Apple Inc.

Nvidia Corporation

Craig M Okruhlica

Brijesh Tripathi

Guy Cote

Nitin Bhargava

Edward Anthony Hutchins

Michael J M Toksvig

Tyson J Bergland

Justin Michael Mahan

Jim Chen Chou

Timothy John Millet

Peter F Holland

Joseph P Bratt

Wolfgang Roethig

Mark P Rygh

Joseph J Cheng

Gaurav S Gulati

Nitin Bhargawa

A block processing pipeline in which macroblocks are input to and processed according to row groups so that adjacent macroblocks on a row are not concurrently at adjacent stages of the pipeline. The input method may allow chroma processing to be postponed until after luma processing. One or more upstream stages of the pipeline may process luma elements of each macroblock to generate luma results such as a best mode for processing the luma elements. Luma results may be provided to one or more downstream stages of the pipeline that process chroma elements of each macroblock. The luma results may be used to determine processing of the chroma elements. For example, if the best mode for luma is an intra-frame mode, then a chroma processing stage may determine a best intra-frame mode for chroma and reconstruct the chroma elements according to the best chroma intra-frame mode.

Delayed chroma processing in block processing pipelines

A graphics processing circuit and method for power savings in the same is disclosed. In one embodiment, a graphics processing circuit includes a number of channels. The number of channels includes a number of color component channels that are each configured to process color components of pixel values of an incoming frame of graphics information. The number of channels also includes an alpha scaling channel configured to process alpha values (indicative of a level of transparency) for the incoming and/or outgoing frames. The graphics processing circuit also includes a control circuit. The control circuit is configured to place the alpha scaling channel into a low-power state responsive to determining that at least one of the incoming or outgoing frames does not include alpha values.

Alpha channel power savings in graphics unit

Block processing pipeline methods and apparatus in which reference data are stored to a memory according to tile formats to reduce memory accesses when fetching the data from the memory. When the pipeline stores reference data from a current frame being processed to memory as a reference frame, the reference samples are stored in macroblock sequential order. Each macroblock sample set is stored as a tile. Reference data may be stored in tile formats for luma and chroma. Chroma reference data may be stored in tile formats for chroma 4:2:0, 4:2:2, and/or 4:4:4 formats. A stage of the pipeline may write luma and chroma reference data for macroblocks to memory according to one or more of the macroblock tile formats in a modified knight's order. The stage may delay writing the reference data from the macroblocks until the macroblocks have been fully processed by the pipeline.

Data storage and access in block processing pipelines

The video encoders described herein may make an initial determination to designate a macroblock as a skip macroblock, but may subsequently reverse that decision based on additional information. For example, an initial skip mode decision may be based on aggregate distortion metrics for the luma component of the macroblock (e.g., SAD, SATD, or SSD), then reversed based on an individual pixel difference metric, an aggregate or individual pixel metric for a chroma component of the macroblock, or on the position of the macroblock within a macroblock row. The final skip mode decision may be based, at least in part, on the maximum difference between any pixel in the macroblock (or in a region of interest within the macroblock) and the corresponding pixel in a reference frame. The initial skip mode decision may be made during an early stage of a pipelined video encoding process and reversed in a later stage.

Skip thresholding in pipelined video encoders

A block processing pipeline in which blocks are input to and processed according to row groups so that adjacent blocks on a row are not concurrently at adjacent stages of the pipeline. A stage of the pipeline may process a current block according to neighbor pixels from one or more neighbor blocks. Since adjacent blocks are not concurrently at adjacent stages, the left neighbor of the current block is at least two stages downstream from the stage. Thus, processed pixels from the left neighbor can be passed back to the stage for use in processing the current block without the need to wait for the left neighbor to complete processing at a next stage of the pipeline. In addition, the neighbor blocks may include blocks from the row above the current block. Information from these neighbor blocks may be passed to the stage from an upstream stage of the pipeline.

Neighbor context processing in block processing pipelines

A system includes one or more video processing components and a display processing unit. The display processing unit may include one or more processing pipelines that generate read requests to fetch stored pixel data from a memory for subsequent display on a display unit. The display processing unit may also include a timing control unit that may generate an indication that indicates that the display unit will enter an inactive state. In response to receiving the indication, one or more of the video processing components may enter a low power state.

