-
公开(公告)号:US09111369B2
公开(公告)日:2015-08-18
申请号:US13782081
申请日:2013-03-01
Applicant: Broadcom Corporation
Inventor: Alexander G. MacInnis , Chengfuh Jeffrey Tang , Xiadong Xie , James T. Patterson , Greg A. Kranawetter
IPC: G06T1/60 , G06T1/20 , G06T9/00 , G09G5/00 , G09G5/06 , G09G5/12 , G09G5/14 , G09G5/28 , G09G5/34 , G09G5/36 , H04N5/04 , H04N5/14 , H04N5/445 , H04N5/45 , H04N7/01 , H04N9/45 , H04N9/64 , H04N21/422 , H04N21/426 , H04N21/43 , H04N21/431 , H04N21/434 , H04N21/4402 , H04N21/443 , H04N21/47 , H04N19/61 , H04N19/44 , H04N19/42 , H04N19/423 , H04N19/436 , H04N19/59 , G09G5/02 , H04N5/12 , H04N5/44 , H04N5/46 , H04N11/14 , H04N11/20
CPC classification number: G06F3/0611 , G06F3/0653 , G06F3/0659 , G06F3/0683 , G06T1/20 , G06T1/60 , G06T9/00 , G06T9/007 , G09G5/001 , G09G5/02 , G09G5/024 , G09G5/026 , G09G5/06 , G09G5/12 , G09G5/14 , G09G5/28 , G09G5/346 , G09G5/36 , G09G5/363 , G09G2310/0224 , G09G2320/0247 , G09G2340/02 , G09G2340/0407 , G09G2340/10 , G09G2340/125 , G09G2360/02 , G09G2360/121 , G09G2360/125 , G09G2360/126 , G09G2360/128 , H04N5/04 , H04N5/126 , H04N5/14 , H04N5/4401 , H04N5/44504 , H04N5/44508 , H04N5/44591 , H04N5/45 , H04N5/46 , H04N7/0122 , H04N7/0135 , H04N9/45 , H04N9/641 , H04N9/642 , H04N11/143 , H04N11/20 , H04N19/42 , H04N19/423 , H04N19/436 , H04N19/44 , H04N19/59 , H04N19/61 , H04N21/42204 , H04N21/42653 , H04N21/4305 , H04N21/4316 , H04N21/434 , H04N21/440263 , H04N21/4438 , H04N21/47
Abstract: Disclosed herein are various embodiments of a graphics accelerator, which may include an integrated circuit. The integrated circuit may include a local memory; a direct memory access (DMA) engine; a processor; and one or more processing pipelines. The local memory stores graphics data that includes a plurality of pixels. The DMA engine transfers the graphics data between the local memory and an external memory. The processor performs at least one operation, in parallel, on components of at least a portion of the pixels. The one or more processing pipelines process the graphics data. The graphics accelerator works on operands and produces outputs for one set of pixels while the DMA engine is bringing in operands for a future set of pixel operations, and transfers data from the external memory to the one or more processing pipelines by directing data to the one or more pipelines.
Abstract translation: 这里公开了图形加速器的各种实施例,其可以包括集成电路。 集成电路可以包括本地存储器; 直接存储器访问(DMA)引擎; 处理器 和一个或多个处理管道。 本地存储器存储包括多个像素的图形数据。 DMA引擎在本地存储器和外部存储器之间传输图形数据。 处理器并行地执行至少一部分像素的组件上的至少一个操作。 一个或多个处理管线处理图形数据。 图形加速器对操作数起作用,并产生一组像素的输出,而DMA引擎为未来的像素运算组提供操作数,并通过将数据指向一个或多个处理流水线将数据从外部存储器传输到一个或多个处理流水线 或更多管道。
-
公开(公告)号:US20140078155A1
公开(公告)日:2014-03-20
申请号:US13782081
申请日:2013-03-01
Applicant: Broadcom Corporation
Inventor: Alexander G. MacInnis , Chengfuh Jeffrey Tang , Xiadong Xie , James T. Patterson , Greg A. Kranawetter
IPC: G06T1/20
CPC classification number: G06F3/0611 , G06F3/0653 , G06F3/0659 , G06F3/0683 , G06T1/20 , G06T1/60 , G06T9/00 , G06T9/007 , G09G5/001 , G09G5/02 , G09G5/024 , G09G5/026 , G09G5/06 , G09G5/12 , G09G5/14 , G09G5/28 , G09G5/346 , G09G5/36 , G09G5/363 , G09G2310/0224 , G09G2320/0247 , G09G2340/02 , G09G2340/0407 , G09G2340/10 , G09G2340/125 , G09G2360/02 , G09G2360/121 , G09G2360/125 , G09G2360/126 , G09G2360/128 , H04N5/04 , H04N5/126 , H04N5/14 , H04N5/4401 , H04N5/44504 , H04N5/44508 , H04N5/44591 , H04N5/45 , H04N5/46 , H04N7/0122 , H04N7/0135 , H04N9/45 , H04N9/641 , H04N9/642 , H04N11/143 , H04N11/20 , H04N19/42 , H04N19/423 , H04N19/436 , H04N19/44 , H04N19/59 , H04N19/61 , H04N21/42204 , H04N21/42653 , H04N21/4305 , H04N21/4316 , H04N21/434 , H04N21/440263 , H04N21/4438 , H04N21/47
Abstract: Disclosed herein are various embodiments of a graphics accelerator, which may include an integrated circuit. The integrated circuit may include a local memory; a direct memory access (DMA) engine; a processor; and one or more processing pipelines. The local memory stores graphics data that includes a plurality of pixels. The DMA engine transfers the graphics data between the local memory and an external memory. The processor performs at least one operation, in parallel, on components of at least a portion of the pixels. The one or more processing pipelines process the graphics data. The graphics accelerator works on operands and produces outputs for one set of pixels while the DMA engine is bringing in operands for a future set of pixel operations, and transfers data from the external memory to the one or more processing pipelines by directing data to the one or more pipelines.
Abstract translation: 这里公开了图形加速器的各种实施例,其可以包括集成电路。 集成电路可以包括本地存储器; 直接存储器访问(DMA)引擎; 处理器 和一个或多个处理管道。 本地存储器存储包括多个像素的图形数据。 DMA引擎在本地存储器和外部存储器之间传输图形数据。 处理器并行地执行至少一部分像素的组件上的至少一个操作。 一个或多个处理管线处理图形数据。 图形加速器对操作数起作用,并产生一组像素的输出,而DMA引擎为未来的像素运算组提供操作数,并通过将数据指向一个或多个处理流水线将数据从外部存储器传输到一个或多个处理流水线 或更多管道。
-