MULTI-RATE LDPC DECODING

• US 20130031438A1
• Filed: 01/04/2012
• Published: 01/31/2013
• Est. Priority Date: 07/29/2011
• Status: Active Grant

• Current Assignee: JP Morgan Chase Bank, N.A. (JP Morgan Chase & Co)
• Original Assignee: STEC Incorporated (Western Digital Corporation)
• Inventors: HU, Xinde, Dash, Dillip K., Jakab, Levente Peter, Komatineni, Rohit
• Granted Patent: US 8,566,667 B2
• Time in Patent Office: Days
• US Class Current: 714/752
• CPC Class Codes: G06F 11/1012 using codes or arrangements adapted for a specific type of error G06F11/1048 takes precedence
• CPC Class Codes: H03M 13/1111 Soft-decision decoding, e.g. by means of message passing or belief propagation algorithms
• CPC Class Codes: H03M 13/116 Quasi-cyclic LDPC [QC-LDPC] codes, i.e. the parity-check matrix being composed of permutation or circulant sub-matrices
• CPC Class Codes: H03M 13/6505 Memory efficient implementations
• CPC Class Codes: H03M 13/6516 Support of multiple code parameters, e.g. generalized Reed-Solomon decoder for a variety of generator polynomials or Galois fields

Abstract

The subject technology provides a decoding solution that supports multiple choices of code rates. A decoder may be configured to receive a selected code rate from a plurality of code rates. On the selection of the code rate, the decoder may determine a circulant size based on the code rate, and, on receiving the codeword, update, during one or more parity-check operations, a number of confidence values proportional to the circulant size in each of a plurality of memory units, each number of confidence values corresponding to a portion of the codeword.

Claims

1. A computer-implemented method for performing a parity-check operation in a variable rate decoder, the method comprising:

selecting one of a plurality of code rates;

receiving a codeword encoded at the selected code rate;

determining a circulant size based on the code rate; and

updating, during one or more parity check operations, a number of confidence values proportional to the circulant size in each of a plurality of memory units, each number of confidence values corresponding to a portion of the codeword.

2. The computer-implemented method of claim 1, the method further comprising:

determining a plurality of circulants based on the circulant size and the code rate;

on a first code rate being selected, for each circulant, storing confidence values corresponding to the circulant in a single memory unit;

on a second code rate being selected, for each circulant, merging two memory units to store confidence values corresponding to the circulant;

on a third code rate being selected, for a pair of circulants, splitting one memory unit into a first section and a second section to store confidence values corresponding to the pair of circulants, half of the confidence values being stored in the first section and another half of the confidence values being stored in the second section.

3. The computer-implemented method of claim 1, the method further comprising:

on a first code rate being selected, automatically mapping a data buffer corresponding to a number of first-sized circulants to a first set of memory units to facilitate updating a number of confidence values equal to a size of a respective first-sized circulant in each of the first set of memory units; and

second code rate being selected, automatically mapping a data buffer corresponding to a number of second-sized circulants to the first set of memory units to facilitate updating a number of confidence values corresponding to half of the size of a respective second-sized circulant in each memory unit.

4. The computer-implemented method of claim 3, wherein the data buffer comprises one data bit for each corresponding circulant, and

wherein the number of first-sized circulants equals a number of the first set of memory units, and

wherein the number of second-sized circulants is half the number of first-sized circulants.

5. The computer-implemented method of claim 3, the method further comprising:

on a third code rate being selected, automatically mapping a data buffer corresponding to a number of third-sized circulants to a second set of memory units to facilitate updating a number of confidence values equal to a size of a respective third-sized circulant in each of the second set of memory units,

wherein the number of third-sized circulants equals a number of the second set of memory units; and

wherein the number of third-sized circulants is less than a number of first-sized circulants.

6. The computer-implemented method of claim 5, the method further comprising:

on a fourth code rate mode being selected, automatically mapping a data buffer corresponding to a number of fourth-sized circulants to the second set of memory units to facilitate updating a first number of confidence values in a first section of each of the second set of memory units, and a second number of confidence values in a second section of each of the second set of memory units, the first number and second number of confidence values equal to a size of a respective fourth-sized circulant,

the number of fourth-sized circulants being equal to twice the number of third-sized circulants and twice the number of the second set of memory units.

