diffraction. The martensitic transformations evidenced by thermal scans are discussed in correlation with the relative phase content obtained from the refinement of the X-ray diffraction patterns.     

Regulation of memory accuracy with multiple answers: The plurality option.

We report two experiments that investigated the regulation of memory accuracy with a new regulatory mechanism: the plurality option. This mechanism is closely related to the grain-size option but involves control over the number of alternatives contained in an answer rather than the quantitative boundaries of a single answer. Participants were presented with a slideshow depicting a robbery (Experiment 1) or a murder (Experiment 2), and their memory was tested with five-alternative multiple-choice questions. For each question, participants were asked to generate two answers: a single answer consisting of one alternative and a plural answer consisting of the single answer and two other alternatives. Each answer was rated for confidence (Experiment 1) or for the likelihood of being correct (Experiment 2), and one of the answers was selected for reporting. Results showed that participants used the plurality option to regulate accuracy, selecting single answers when their accuracy and confidence were high, but opting for plural answers when they were low. Although accuracy was higher for selected plural than for selected single answers, the opposite pattern was evident for confidence or likelihood ratings. This dissociation between confidence and accuracy for selected answers was the result of marked overconfidence in single answers coupled with underconfidence in plural answers. We hypothesize that these results can be attributed to overly dichotomous metacognitive beliefs about personal knowledge states that cause subjective confidence to be extreme. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Collaborative remembering in older adults: Age-invariant outcomes in the context of episodic recall deficits.

Rapidly growing research reveals complex yet systematic consequences of collaboration on memory in young adults, but much less is known about this phenomenon in older adults. Young and older adults studied a list of categorized words and took three successive recall tests. Test 1 and 3 were always taken individually, and Test 2 was done either in triads or alone. Despite older adults recalling less overall than young adults, both age groups exhibited similar costs and benefits of collaboration: Collaboration reduced both correct and false recall during collaborative remembering, was associated with more positive beliefs about its value, and produced reminiscence, collective memory, and some forgetting in its cascading effects on postcollaborative recall. We examine the role of retrieval organization in these effects. As environmental support may play a substantial role in healthy aging, the relatively preserved effects of collaboration on memory in older adults hold promise for testing judicious uses of group remembering in aging. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

A core knowledge architecture of visual working memory: Correction to Wood (2011).

Reports an error in A core knowledge architecture of visual working memory by Justin N. Wood (Journal of Experimental Psychology: Human Perception and Performance, 2011[Apr], Vol 37[2], 357-381). The supplemental materials DOI is incorrect. The correct DOI for the supplemental materials is provided in the erratum. (The following abstract of the original article appeared in record 2011-06411-002.) Visual working memory (VWM) is widely thought to contain specialized buffers for retaining spatial and object information: a 'spatial-object architecture.' However, studies of adults, infants, and nonhuman animals show that visual cognition builds on core knowledge systems that retain more specialized representations: (1) spatiotemporal representations for object tracking, (2) object identity representations for object recognition, and (3) view-dependent snapshots for place recognition. In principle, these core knowledge systems may retain information separately from one another. Consistent with this hypothesis, this study provides evidence that these three types of information are subject to independent working memory storage limits. These results, combined with those from previous studies, indicate that VWM contains three specialized buffers for retaining spatiotemporal information, object identity information, and snapshot information. Thus, VWM buffers parallel core knowledge systems. This 'core knowledge architecture' links the study of visual working memory to the study of the biological foundations of visual cognition. (PsycINFO Database Record (c) 2011 APA, all rights reserved)