Method and system for processing pixels utilizing scoreboarding
    1.
    发明授权
    Method and system for processing pixels utilizing scoreboarding 有权
    使用记分板处理像素的方法和系统

    公开(公告)号:US08854384B2

    公开(公告)日:2014-10-07

    申请号:US12953756

    申请日:2010-11-24

    摘要: In a graphics processing device, a plurality of processors write fragment shading results for order-dependent data to a buffer, according to the order in which the data is received. Fragment shading results for order-neutral data is written to the buffer one batch at a time. The order-dependent data comprises spatially overlapping data. Order-neutral data may not overlap. A scheduler controls the order of reception of one batch of data at a time by the processors. The order for receiving the order-dependent data may be determined. The plurality of processors may process the data in parallel. A writing order for writing results to a buffer from the processing in parallel, may be enforced. A portion of the processors may be instructed to wait before writing results to the buffer in a specified order. Processors signal when writing results to the buffer is complete.

    摘要翻译: 在图形处理装置中,多个处理器根据接收到数据的顺序将针对订单的数据写入缓冲器的片段着色结果。 一次性将批次中性数据的片段阴影结果写入缓冲区。 依赖于顺序的数据包括空间重叠的数据。 订单中性数据可能不重叠。 调度器一次由处理器控制一批数据的接收顺序。 可以确定接收订单相关数据的顺序。 多个处理器可以并行处理数据。 可以执行从处理并行写入结果到缓冲器的写入顺序。 可以指示处理器的一部分以指定的顺序将结果写入缓冲器之前等待。 当缓冲区写入结果完成时,处理器发出信号。

    Method And System For Controlling A 3D Processor Using A Control List In Memory
    2.
    发明申请
    Method And System For Controlling A 3D Processor Using A Control List In Memory 有权
    使用控制列表在内存中控制3D处理器的方法和系统

    公开(公告)号:US20110221743A1

    公开(公告)日:2011-09-15

    申请号:US12942626

    申请日:2010-11-09

    IPC分类号: G06T15/00

    CPC分类号: G06T15/005

    摘要: A graphics processing device generates control lists for controlling processing of 3D graphics. Control lists comprise primitive data for the 3D graphics, pointers to primitive data and control data. Tags that correspond to records within the control lists may be utilized for the control of processing. The graphics processing device may comprise a 3D pipeline comprising parallel processors. Processing is performed on a tile by tile basis and comprises a tile binning phase and/or a tile rendering phase. The two phases of processing may be performed in parallel on different sets of data. Control lists comprise a main list and/or sub-lists, for example, tile lists. Control lists may comprise links to other lists. Processing control may advance through the control lists without interaction from a driver and/or from a processor that is external to the graphics processing device.

    摘要翻译: 图形处理装置产生用于控制3D图形处理的控制列表。 控制列表包括3D图形的原始数据,指向原始数据和控制数据的指针。 对应于控制列表内的记录的标签可以用于处理的控制。 图形处理装置可以包括包括并行处理器的3D流水线。 处理是以瓦片为单位执行的,并且包括瓦片合并阶段和/或瓦片渲染阶段。 处理的两个阶段可以在不同的数据集上并行执行。 控制列表包括主列表和/或子列表,例如瓦片列表。 控制列表可以包括到其他列表的链接。 处理控制可以通过控制列表前进,而不需要来自驱动器和/或来自图形处理装置外部的处理器的交互。

    Method And System For Compressing Tile Lists Used For 3D Rendering
    3.
    发明申请
    Method And System For Compressing Tile Lists Used For 3D Rendering 有权
    压缩瓷砖列表的方法和系统用于3D渲染

    公开(公告)号:US20110216069A1

    公开(公告)日:2011-09-08

    申请号:US12953128

    申请日:2010-11-23

    IPC分类号: G06T11/20 G06T15/00

    CPC分类号: G06T11/20 G06T15/00

    摘要: A graphics processing device may generate coordinates for vertices of graphics primitives in a view-space. Tiles are defined within the view-space and are associated with tile lists. Primitives and/or vertices which overlap a tile are determined. Tile lists comprise differentially encoded indices and/or spatial coordinates for overlapping primitives. The differential encoding may or may not be byte aligned. During tile mode graphics rendering, tile lists are utilized to reference vertex attributes and/or primitives. Graphics rendering comprises a tile binning phase and a tile rendering phase. The primitives may comprise a triangle and/or joined triangles that share one or more vertices. For multiple joined primitives, information about shared vertices may be encoded without repetition for each primitive. Coordinates and/or corresponding weights for new vertices are encoded in a tile list and utilized for interpolating properties of the new vertices based on attributes of the original vertices.

