摘要：【中文】本发明公开了一种基于时间序列预测的海量小文件存储性能优化方法，属于信息存储领域，包括：收集带有时间信息的历史文件访问记录，得到数据集；将数据集预处理为离散的时间序列数据后，使用时间窗口在其上滚动产生训练数据集合，任意t时刻的训练数据以t‑n～t时刻的数据为输入数据，并以t+1时刻的数据为标签数据；基于循环神经网络建立时间序列预测模型，并利用由训练数据集合划分得到的训练集、验证集和测试集依次进行训练、验证与测试，从而得到目标模型；利用目标模型预测文件大小的变化趋势，以识别出其中的大文件和小文件；将大文件直接存储，将小文件基于时间序列聚合后存储。本发明能够优化海量小文件在分布式存储系统中的存储性能。 【EN】The invention discloses a method for optimizing the storage performance of a large number of small files based on time sequence prediction, which belongs to the field of information storage and comprises the following steps: collecting historical file access records with time information to obtain a data set; preprocessing a data set into discrete time sequence data, rolling the discrete time sequence data on the discrete time sequence data by using a time window to generate a training data set, wherein the training data at any time t takes the data at the time t-n-t as input data and takes the data at the time t +1 as label data; establishing a time sequence prediction model based on a recurrent neural network, and sequentially training, verifying and testing a training set, a verifying set and a testing set which are obtained by dividing a training data set so as to obtain a target model; predicting the change trend of the file size by using the target model so as to identify a large file and a small file in the file; the large files are directly stored, and the small files are aggregated based on time series and then stored. The invention can optimize the storage performance of the mass small files in the distributed storage system.

摘要：【中文】本发明公开了一种卷积神经网络加速引擎、卷积神经网络加速系统及方法，属于异构计算加速领域，引擎包括：全局缓存和物理PE矩阵；物理PE矩阵包括多个物理PE单元，物理PE单元用于执行行卷积操作以及相关的部分和累加操作；XY互联总线用于将输入特征图像数据、输出特征图像数据和卷积核参数从全局缓存传输到物理PE矩阵上，或者将物理PE矩阵产生的操作结果传输到全局缓存中；邻接互联总线用于在同一列物理PE单元之间传输中间结果；系统包括：3D‑Memory，其中每个Vault单元的内存控制器中集成了卷积神经网络加速引擎，用于完成卷积神经网络计算任务的一个子集；方法在系统的基础上进行逐层优化。本发明能够提高卷积神经网络的性能和能耗。 【EN】The invention discloses a convolutional neural network acceleration engine, a convolutional neural network acceleration system and a convolutional neural network acceleration method, which belong to the field of heterogeneous computation acceleration, and the engine comprises: a global cache and a physical PE matrix; the physical PE matrix comprises a plurality of physical PE units, and the physical PE units are used for executing row convolution operation and related partial sum accumulation operation; the XY interconnection bus is used for transmitting the input characteristic image data, the output characteristic image data and the convolution kernel parameter from the global cache to the physical PE matrix, or transmitting an operation result generated by the physical PE matrix to the global cache; the adjacent interconnection buses are used for transmitting intermediate results among the physical PE units in the same column; the system comprises: the system comprises a 3D-Memory, wherein a convolutional neural network acceleration engine is integrated in a Memory controller of each Vault unit and is used for completing a subset of convolutional neural network calculation tasks; the method performs layer-by-layer optimization on the basis of the system. The invention can improve the performance and energy consumption of the convolutional neural network.

