Read-detection in solid-state storage devices

• US 8,938,665 B2
• Filed: 02/01/2013
• Issued: 01/20/2015
• Est. Priority Date: 02/29/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Mittelholzer, Thomas, Papandreou, Nikolaos, Pozidis, Charalampos
• Primary Examiner(s): Merant, Guerrier
• Application Number:
  US13/756,894
• Publication Number:
  US 20130227380A1
• Time in Patent Office: 718 Days
• US Class Current: 714/807
• CPC Class Codes: G06F 11/1004 to protect a block of data words, e.g. CRC or checksum G06F11/1076 takes precedence; security arrangements for protecting computers or computer systems against unauthorized activity G06F21/00
• CPC Class Codes: G06F 11/1072 in multilevel memories
• CPC Class Codes: G11C 11/5642 Sensing or reading circuits; Data output circuits
• CPC Class Codes: G11C 16/26 Sensing or reading circuits; Data output circuits
• CPC Class Codes: G11C 7/1006 Data managing, e.g. manipulating data before writing or reading out, data bus switches or control circuits therefor

Abstract

A method for detecting codewords of a length-N, q^ary-symbol code, the symbols of each codeword stored in respective q-level cells of solid-state memory, where q>2, includes reading from memory cells storing a group of codewords to obtain respective read signals each comprising N signal components corresponding to respective symbols of a codeword. The signal components of the read signals are ordered according to signal level to produce an ordered component sequence. The ordered component sequence is partitioned to obtain segments corresponding to respective memory cell levels. Each segment contains a number of components dependent on predetermined frequency data indicative of expected frequency of occurrence of the corresponding level in use of the code. A reference signal level corresponding to each q memory cell level is determined in dependence on the signal components. The codeword corresponding to each read signal is then detected in dependence on the reference signal levels.

Claims

1. A method for detecting codewords of a length-N, q^ary-symbol code, the symbols of each codeword being stored in respective q-level cells of solid-state memory, where q>

2, the method comprising;

reading the memory cells storing a group of codewords to obtain respective read signals each comprising N signal components corresponding to respective symbols of a codeword;

ordering the signal components of the group of read signals according to signal level to produce an ordered component sequence;

partitioning the ordered component sequence to obtain segments corresponding to respective memory cell levels, each segment containing a number of components dependent on predetermined frequency data indicative of expected frequency of occurrence of the corresponding level in use of the code;

determining a reference signal level corresponding to each of the q memory cell levels in dependence on the signal components in at least a plurality of the segments; and

detecting the codeword corresponding to each read signal in dependence on the reference signal levels.

2. The method as claimed in claim 1, further comprising including partitioning the ordered component sequence into q the segments corresponding to respective memory cell levels.

3. The method as claimed in claim 2, further comprising determining the reference signal level for each of the q memory cell levels in dependence on the signal components in the corresponding segment.

4. The method as claimed in claim 1, further comprising:

determining the reference signal level for each of a plurality the q memory cell levels in dependence on the signal components in the corresponding segment; and

determining the reference signal level for each other memory cell level by interpolation from reference signal levels determined for the plurality of memory cell levels.

5. The method as claimed in claim 3, further comprising averaging the signal components in the corresponding segment to determine the reference signal level for the memory cell level corresponding to that segment.

6. The method as claimed in claim 3, further comprising selecting the median of the signal components in the corresponding segment to determine the reference signal level for the memory cell level corresponding to that segment.

7. The method as claimed in claim 1 wherein detecting the codeword corresponding to each read signal further comprises:

producing a codeword signal corresponding to each codeword of the code by replacing each symbol of the codeword by the reference signal level for the memory cell level corresponding to the value of that symbol; and

identifying the codeword signal which is closest to each read signal.

8. The method as claimed in claim 7, further comprising:

calculating statistical data for the distribution of the read signal components corresponding to each of the q memory cell levels from the read signals and the codeword signals identified therefor; and

detecting the codeword corresponding to each read signal in dependence on the statistical data.

9. The method as claimed in claim 2, further comprising:

calculating statistical data for the distribution of the read signal components corresponding to each of the q memory cell levels from the signal components in the segment corresponding to that memory cell level; and

detecting the codeword corresponding to each read signal in dependence on the statistical data.

10. The method as claimed in claim 8, further comprising detecting the codewords in dependence on the statistical data using one of a maximum-a-posteriori detection method and a maximum-likelihood detection method.

11. The method as claimed in claim 10, further comprising detecting the codewords by trellis decoding using the statistical data.

12. The method as claimed in claim 1 wherein each of the codewords of the code is a permutation of an N-symbol vector of a predetermined set of N-symbol vectors whose symbols are ordered according to symbol value, and wherein the step of detecting the codeword corresponding to each read signal includes:

ordering the components of each read signal according to signal level produce an ordered read signal;

producing a vector signal corresponding to each the vector by replacing each symbol of the vector by the reference signal level for the memory cell level corresponding to the value of that symbol; and

identifying the vector of which each codeword in the group is a permutation by determining which vector signal is closest to the ordered read signal for that codeword.

13. The method as claimed in claim 1 wherein the group of codewords are read substantially simultaneously.

14. The method as claimed in claim 1 wherein the group of codewords comprises codewords written substantially simultaneously to the solid state memory.

15. An apparatus for detecting codewords of a length-N, qary-symbol code, the symbols of each codeword being stored in respective q-level cells of solid-state memory, where q>

2, comprising;

a memory controller configured to read the memory cells storing a group of codewords to obtain respective read signals each comprising N signal components corresponding to respective symbols of a codeword, and a codeword detector comprising control logic configured to;

order the signal components of the group of read signals according to signal level to produce an ordered component sequence;

partition the ordered component sequence to obtain segments corresponding to respective memory cell levels, each segment containing a number of components dependent on predetermined frequency data indicative of expected frequency of occurrence of the corresponding level in use of the code;

determine a reference signal level corresponding to each of the q memory cell levels in dependence on the signal components in at least a plurality of the segments; and

detect the codeword corresponding to each read signal in dependence on the reference signal levels.

16. The apparatus as claimed in claim 15, wherein the control logic is further configured to partition the ordered component sequence into q the segments corresponding to respective memory cell levels.

17. The apparatus as claimed in claim 16, wherein the control logic is further configured to determine the reference signal level for each of the q memory cell levels in dependence on the signal components in the corresponding segment.

18. The apparatus as claimed in claim 15, wherein the control logic is further configured to:

determine the reference signal level for each of a plurality the q memory cell levels in dependence on the signal components in the corresponding segment; and

determine the reference signal level for each other memory cell level by interpolation from reference signal levels determined for the plurality of memory cell levels.

19. The apparatus as claimed in claim 17, wherein the control logic is further configured to average the signal components in the corresponding segment to determine the reference signal level for the memory cell level corresponding to that segment.

20. The apparatus as claimed in claim 17, wherein the control logic is further configured to select the median of the signal components in the corresponding segment to determine the reference signal level for the memory cell level corresponding to that segment.