Coded data generation or conversion – Digital code to digital code converters
Reexamination Certificate
2000-05-12
2001-11-20
Phan, Trong (Department: 2819)
Coded data generation or conversion
Digital code to digital code converters
C341S051000
Reexamination Certificate
active
06320521
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to the field of data compression systems, and more specifically to an improved method and format for organizing, storing and transmitting compressed data.
2. Discussion of the Prior Art
Data compression has become increasingly vital in today's computer systems due to the high demand for data transmission and storage capacity. In particular, main memory compression is now both feasible and desirable with the advent of parallel compression using a cooperative dictionary, as described in commonly-owned U.S. Pat. No. 5,729,228 to Franaszek et al. entitled PARALLEL COMPRESSION AND DECOMPRESSION USING A COOPERATIVE DICTIONARY, incorporated herein by reference. Parallel compression is a relatively new art in the field of compression. Its main concept is to divide a block of uncompressed data into multiple sectors and then assign them to individual engines for both compression and decompression with all engines sharing a cooperative dictionary such that the compression ratio is close to that of a single-engine design. This results in much better latency and throughput than the previous single-engine designs, thus making main memory compression feasible. It is the case however, that latency and throughput objectives may be better achieved provided there is implemented a highly-efficient compressed data format.
It would thus be highly desirable to provide a system and method for organizing compressed data efficiently, particularly, compressed data in parallel format, in order to enhance compression throughput and reduce decompression latency in data storage and data transmission systems.
SUMMARY OF THE INVENTION
It is an object of the invention to provide for a computer memory system implementing data compression, a highly-efficient compression output mechanism for organizing, storing and/or transmitting compressed data efficiently, particularly, by processing compressed data in parallel in order to enhance compression throughput and reduce decompression latency.
It is another object of the invention to provide a system and methodology for efficiently merging compressed data in a manner such that a compressor mechanism is enabled to write out parallel streams of compressed data from multiple compression engines quickly and with minimal loss of compression ratio.
Thus, according to the principles of the invention, there is provided a data compression system comprising: one or more compression engines executing in parallel for receiving uncompressed information words and compressing said information words into one or more compressed information words; a temporary storage queue associated with each compression engine for temporarily storing one or more compressed information words from its respective compression engine; a pre-fetch mechanism for pre-fetching an amount of compressed information words in parallel from each said temporary storage queue; a first processing device for receiving said pre-fetched compressed information words and successively forming a merged word in queue sequence order for output each processing cycle; and, a second processing device for receiving said pre-fetched compressed information words and generating a number of consecutive next available compressed words in queue sequence order when one or more temporary storage queues have exhausted its storage of compressed information words at any time during a processing cycle and, at such time, utilizing said number of consecutive next available compressed words to form said merged word output, whereby output of a complete merged word in each successive cycle is ensured.
Advantageously, the compressed data format enables a compressor to write out parallel streams of compressed data from multiple engines quickly and facilitates a decompression engine to extract the individual streams of compressed data easily without separate data delimiters and control. Thus, not only are the advantages of parallel processing with the compression and decompression engines exploited, but also the advantages of enhanced compression throughput and reduced decompression latency.
REFERENCES:
patent: 4049917 (1977-09-01), Copperi et al.
patent: 5729228 (1998-03-01), Franaszek et al.
Har David
Mak Kwok-Ken
Schulz Charles O.
International Business Machines - Corporation
Phan Trong
Scully Scott Murphy & Presser
Tassinari, Esq. Robert P.
LandOfFree
System and method for efficiently merging compressed data does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for efficiently merging compressed data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for efficiently merging compressed data will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2597616