System of lanes of processing units receiving instructions via shared memory units for data-parallel or task-parallel operations
    3.
    发明授权
    System of lanes of processing units receiving instructions via shared memory units for data-parallel or task-parallel operations 有权
    处理单元的通道系统通过用于数据并行或任务并行操作的共享存储器单元接收指令

    公开(公告)号:US08180998B1

    公开(公告)日:2012-05-15

    申请号:US12208231

    申请日:2008-09-10

    IPC分类号: G06F15/16

    摘要: A system for performing data-parallel operations and task-parallel operations. A first switch fabric node (SFN) includes first and second lane processing engines (LPEs). The first LPE includes a first set of lane processing units (LPUs) configured to perform data-parallel operations, where each LPU performs a set of operations, and each LPU uses a different set of data for the set of operations, and each LPU within the first set of LPUs uses a different set of data for the set of operations. The second LPE includes a second set of LPUs configured to perform task-parallel operations, where each LPU performs a different set of operations. A processing control engine (PCE) is configured to distribute instructions and data to the first LPE and the second LPE. Advantageously, data parallel operations and task parallel operations are able to be performed on the same processor simultaneously.

    摘要翻译: 用于执行数据并行操作和任务并行操作的系统。 第一交换机节点(SFN)包括第一和第二通道处理引擎(LPE)。 第一LPE包括配置成执行数据并行操作的第一组通道处理单元(LPU),其中每个LPU执行一组操作,并且每个LPU对于该组操作使用不同的一组数据,并且每个LPU在 第一组接口板使用不同的数据集作为该组操作。 第二LPE包括被配置为执行任务并行操作的第二组LPU,其中每个LPU执行不同的操作集合。 处理控制引擎(PCE)被配置为将指令和数据分发到第一LPE和第二LPE。 有利地,数据并行操作和任务并行操作能够同时在同一处理器上执行。

    Physics processing unit
    4.
    发明授权
    Physics processing unit 有权
    物理处理单位

    公开(公告)号:US07895411B2

    公开(公告)日:2011-02-22

    申请号:US10715440

    申请日:2003-11-19

    IPC分类号: G06F19/00

    CPC分类号: G06T1/20

    摘要: One embodiment of the invention sets forth a hardware-based physics processing unit (PPU) having unique architecture designed to efficiently generate physics data. The PPU includes a PPU control engine (PCE), a data movement engine and a floating point engine (FPE). The PCE manages the overall operation of the PPU by allocating memory resources and transmitting graphics processing commands to the FPE and data movement commands to the DME. The FPE includes multiple vector processors that operate in parallel and perform floating point operations on data received from a host unit to generate physics simulation data. The DME facilitates the transmission of data between the host unit and the FPE by performs data movement operations between memories internal and external to the PPU.

    摘要翻译: 本发明的一个实施例阐述了一种基于硬件的物理处理单元(PPU),其具有设计成有效地生成物理数据的独特架构。 PPU包括PPU控制引擎(PCE),数据移动引擎和浮点引擎(FPE)。 PCE通过分配内存资源并向FPE发送图形处理命令和向DME发送数据移动命令来管理PPU的整体操作。 FPE包括并行操作的多个矢量处理器,并对从主机单元接收的数据进行浮点运算,以产生物理模拟数据。 DME通过在PPU内部和外部的存储器之间执行数据移动操作来促进主机和FPE之间的数据传输。

    Method and device for distributing bandwidth
    5.
    发明授权
    Method and device for distributing bandwidth 失效
    分配带宽的方法和设备

    公开(公告)号:US06810031B1

    公开(公告)日:2004-10-26

    申请号:US09515028

    申请日:2000-02-29

    IPC分类号: H04L1228

    摘要: A method and device for controlling bandwidth distribution through a switch fabric is provided wherein a plurality of line cards and processor cards are connected through a switch fabric for parallel processing of transmission requests, along with the provision of transmission “credits” allowing for transmitting additional data bytes during a given cycle, which provides efficient and speedy bandwidth distribution, as well as resolution of output contentions. The processors maintain a credit balance which allows flexibility in granting transmission requests to accommodate transmission scheduling and “bursty” transmissions. Processors on both of the line cards and the processor cards normalize the data transmission requirements for both inputs and outputs connected by the switch fabric. Smoothing of data transmission is provided using a time-weighted buffer.

    摘要翻译: 提供了一种用于控制通过交换结构的带宽分配的方法和设备,其中通过交换结构连接多个线路卡和处理器卡,用于传输请求的并行处理,以及提供允许发送附加数据的传输“信用” 字节在给定的周期,提供有效和快速的带宽分布,以及输出争用的分辨率。 处理器保持信用余额,允许灵活地授予传输请求以适应传输调度和“突发”传输。 两个线卡和处理器卡上的处理器标准化了由交换结构连接的输入和输出的数据传输要求。 使用时间加权缓冲器提供数据传输的平滑。

    Method of operation for parallel LCP solver
    8.
    发明授权
    Method of operation for parallel LCP solver 有权
    并行LCP求解器的操作方法

    公开(公告)号:US07526456B2

    公开(公告)日:2009-04-28

    申请号:US10793856

    申请日:2004-03-08

    IPC分类号: G06F15/00

    CPC分类号: G06F17/11

    摘要: A method of operating a Linear Complementarity Problem (LCP) solver is disclosed, where the LCP solver is characterized by multiple execution units operating in parallel to implement a competent computational method adapted to resolve physics-based LCPs in real-time.

    摘要翻译: 公开了一种操作线性互补问题(LCP)求解器的方法,其中LCP求解器的特征在于并行操作的多个执行单元,以实现适合于实时解决基于物理的LCP的胜任计算方法。