Parallelization of variable length decoding
AssigneeSTMicroelectronics International N.V.
InventorsSurinder Pal Singh, Aneesh Bhasin, Kaushik Saha
Life of this patent
Estimating the term…
Abstract
Parallelization of decoding of a data stream encoded with a variable length code includes determining one or more markers, each of which indicates a position within the encoded data stream. The determined markers are included into the encoded data stream together with the encoded data. At the decoder side, the markers are parsed from the encoded data stream and based on the extracted markers. The encoded data is separated into partitions, which are decoded separately and in parallel.
Claims15
A method for encoding input data into a data stream, the method comprising:
- encoding with a variable-length-code encoder, the input data using a variable length code;
- determining with a marker setting unit, marker data indicating a position within the input data encoded with the variable length code;
- including with a data stream generator, into the data stream the encoded input data and the determined marker data;
- and
- marking boundaries between partitions of entropy coded data by adding partition markers to the encoded data stream so that they are decoded independently, and so that their decoding results in an approximately equal number of computations on parallel decoders that enables load balancing in a parallel processing environment.
An encoder for encoding input data into a data stream, the encoder comprising:
- a variable length coder for encoding the input data using a variable length code;
- a marker setting unit for determining marker data indicating a position within the input data encoded with the variable length code;
- and
- a data steam generator for including into the data stream the encoded input data and the determined marker data,
- wherein the encoder is configured for marking boundaries between partitions of entropy coded data by adding partition markers to the encoded data stream so that they are decoded independently, and so that their decoding results in an approximately equal number of computations on parallel decoders that enables load balancing in a parallel processing environment.
Dependencies are inferred from the claim text.