US Patent Application 18340464. METHOD AND APPARATUS FOR COMPRESSING AND DECOMPRESSING SPARSE DATA SETS simplified abstract
Contents
METHOD AND APPARATUS FOR COMPRESSING AND DECOMPRESSING SPARSE DATA SETS
Organization Name
Microsoft Technology Licensing, LLC
Inventor(s)
Mattheus C. Heddes of Redmond WA (US)
Ankit More of San Mateo CA (US)
Nishit Shah of Sunnyvale CA (US)
Torsten Hoefler of Zurich (CH)
METHOD AND APPARATUS FOR COMPRESSING AND DECOMPRESSING SPARSE DATA SETS - A simplified explanation of the abstract
- This abstract for appeared for US patent application number 18340464 Titled 'METHOD AND APPARATUS FOR COMPRESSING AND DECOMPRESSING SPARSE DATA SETS'
Simplified Explanation
The present disclosure describes a digital circuit and method for compressing and decompressing digital data values. The compression algorithm used is a multi-stage process, where the compressed values are stored in a memory. The decompression circuit receives these values and performs a partial decompression. The partially decompressed values are then passed on to a processor, which completes the final decompression. In one specific embodiment, a vector of N length compressed values is decompressed using a first bit mask, resulting in two sets of N length with non-zero values. These two sets are further decompressed using two M length bit masks, resulting in M length sparse vectors, each containing non-zero values.
Original Abstract Submitted
Embodiments of the present disclosure include a digital circuit and method for multi-stage compression. Digital data values are compressed using a multi-stage compression algorithm and stored in a memory. A decompression circuit receives the values and performs a partial decompression. The partially compressed values are provided to a processor, which performs the final decompression. In one embodiment, a vector of N length compressed values are decompressed using a first bit mask into two N length sets having non-zero values. The two N length sets are further decompressed using two M length bit masks into M length sparse vectors, each having non-zero values.