    摘要翻译: 图形处理装置可以在视图空间中生成图形基元的顶点的坐标。 瓷砖在视图空间内定义,并与瓷砖列表相关联。 确定与瓦片重叠的基元和/或顶点。 平铺列表包含用于重叠基元的差分编码索引和/或空间坐标。 差分编码可以是或不是字节对齐。 在瓦片模式图形渲染期间,瓦片列表用于引用顶点属性和/或图元。 图形渲染包括瓦片合并阶段和瓦片渲染阶段。 原语可以包括共享一个或多个顶点的三角形和/或连接的三角形。 对于多个连接的原语,关于共享顶点的信息可以被编码而不对每个基元重复。 用于新顶点的坐标和/或对应的权重被编码在瓦片列表中,并且用于根据原始顶点的属性内插新顶点的属性。

    Method And System For Processing Pixels Utilizing Scoreboarding
    4.
    发明申请
    Method And System For Processing Pixels Utilizing Scoreboarding 有权
    处理使用记分板的像素的方法和系统

    公开(公告)号:US20110242113A1

    公开(公告)日:2011-10-06

    申请号:US12953756

    申请日:2010-11-24

    IPC分类号: G06F15/80

    摘要: In a graphics processing device, a plurality of processors write fragment shading results for order-dependent data to a buffer, according to the order in which the data is received. Fragment shading results for order-neutral data is written to the buffer one batch at a time. The order-dependent data comprises spatially overlapping data. Order-neutral data may not overlap. A scheduler controls the order of reception of one batch of data at a time by the processors. The order for receiving the order-dependent data may be determined. The plurality of processors may process the data in parallel. A writing order for writing results to a buffer from the processing in parallel, may be enforced. A portion of the processors may be instructed to wait before writing results to the buffer in a specified order. Processors signal when writing results to the buffer is complete.

    摘要翻译: 在图形处理装置中,多个处理器根据接收到数据的顺序将针对订单的数据写入缓冲器的片段着色结果。 一次性将批次中性数据的片段阴影结果写入缓冲区。 依赖于顺序的数据包括空间重叠的数据。 订单中性数据可能不重叠。 调度器一次由处理器控制一批数据的接收顺序。 可以确定接收订单相关数据的顺序。 多个处理器可以并行处理数据。 可以执行从处理并行写入结果到缓冲器的写入顺序。 可以指示处理器的一部分以指定的顺序将结果写入缓冲器之前等待。 当缓冲区写入结果完成时,处理器发出信号。

    Method and System For a Shader Processor With Closely-Coupled Peripherals
    5.
    发明申请
    Method and System For a Shader Processor With Closely-Coupled Peripherals 审中-公开
    具有紧耦合外设的着色器处理器的方法和系统

    公开(公告)号:US20110227920A1

    公开(公告)日:2011-09-22

    申请号:US12869900

    申请日:2010-08-27

    IPC分类号: G06T15/50

    CPC分类号: G06T15/005

    摘要: A method and system are provided in which a first instruction associated with a graphics rendering operation may be executed in a shader processor, the shader processor may receive result information associated with an intermediate portion of the graphics rendering operation performed by a peripheral device operably coupled to a register file bus in the shader processor, and the shader processor may execute a second instruction associated with the graphics rendering operation based on the received result information. The register file bus may be utilized for handling execution of intermediate instructions associated with the intermediate portion of the graphics rendering operation. The peripheral device may be accessed via one or more register file addresses associated with the peripheral device. The peripheral device may be operably coupled to the shader processor via a FIFO.

    摘要翻译: 提供了一种方法和系统,其中可以在着色器处理器中执行与图形呈现操作相关联的第一指令,着色器处理器可以接收与由外围设备执行的图形呈现操作的中间部分相关联的结果信息,该外围设备可操作地耦合到 着色器处理器中的寄存器文件总线,并且着色器处理器可以基于接收到的结果信息执行与图形呈现操作相关联的第二指令。 寄存器文件总线可以用于处理与图形呈现操作的中间部分相关联的中间指令的执行。 可以通过与外围设备相关联的一个或多个寄存器文件地址访问外围设备。 外围设备可以经由FIFO可操作地耦合到着色器处理器。

    Method and System For Tile Mode Renderer With Coordinate Shader
    6.
    发明申请
    Method and System For Tile Mode Renderer With Coordinate Shader 有权
    具有坐标着色器的平铺模式渲染器的方法和系统

