基于NR-Transformer的集群作业运行时间预测

高性能集群的作业调度通常使用作业调度系统来实现，准确填写作业运行时间能在很大程度上提升作业调度效率。现有的研究通常使用机器学习的预测方式，在预测精度和实用性上还存在一定的提升空间。为了进一步提高集群作业运行时间预测的准确率，考虑先对集群作业日志进行聚类，将作业类别信息添加到作业特征中，再使用基于注意力机制的NR-Transformer网络对作业日志数据建模和预测。在数据处理上，根据与预测目标的相关性、特征的完整性和数据的有效性，从历史日志数据集中筛选出7维特征，并按作业运行时间的长度将其划分为多个作业集，再对各作业集分别进行训练和预测。实验结果表明，相比于传统机器学习和BP神经网络，时序神经网络结构有更好的预测性能，其中NR-Transformer在各作业集上都有较好的性能。     

格点QCD基础求解器及其异构计算实现的性能优化

格点量子色动力学(格点QCD)是研究夸克、胶子等微观粒子间相互作用的重要理论和方法. 通过将时空离散化为四维结构网格, 并将量子色动力学的基本场量定义在网格上, 让研究人员可以使用数值模拟方法, 从第一性原理出发研究强子间相互作用和性质, 但这个过程中的计算量极大, 需要进行大规模并行计算. 格点QCD计算的核心基础为格点QCD求解器, 是程序运行主要的计算热点模块. 本文研究在国产异构计算平台下格点QCD求解器的实现与优化, 提出一套格点QCD求解器的设计实现, 实现了BiCGSTAB求解器, 显著降低了迭代次数; 通过对奇偶预处理技术, 降低了所求问题的计算规模; 针对国产异构加速卡的特点, 优化了Dslash模块的访存操作. 实验测试表明, 相比优化前的求解器获得了约30倍的加速比, 为国产异构超算下格点QCD软件性能优化提供了有益的参考价值.     

Mixing characteristics for the single and air-water two-phase flows in miniaturized parallel multichannel-based devices

Investigation on the miniaturized parallel multichannel-based devices packed with glass beads to improve the mass exchange execution is the critical focal point of the current study. One of the essential parameters to specify the miniaturized devices' flow distribution is the residence time distribution (RTD). In the present context, the RTDs of a liquid tracer were investigated for the air-water multiphase flows (concurrent) across the multichannel-based miniaturized devices (comprising of 11 similar dimensional parallel channels). The devices were variable in height and packed with glass beads. The conductivity estimations generated the RTD curves and were addressed by the axial dispersion model (ADM). The fluid-flow rates differed within the range of 5–23 ml min⁻¹. The axial dispersion coefficients and the rate of the specific energy dispersion were investigated. The effects of pressure difference and geometry on the hydrodynamic attributes and mixing properties were well-illustrated, and the new correlations were suggested.     

Designing amorphous formulations of polyphenols with naringin by spray-drying for enhanced solubility and permeability

Naringin (NAR), a major flavanone (FVA) glycoside, is a component of food mainly obtained from grapefruit. We used NAR as a food additive to improve the solubility and permeability of hydrophobic polyphenols used as supplements in the food industry. The spray-dried particles (SDPs) of NAR alone show an amorphous state with a glass transition temperature (T_g) at 93.2 °C. SDPs of hydrophobic polyphenols, such as flavone (FVO), quercetin (QCT), naringenin (NRG), and resveratrol (RVT) were prepared by adding varying amounts of NAR. All SDPs of hydrophobic polyphenols with added NAR were in an amorphous state with a single T_g, but SDPs of hydrophobic polyphenols without added NAR showed diffraction peaks derived from each crystal. The SDPs with NAR could keep an amorphous state after storage at a high humidity condition for one month, except for SDPs of RVT/NAR. SDPs with NAR enhanced the solubility of hydrophobic polyphenols, especially NRG solubility, which was enhanced more than 9 times compared to NRG crystal. The enhanced solubility resulted in the increased membrane permeability of NRG. The antioxidant effect of the hydrophobic NRG was also enhanced by the synergetic effect of NAR. The findings demonstrated that NAR could be used as a food additive to enhance the solubility and membrane permeability of hydrophobic polyphenols.     

接触器触头系统振动机理分析及算法优化

为了解决合闸过程中动、静触头接触引起的振动弹跳问题.本文建立了接触系统的二自由度运动微分方程,并利用遗传算法对交流接触器吸合过程进行优化,同时通过高速摄影实验对接触弹跳的全过程进行了观察和分析.结果表明:理论与实验结果高度一致,铁心在触头分离前发生碰撞,进一步加剧触头弹跳;铁芯弹跳再次碰撞时,触头的弹跳不受影响;在接触器运行过程中,动铁芯的运动会引起系统轻微振动;采用遗传算法优化的接触器触头弹跳时间和最大振幅均减小.研究结果为进一步控制和减小接触弹跳提供了理论依据.     