System and method of reducing power using a display inactive indication

Methods and apparatuses for performing lossless normalization of input pixel component values. The apparatus includes a normalization unit for converting pixel values from a range of 0 to (2&#x2212;1) to a normalized range from 0.0 to 1.0. The step size between adjacent values of the normalized range is 1/(2&#x2212;1), and a maximum input value of (2&#x2212;1) corresponds to a normalized value of 1. The normalization unit divides each input pixel component value by (2&#x2212;1) in order to preserve the fidelity of the color information contained in the input pixel component value.

Pixel normalization

Methods and graphics processing pipelines for performing inline chroma downsampling of pixel data. The graphics processing pipeline includes a chroma downsampling unit for performing buffer-free downsampling of chroma pixel components. A vertical column of chroma pixel components is received in each clock cycle by the chroma downsampling unit, and downsampled chroma pixel components are generated on every clock cycle or every other clock cycle. Vertical, horizontal, and vertical and horizontal downsampling can be performed without buffers by the chroma downsampling unit. A programmable configuration register in the chroma downsampling unit determines the type of downsampling that is implemented.

Performing inline chroma downsampling with reduced power consumption

A dither unit with a programmable kernel matrix in which each indexed location/entry may store one or more dither values. Each dither value in a respective entry of the kernel matrix may correspond to the number of bits that are truncated during dithering. During dithering of each pixel of an image, entries in the kernel matrix may be indexed according to the relative coordinates of the pixel within the image. A dither value for the pixel may be selected from the indexed entry based on the truncated least significant bits of the pixel component value. When the kernel matrix is storing more than one dither value per entry, the dither value may be selected based further on the number of truncated least significant bits. A dithered pixel component value may then be generated according to the dither value and the remaining most significant bits of the pixel component value.

Non-real-time dither using a programmable matrix

Methods and apparatus for performing an inline rotation of an image. The apparatus includes a rotation unit for reading pixels from a source image in an order based on a specified rotation to be performed. The source image is partitioned into multiple tiles, the tiles are processed based on where they will be located within the rotated image, and each tile is stored in a tile buffer. The target pixel addresses within a tile buffer are calculated and stored in a lookup table, and when the pixels are retrieved from the source image by the rotation unit, the lookup table is read to determine where to write the pixels within a corresponding tile buffer.

Inline image rotation

Sourcing immediate values from a very long instruction word includes determining if a VLIW sub-instruction expansion condition exists. If the sub-instruction expansion condition exists, operation of a portion of a first arithmetic logic unit component is minimized. In addition, a part of a second arithmetic logic unit component is expanded by utilizing a block of a very long instruction word, which is normally utilized by the first arithmetic logic unit component, for the second arithmetic logic unit component if the sub-instruction expansion condition exists.

Techniques for sourcing immediate values from a VLIW

A parallel scaler unit for simultaneously scaling multiple pixels from a source image. The scaler unit includes multiple vertical scalers and multiple horizontal scalers. A column of pixels from the source image is presented to the vertical scalers, and each vertical scaler selects appropriate pixels from the column of pixels for scaling. Each vertical scaler scales the selected pixels in a vertical direction and then conveys the vertically scaled pixels to a corresponding horizontal scaler. Each horizontal scaler scales the received pixels in a horizontal direction.

Parallel scaler processing

An arithmetic logic stage in a graphics processor unit includes arithmetic logic units (ALUs) and global registers. The registers contain global values for a group of pixels. Global values may be read from any of the registers, regardless of which of the pixels is being operated on by the ALUs. However, when writing results of the ALU operations, only some of the global registers are candidates to be written to, depending on the pixel number. Accordingly, overwriting of data is prevented.

Shared readable and writeable global values in a graphics processor unit pipeline

An arithmetic logic stage in a graphics processor unit pipeline includes a number of arithmetic logic units (ALUs) and at least one buffer that stores pixel data for a group of pixels. Each clock cycle, the buffer stores one row of a series of rows of pixel data. A deserializer deserializes the rows of pixel data before the pixel data is placed in the buffer. After the buffer accumulates all rows of pixel data for a pixel, then the pixel data for the pixel can be operated on by the ALUs.

Growing community of inventors

San Jose, CA, United States of America

Craig M Okruhlica