    公开(公告)号:US20110148901A1

    公开(公告)日:2011-06-23

    申请号:US12868508

    申请日:2010-08-25

    IPC分类号: G09G5/02

    CPC分类号: G06T15/005 G06T11/40

    摘要: A method and system are provided in which one or more processors and/or circuits are operable to generate position information for a plurality of primitives utilizing a coordinate shader, one or more lists based on the generated position information, and rendering information for the plurality of primitives utilizing a vertex shader and the generated one or more lists. The generated one or more lists may comprise indices associated with one or more primitives from the plurality of primitives and with one or more tiles from a plurality of tiles in a screen plane. The position information and the one or more lists may be generated during a first rendering phase, and the rendering information may be generated during a second rendering phase different from the first rendering phase. The coordinate shader may perform a subset of the operations supported by the vertex shader.

    摘要翻译: 提供了一种方法和系统,其中一个或多个处理器和/或电路可操作以利用坐标着色器生成多个图元的位置信息,基于生成的位置信息生成一个或多个列表,以及针对多个 使用顶点着色器的基元和生成的一个或多个列表。 生成的一个或多个列表可以包括与来自多个图元的一个或多个图元相关联的索引以及来自屏幕平面中的多个瓦片的一个或多个瓦片。 可以在第一渲染阶段期间生成位置信息和一个或多个列表,并且可以在与第一渲染阶段不同的第二渲染阶段期间生成渲染信息。 坐标着色器可以执行顶点着色器支持的操作的一个子集。

    Method and system for controlling a 3D processor using a control list in memory
    7.
    发明授权
    Method and system for controlling a 3D processor using a control list in memory 有权
    用于使用存储器中的控制列表来控制3D处理器的方法和系统

    公开(公告)号:US09058685B2

    公开(公告)日:2015-06-16

    申请号:US12942626

    申请日:2010-11-09

    IPC分类号: G06T15/00

    CPC分类号: G06T15/005

    摘要: A graphics processing device generates control lists for controlling processing of 3D graphics. Control lists comprise primitive data for the 3D graphics, pointers to primitive data and control data. Tags that correspond to records within the control lists may be utilized for the control of processing. The graphics processing device may comprise a 3D pipeline comprising parallel processors. Processing is performed on a tile by tile basis and comprises a tile binning phase and/or a tile rendering phase. The two phases of processing may be performed in parallel on different sets of data. Control lists comprise a main list and/or sub-lists, for example, tile lists. Control lists may comprise links to other lists. Processing control may advance through the control lists without interaction from a driver and/or from a processor that is external to the graphics processing device.

    摘要翻译: 图形处理装置产生用于控制3D图形处理的控制列表。 控制列表包括3D图形的原始数据,指向原始数据和控制数据的指针。 对应于控制列表内的记录的标签可以用于处理的控制。 图形处理装置可以包括包括并行处理器的3D流水线。 处理是以瓦片为单位执行的,并且包括瓦片合并阶段和/或瓦片渲染阶段。 处理的两个阶段可以在不同的数据集上并行执行。 控制列表包括主列表和/或子列表,例如瓦片列表。 控制列表可以包括到其他列表的链接。 处理控制可以通过控制列表前进,而不需要来自驱动器和/或来自图形处理装置外部的处理器的交互。

    Method and system for tile mode renderer with coordinate shader
    8.
    发明授权
    Method and system for tile mode renderer with coordinate shader 有权
    具有坐标着色器的瓦片模式渲染器的方法和系统

    公开(公告)号:US08692848B2

    公开(公告)日:2014-04-08

    申请号:US12868508

    申请日:2010-08-25

    IPC分类号: G09G5/00

    CPC分类号: G06T15/005 G06T11/40

    摘要: A method and system are provided in which one or more processors and/or circuits are operable to generate position information for a plurality of primitives utilizing a coordinate shader, one or more lists based on the generated position information, and rendering information for the plurality of primitives utilizing a vertex shader and the generated one or more lists. The generated one or more lists may comprise indices associated with one or more primitives from the plurality of primitives and with one or more tiles from a plurality of tiles in a screen plane. The position information and the one or more lists may be generated during a first rendering phase, and the rendering information may be generated during a second rendering phase different from the first rendering phase. The coordinate shader may perform a subset of the operations supported by the vertex shader.