平面并联机构工作空间的三维螺旋扫描CAD求解

针对平面并联机构无奇异位置工作空间求解困难、过程繁琐、计算量大等问题，提出了基于CAD求解平面并联机构工作空间的三维螺旋扫描方法。将[n]自由度平面并联机构分解成[n]条支链进行独立分析，得到每条支链下末端执行器的可达区域，再将所有支链可达区域取交集即为平面并联机构工作空间。应用SolidWorks软件建立平面并联机构模型，进行几何特征处理，通过自动求解器求解，将求解过程图形化，快速得到同轴布局5R机构和平面3-RPR并联机构的无奇异位置工作空间。通过同轴布局5R机构的运动学实验，验证了该求解方法的可行性。     

Full Eulerian deformable solid-fluid interaction scheme based on building-cube method for large-scale parallel computing

We propose a full Eulerian incompressible solid-fluid interaction scheme capable of achieving high parallel efficiency and easily generating meshes for complex solid geometries. While good scalability of a full Eulerian solid-fluid interaction formulation has been reported by Sugiyama et al, their analysis was carried out using uniform Cartesian mesh and an artificial compressibility method. Typically, it is more challenging to achieve good scalability for hierarchical Cartesian meshes and a fully incompressible formulation. In addition, the conventional full Eulerian methods require a large computational cost to resolve complex solid geometries due to the usage of uniform Cartesian meshes. In an attempt to overcome the aforementioned issues, we employ the building-cube method, where the computational domain is divided into cubic regions called cubes. Each cube is divided at equal intervals, the same number of cubes is assigned to each core, and the spatial loop processing is executed for each cube. The numerical method is verified by computing five numerical examples. In the weak scaling test, the parallel efficiency at 32768 cores with 32 cores as a reference is 93.6%. In the strong scaling test, the parallel efficiency at 32768 cores with 128 cores as a reference is 70.2%.     

雷达发射机浮动板调制器与故障检测电路的设计

提出了一种改进的浮动板调制器和对正负偏电压的故障检测电路,利用MOSFET寄生电容特性,通过固定脉宽窄脉冲控制调制脉宽,通过增加负偏MOSFET提高输出负偏电压。对故障检测电路进行仿真实验,通过模拟故障,验证故障检测报警信号发生时间均在微秒级别。完成样机的搭建,利用多重方式解决高压绝缘问题,并对其加电进行试验验证,分析实验波形,证明此电路满足设计要求并且具有可行性。     

移动物料信息分布式传输方法及应用

在流程工业生产中,移动物料计量信息的传输普遍采用电台和无线模块。接收电台周围环境严重影响计量信息的传输效果,尤其是钢结构厂房、电磁干扰区内,数据传输过程中的丢包率、误码率相当严重。本文提出一种多节点分布式传输方法,根据各个生产作业区环境条件设置相应的WIFI节点数,实现对作业区的无线信号覆盖。网络覆盖区内无线模块、PDA终端(PDA)、终端查询设备通过节点工作频段和阵列天线自适应调节,实现数据传输最佳化。铜熔炼生产现场测试表明,此种传输方法可以有效消除周围环境的电磁干扰,数据丢包率低于0.1%。     

Compiler compatible 5.66 Mb/mm2 8T 1R1W register file in 14 nm FinFET technology

1-read/1-write (1R1W) register file (RF) is a popular memory configuration in modern feature rich SoCs requiring significant amount of embedded memory. A memory compiler is constructed using the 8T RF bitcell spanning a range of instances from 32 b to 72 Kb. An 8T low-leakage bitcell of 0.106 μm² is used in a 14 nm FinFET technology with a 70 nm contacted gate pitch for high-density (HD) two-port (TP) RF memory compiler which achieves 5.66 Mb/mm² array density for a 72 Kb array which is the highest reported density in 14 nm FinFET technology. The density improvement is achieved by using techniques such as leaf-cell optimization (eliminating transistors), better architectural planning, top level connectivity through leaf-cell abutment and minimizing the number of unique leaf-cells. These techniques are fully compatible with memory compiler usage over the required span. Leakage power is minimized by using power-switches without degrading the density mentioned above. Self-induced supply voltage collapse technique is applied for write and a four stack static keeper is used for read Vmin improvement. Fabricated test chips using 14 nm process have demonstrated 2.33 GHz performance at 1.1 V/25 °C operation. Overall V_min of 550 mV is achieved with this design at 25 °C. The inbuilt power-switch improves leakage power by 12x in simulation. Approximately 8% die area of a leading 14 nm SoC in commercialization is occupied by these compiled RF instances.