摘要：【中文】本发明公开了一种多通道X射线光机能谱测量系统及测量方法，本发明中的剂量探头设计独特，采用N层吸收材料层、N个探测器圆片错位分布的结构，由于不同吸收材料层位置的剂量率不同，因此不同层的探测器圆片会产生不同强度的荧光信号，荧光信号经光纤阵列传输到相应的硅光电二极管阵列，其信号再经放大电路放大后，由数据采集卡送入PC机中处理，最终获得光谱的原级谱分布。本发明还通过激光器发出的激光开展光释光测量，用于测量x射线在探测器圆片中产生的累积剂量。且本发明能获得辐射发光和光释光信号，由此可以同时获得实时辐射剂量和累积剂量，同时用于计算X射线能谱分布，两者的结果可用于比较和分析。 【EN】The invention discloses a multichannel X-ray optical machine energy spectrum measuring system and a measuring method, wherein a dosage probe is uniquely designed, a structure that N layers of absorption material layers and N detector wafers are distributed in a staggered manner is adopted, because dosage rates at different positions of the absorption material layers are different, the detector wafers at different layers can generate fluorescence signals with different intensities, the fluorescence signals are transmitted to corresponding silicon photodiode arrays through optical fiber arrays, the signals are amplified through an amplifying circuit and then sent into a PC (personal computer) through a data acquisition card for processing, and the primary spectrum distribution of a spectrum is finally obtained. The invention also provides a photoluminescence measurement by laser emitted by the laser, which is used for measuring the cumulative dose of the x-ray generated in the detector wafer. And the invention can obtain the radiation luminescence and luminescence signals, thereby simultaneously obtaining the real-time radiation dose and the accumulated dose, and simultaneously calculating the X-ray energy spectrum distribution, and the results of the two can be used for comparison and analysis.

摘要：【中文】一种膨胀循环液氧甲烷上面级发动机系统，包括推进剂供应系统、点火系统和推力室。推进剂供应系统包括甲烷供应系统和液氧供应系统，甲烷供应系统对来自外部贮箱的低温液甲烷进行增压处理后提供给推力室；液氧供应系统对来自外部贮箱的低温液氧进行增压处理后提供给推力室；点火系统位于推力室头部，在控制系统控制下点火，引燃进入推力室的液氧和甲烷，产生的高温燃气从推力室喷口喷出，产生推力。本发明以液氧和液甲烷作为推进剂组合，采用闭式膨胀循环系统方案，系统简单、固有可靠性高、比冲性能高、易于实现多次起动。 【EN】An expansion cycle liquid oxymethane upper stage engine system includes a propellant supply system, an ignition system, and a thrust chamber. The propellant supply system comprises a methane supply system and a liquid oxygen supply system, wherein the methane supply system is used for pressurizing low-temperature liquid methane from an external storage tank and then supplying the low-temperature liquid methane to the thrust chamber; the liquid oxygen supply system is used for pressurizing low-temperature liquid oxygen from the external storage tank and then supplying the low-temperature liquid oxygen to the thrust chamber; the ignition system is positioned at the head of the thrust chamber, the ignition system ignites under the control of the control system, liquid oxygen and methane entering the thrust chamber are ignited, and the generated high-temperature fuel gas is sprayed out from a nozzle of the thrust chamber to generate thrust. The invention takes liquid oxygen and liquid methane as propellant combination, adopts a scheme of a closed expansion circulating system, has simple system, high inherent reliability and high specific impulse performance, and is easy to realize multiple starting.

摘要：【中文】本发明公开了一种基于辐射粒子事件的采集器，通过探测器将探测到的粒子事件转化为电脉冲信号，电脉冲信号经过前置放大电路后分为两路，分别进行时间甄别和能量甄别，时间信息和能量信息在FPGA中打包成辐射粒子事件信息包，并传输到上位机中处理。本发明还公开了一种辐射粒子事件快速核素识别方法，建立某种放射性核素的特征伽马射线粒子事件模型和在环境本底下的粒子事件模型，指定判别函数及判别阈值，对采集到的粒子事件进行筛选、计算核素的特征粒子事件模型参数并更新判别函数值，将决策函数值与设定的判别阈值作比较，如果高于高阈值，判定放射性核素存在，如果小于低阈值，则判定放射性核素不存在，否则重复前面过程。 【EN】The invention discloses a collector based on radiation particle events, which converts detected particle events into electric pulse signals through a detector, the electric pulse signals are divided into two paths after passing through a preamplifier circuit and are respectively subjected to time discrimination and energy discrimination, and time information and energy information are packaged into radiation particle event information packets in an FPGA (field programmable gate array) and are transmitted to an upper computer for processing. The invention also discloses a method for quickly identifying nuclides of radiation particle events, which comprises the steps of establishing a characteristic gamma ray particle event model of a certain radionuclide and a particle event model under the environment background, appointing a discrimination function and a discrimination threshold value, screening the collected particle events, calculating characteristic particle event model parameters of the nuclides, updating the discrimination function value, comparing the decision function value with the set discrimination threshold value, judging that the radionuclide exists if the decision function value is higher than the high threshold value, judging that the radionuclide does not exist if the decision function value is lower than the low threshold value, and repeating the previous process if the decision function value is not higher than the high threshold value.