    摘要翻译: 提供了一种方法和系统,其中一个或多个处理器和/或电路可操作以利用坐标着色器生成多个图元的位置信息,基于生成的位置信息生成一个或多个列表,以及针对多个 使用顶点着色器的基元和生成的一个或多个列表。 生成的一个或多个列表可以包括与来自多个图元的一个或多个图元相关联的索引以及来自屏幕平面中的多个瓦片的一个或多个瓦片。 可以在第一渲染阶段期间生成位置信息和一个或多个列表,并且可以在与第一渲染阶段不同的第二渲染阶段期间生成渲染信息。 坐标着色器可以执行顶点着色器支持的操作的一个子集。

    Method and system for compressing tile lists used for 3D rendering
    9.
    发明授权
    Method and system for compressing tile lists used for 3D rendering 有权
    用于压缩用于3D渲染的瓦片列表的方法和系统

    公开(公告)号:US08619085B2

    公开(公告)日:2013-12-31

    申请号:US12953128

    申请日:2010-11-23

    IPC分类号: G06T11/20 G06T1/00

    CPC分类号: G06T11/20 G06T15/00

    摘要: A graphics processing device may generate coordinates for vertices of graphics primitives in a view-space. Tiles are defined within the view-space and are associated with tile lists. Primitives and/or vertices which overlap a tile are determined. Tile lists comprise differentially encoded indices and/or spatial coordinates for overlapping primitives. The differential encoding may or may not be byte aligned. During tile mode graphics rendering, tile lists are utilized to reference vertex attributes and/or primitives. Graphics rendering comprises a tile binning phase and a tile rendering phase. The primitives may comprise a triangle and/or joined triangles that share one or more vertices. For multiple joined primitives, information about shared vertices may be encoded without repetition for each primitive. Coordinates and/or corresponding weights for new vertices are encoded in a tile list and utilized for interpolating properties of the new vertices based on attributes of the original vertices.

    摘要翻译: 图形处理装置可以在视图空间中生成图形基元的顶点的坐标。 瓷砖在视图空间内定义,并与瓷砖列表相关联。 确定与瓦片重叠的基元和/或顶点。 平铺列表包含用于重叠基元的差分编码索引和/或空间坐标。 差分编码可以是或不是字节对齐。 在瓦片模式图形渲染期间,瓦片列表用于引用顶点属性和/或图元。 图形渲染包括瓦片合并阶段和瓦片渲染阶段。 原语可以包括共享一个或多个顶点的三角形和/或连接的三角形。 对于多个连接的原语,关于共享顶点的信息可以被编码而不对每个基元重复。 用于新顶点的坐标和/或对应的权重被编码在瓦片列表中,并且用于根据原始顶点的属性内插新顶点的属性。

    Method and system for decomposing complex shapes into curvy RHTs for rasterization
    10.
    发明授权
    Method and system for decomposing complex shapes into curvy RHTs for rasterization 有权
    将复杂形状分解为弯曲RHT以进行光栅化的方法和系统

    公开(公告)号:US08797325B2

    公开(公告)日:2014-08-05

    申请号:US12953739

    申请日:2010-11-24

    IPC分类号: G06T11/20

    CPC分类号: G06T11/40 G06T11/203

    摘要: A complex curved primitive is decomposed into curvy RHTs comprising a curved portion and horizontal and vertical lines. Pixel rows covered by curvy RHTs are determined. RHT pixels covered by the curved primitive are determined with counters. Tile based color rendering is performed for covered pixels. The primitive path is decomposed into Bezier curves. Curvy RHTs may overlap and may cover pixels that are not covered by the curved primitive. Pixel rows are located by traversing an RHT path and the direction of traversing may determine pixel counts. Pixel coverage and tile coverage information is stored in memory. Pixels may be rendered in parallel. The curved primitive is rendered in a tile binning phase and a tile rendering phase. Tile rendering comprises a pixel coverage first pass which determines pixels covered by said curved primitive and a color rendering second pass.

    摘要翻译: 复杂的弯曲原始体被分解为弯曲的RHT,其包括弯曲部分和水平和垂直线。 确定由弯曲RHT覆盖的像素行。 使用计数器确定弯曲原始图案覆盖的RHT像素。 对覆盖的像素执行基于平铺的颜色渲染。 原始路径被分解为贝塞尔曲线。 弯曲的RHT可以重叠并且可以覆盖未被弯曲原始图案覆盖的像素。 通过遍历RHT路径定位像素行,并且遍历方向可以确定像素数。 像素覆盖和瓦片覆盖信息存储在内存中。 像素可以并行渲染。 弯曲的原始图像呈现在瓦片合并阶段和瓦片渲染阶段。 平铺渲染包括确定由所述弯曲图元覆盖的像素和显色第二遍的像素覆盖率第一遍。