摘要:
An electronic device is provided which comprises a plurality of processing units (IP1-IP6) and a flit-synchronous network-based interconnect (N) for coupling the processing units (IP1-IP6). The network-based interconnect (N) comprises at least one first and at least one second link. The at least one second link comprises N pipeline stages. The communication via the at least one second link and the N pipeline stages constitutes a word-asynchronous communication.
摘要:
A processing system includes a processor and a physical memory (500) with a single-size memory port (505) for accessing data in the memory. The processor is arranged to operate on data of at least a first data size and a smaller second data size. The first data size is equal to or smaller than the size of memory port. The processing system including at least one data register (514) of the first data size connected to the memory port (505), and at least one data port (525) of the second data size connected to the data register (525) and the processor for enabling access to data elements of the second size.
摘要:
A processor includes a memory port for accessing a physical memory under control of an address. A processing unit executing instructions stored in the memory and/or operates on data stored in the memory. An address generation unit (“AGU”) generates address for controlling access to the memory; the AGU being associated with a plurality of N registers enabling the AGU to generate the address under control of an address generation mechanism. A memory unit is operative to save/load k of the N registers, where 2
摘要翻译:处理器包括用于在地址控制下访问物理存储器的存储器端口。 执行存储在存储器中的指令的处理单元和/或对存储在存储器中的数据进行操作。 地址生成单元(“AGU”)生成用于控制对存储器的访问的地址; AGU与多个N个寄存器相关联,使得AGU能够在地址生成机制的控制下生成地址。 一个存储器单元用于保存/加载N个寄存器的k,其中2 <= k <= N,由一个操作触发。 为此,存储器单元包括用于将k个寄存器与通过存储器端口写入存储器的一个存储器字串接的级联器和用于将通过存储器端口从存储器读取的字分离成k个寄存器的分离器。
摘要:
A scalar/vector processor includes a plurality of functional units (252, 260, 262, 264, 266, 268, 270). At least one of the functional units includes a vector section (210) for operating on at least one vector and a scalar section (220) for operating on at least one scalar. The vector section and scalar section of the functional unit co-operate by the scalar section being arranged to provide and/or consume at least one scalar required by and/or supplied by the vector section of the functional unit.
摘要:
The present invention relates to a memory circuit and a method of controlling data retention in the memory circuit, wherein a supply signal is selectively switched to a respective one of at least two virtual supply lines (24) each shared by a respective one of a plurality of groups (30-1 to 30-n) of memory cells (C0,0 to Cy,z). The selective switching is controlled based on a global activity control signal (A), used for setting the memory circuit either into a standby state or into an active state, and a local data retention indication signal (DR1 to DRn) allocated to a dedicated group of memory cells. Thereby, the data retention part of the memory circuit can be adapted to the application and its state, and standby mode leakaged power is only dissipated in those memory cells for which data retentions actually required.
摘要:
The state transition of a linear feedback shift register (LFSR) controlled by a clock (310) with length N and step size W, W being at least two, is accomplished via a next-state function (320). The next-state function deploys a state transition matrix (350). The state vector (330), which represents the contents of the LFSR, is either multiplied sequentially by the state transition matrix or multiplied by the state transition matrix to the power of W (multiple state transition matrix). The method and the LFSR according to the invention are characterized in that the multiple state transition matrix is decomposed in a first matrix (360) and a second matrix (370), the first matrix comprising at most N+W+1 different expressions and the second matrix comprising at most N+W+1 different expressions. The LFSR further comprises means to multiply the state vector by the second matrix and the first matrix, and means for computing the first matrix. The invention overcomes the shortcomings of configurable multi-step linear feedback shift registers because the amount of time needed to generate the output can be reduced significantly.
摘要翻译:通过下一状态功能(320)实现由具有长度N和步长W,W至少为2的时钟(310)控制的线性反馈移位寄存器(LFSR)的状态转换。 下一状态函数部署状态转换矩阵(350)。 代表LFSR的内容的状态向量(330)被状态转移矩阵顺序相乘或乘以状态转移矩阵到W的幂(多状态转移矩阵)。 根据本发明的方法和LFSR的特征在于,多状态转移矩阵在第一矩阵(360)和第二矩阵(370)中被分解,第一矩阵包括至多N + W + 1个不同表达式,并且 第二矩阵包括至多N + W + 1个不同表达式。 LFSR还包括将状态向量乘以第二矩阵和第一矩阵的装置,以及用于计算第一矩阵的装置。 本发明克服了可配置的多级线性反馈移位寄存器的缺点,因为可以显着地减少产生输出所需的时间量。
摘要:
An apparatus includes at least a first hardware part (AP1) and a second hardware part (AP2). Each of the first and second part include a respective processing element (CPU-1, CPU-2) and a respective signal connection to a respective memory element (MEM-1, MEM-2) for providing program code to the processing element of the respective part. The apparatus further includes a third hardware part (AP3) including at least one peripheral element acting as a source and/or destination of data. A fourth hardware part of the apparatus includes an I/O network (AP-4) for enabling communication between elements of the first and third part under control of first configuration data and for enabling communication between elements of the second and third part under control of distinct second configuration data.