7. A computer-implemented method for performing a parity-check operation in a variable rate decoder, the method comprising:

receiving a selected code rate;

receiving a codeword encoded at the selected code rate;

determining a number of circulants for use in one or more parity-check operations based on the code rate;

on receiving a first code rate, for each circulant, storing a number of log-likelihood ratios corresponding to the circulant in a single memory media, each log-likelihood ratio corresponding to a bit of the received codeword;

on receiving a second code rate, splitting one memory media into a first section and a second section to store the number of log-likelihood ratios corresponding to a first circulant and a second circulant, half of the log-likelihood ratios being stored in the first section and another half of the log-likelihood ratios being stored in the second section.

8. The computer-implemented method of claim 7, the method further comprising:

on receiving a third code rate, for each circulant, merging two memory media to store the number of log-likelihood ratios corresponding to the circulant, half of the log-likelihood ratios being stored in a first of the two memory media and another half of the log-likelihood ratios being stored in a second of the two memory media.

9. A low density parity check code decoding system, comprising:

a plurality of memory units;

one or more data buffers;

a plurality of memory mappers, each memory mapper configured to, on being activated, map one or more confidence values stored in one or more respective memory units to the one or more data buffers based on a corresponding active circulant matrix; and

a configuration memory, including instructions for receiving a selected one of a plurality of code rate modes, each code rate mode activating a number of memory mappers and determining how many confidence values are stored in each memory unit based on the selected one of the plurality of code rate modes.

10. The low density parity check code decoding system of claim 9, further comprising:

a circulant memory including storage of data for generating a parity-check matrix comprising one or more rows of circulants, a bit length of each row of circulants corresponding to a codeword length,

wherein each circulant comprises an identity matrix comprising a plurality of non-null values, each non-null value of the identity matrix providing a mapping between a memory unit mapped by an active memory mapper and a bit position in the one or more data buffers.

11. The low density parity check code decoding system of claim 10, wherein each bit position in the one or more data buffers corresponds to a current position of a non-null value located in each identity matrix of an active circulant row.

12. The low density parity check code decoding system of claim 9, wherein the one or more data buffers temporarily stores one or more confidence values mapped by each memory mapper.

13. The low density parity check code decoding system of claim 9, wherein each of the stored confidence values is a log-likelihood ratio corresponding to a bit of a received codeword.

14. The low density parity check code decoding system of claim 9, wherein each of the plurality of memory units provides storage for an equal number of confidence values.

15. The low density parity check code decoding system of claim 9, wherein, on a first code rate mode being selected, each active memory mapper facilitates storing a set of confidence values in one memory unit, a size of the set of confidence values corresponding to a size of the corresponding active circulant matrix.

16. The low density parity check code decoding system of claim 9, wherein, on a code rate mode being selected, each active memory mapper facilitates storing one set of confidence values in each of two of the plurality of memory units, the size of the one set corresponding to a size of half of the corresponding active circulant matrix.

17. The low density parity check code decoding system of claim 9, wherein, on a code rate mode being selected, each active memory mapper facilitates storing in a first section of a memory unit a first set of confidence values, and facilitates storing in a second section of the memory unit a second set of confidence values, the size of the first set and the size of the second set corresponding to a size of the corresponding active circulant matrix.

18. The low density parity check code decoding system of claim 9, further comprising:

wherein, on a first code rate mode being selected, a first set of memory units are mapped by a corresponding first set of memory mappers to facilitate updating a first set of confidence values stored in each of the first set of memory units;

wherein, on a second code rate mode being selected, the first set of memory units are mapped by a second set of memory mappers to facilitate updating a second set of confidence values, the second set of memory mappers being selected from half of the first set of memory mappers, each memory mapper in the second set of memory mappers mapping two memory units.

19. The low density parity check code decoder of claim 18, comprising:

wherein, on a third code rate mode being selected, a second set of memory units are mapped by a third set of memory mappers to facilitate updating a third set of confidence values, at least a portion of the third set of memory mappers being selected from the first set of memory mappers;

wherein, on a fourth code rate mode being selected, the third set of memory mappers map a first section of memory in the second set of memory units to facilitate updating a fourth set of confidence values, and map a second section of memory in the second set of memory units to facilitate updating a fifth set of confidence values, the fourth and fifth sets of confidence values being equal in size.

20. The low density parity check code decoder of claim 19, comprising:

a memory buffer operably connected to a port of each of the second set of memory units,

wherein the fourth set of confidence values are read into or written from the memory buffer during a first clock signal and the fifth set of confidence values are read into or written from the memory buffer on a second clock signal.