READ/WRITE OPERATIONS IN SOLID-STATE STORAGE DEVICES

• US 20130166994A1
• Filed: 12/17/2012
• Published: 06/27/2013
• Est. Priority Date: 12/21/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: POZIDIS, CHARALAMPOS, MITTELHOLZER, THOMAS, PAPANDREOU, NIKOLAOS
• Granted Patent: US 8,972,837 B2
• Time in Patent Office: Days
• US Class Current: 714/805
• CPC Class Codes: G06F 11/1032 Simple parity
• CPC Class Codes: G06F 11/1072 in multilevel memories

Abstract

Methods and apparatus are provided for reading and writing data in q-level cells of solid-state memory, where q>2. Input data is encoded into codewords having N q^ary symbols, wherein the symbols of each codeword satisfy a single-parity-check condition. Each symbol is written in a respective cell of the solid state memory by setting the cell to a level dependent on the q^ary value of the symbol. Memory cells are read to obtain read signals corresponding to respective codewords. The codewords corresponding to respective read signals are detected by relating the read signals to a predetermined set of N-symbol vectors of one of which each possible codeword is a permutation.

Claims

1. A method for reading and writing data in q-level cells of solid-state memory, where q>

2, the method comprising;

encoding input data into codewords having N q^ary symbols, wherein the symbols of each codeword satisfy a single-parity-check condition;

writing each symbol in a respective cell of the solid state memory by setting the cell to a level dependent on the q^ary value of the symbol;

reading memory cells to obtain read signals corresponding to respective codewords; and

detecting the codewords corresponding to respective read signals by relating the read signals to a predetermined set of N-symbol vectors of one of which each possible codeword is a permutation.

2. A method as claimed in claim 1 wherein the codewords are codewords of one of a length-N single-parity-check code or a coset thereof.

3. A method as claimed in claim 2 including:

mapping binary input data into q^ary data symbols; and

encoding groups of (N−

1) q^ary data symbols into respective N-symbol codewords.

4. A method as claimed in claim 2, the method including encoding input data into codewords of a limited codeword set which excludes length-N q^ary-symbol words whose symbols satisfy said single-parity-check condition but which are contained in the set of all permutations of at least one N-symbol vector having a smaller orbit under permutation than vectors in said predetermined set.

5. A method as claimed in claim 1 wherein:

the codeword symbols each comprise a plurality of sub-symbols corresponding to respective code levels; and

in each codeword, the N sub-symbols of at least one code level comply with one of a length-N single-parity-check code or a coset thereof.

6. A method as claimed in claim 5 wherein:

the codeword symbols each comprise a most-significant sub-symbol and a least-significant sub-symbol; and

in each codeword, the N least-significant sub-symbols comply with one of a length-N single parity check code or a coset thereof.

7. A method as claimed in claim 6 wherein, in each codeword, the N most-significant sub-symbols also comply with one of a length-N single parity check code or a coset thereof.

8. A method as claimed in claim 5 including:

mapping binary input data into sub-symbols of said code levels; and

for any code level whose sub-symbols comply with a parity check code or coset thereof, encoding groups of (N−

1) sub-symbols of that code level into the N sub-symbols of that level in respective codewords.

9. A method as claimed in claim 5, the method including encoding input data into codewords of a limited codeword set which excludes length-N q^ary-symbol words whose symbols satisfy said single-parity-check condition but which are contained in the set of all permutations of at least one N-symbol vector having a smaller orbit under permutation than vectors in said predetermined set.

10. A method as claimed in claim 1, wherein the codewords are codewords of a translation-stable code.

11. A method as claimed in claim 1 wherein each read signal comprises N signal components corresponding to respective symbols of a codeword, the method including detecting the codewords corresponding to respective read signals by:

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

relating the components of ordered read signals to symbols of said set of vectors.

12. A method as claimed in claim 11 including:

reading the memory cells storing a group of codewords to obtain read signals for the group;

averaging corresponding components of the ordered read signals for the group to produce an average read signal;

determining a reference signal level corresponding to each of the q levels of the memory cells in dependence on the average read signal and predefined probabilities of occurrence of each symbol value at each symbol position in a said codeword whose symbols are ordered according to symbol value; and

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

13. A method as claimed in claim 12 wherein said predetermined set comprises a plurality of vectors, the method including detecting the codeword corresponding to each read signal by:

producing a vector signal corresponding to each N-symbol vector by replacing each symbol of the vector by the reference signal level corresponding to the symbol value; and

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

14. A method as claimed in claim 13 including:

calculating statistical data for the distribution of the read signal component levels corresponding to each of the q levels of the memory cells from the ordered read signals and the vectors identified for the codewords corresponding thereto.

15. A method as claimed in claim 14 including detecting the codeword corresponding to each read signal by trellis decoding using said statistical data.

16. An apparatus for reading and writing data in q-level cells of solid-state memory, where q>

2, the apparatus comprising;

an encoder for encoding input data into codewords having N q^ary symbols, wherein the symbols of each codeword satisfy a single-parity-check condition;

a read/write controller for writing each symbol in a respective cell of the solid state memory by setting the cell to a level dependent on the q^ary value of the symbol, and for reading memory cells to obtain read signals corresponding to respective codewords; and

a detector for detecting the codewords corresponding to respective read signals by relating the read signals to a predetermined set of N-symbol vectors of one of which each possible codeword is a permutation.