Enhanced multilevel memory
View Patent ↗Subject matter disclosed herein relates to semiconductor memories and, more particularly, to multilevel non-volatile or volatile memories.
1. A method comprising:
storing states in one or more multilevel memory cells, said storing states comprising encoding said states according to a convolutional code to provide symbols representing said states; and
determining said encoded states using soft decision and convolutional decoding, wherein determining further comprises:
reading soft decisions associated with sensed threshold voltages from the multilevel memory cells;
maintaining sets of periodic functions, wherein each set is associated with a different plurality of program distributions of stored states, wherein the stored states associated with a particular periodic function are periodic with respect to one another;
associating a soft decision with a particular set of the periodic functions; and
selecting a program distribution from the particular set of periodic functions based at least partly on a calculation of a most probable encoding path that had been followed for encoding.
2. The method of claim 1 , wherein said determining said encoded states further comprises:
retrieving said stored states from said stored values and applying said soft decision and convolutional decoding to recover said stored states.
3. The method of claim 1 , wherein said soft decision comprises determining a most probable program state among a plurality of program states to determine said encoded states.
4. The method of claim 1 , wherein said multilevel memory cells are capable of storing two or more bits.
5. The method of claim 1 , wherein said values comprise a measurement of a probability that a sensed threshold voltage belongs to a particular program distribution of said one or more multilevel memory cells.
6. A memory device comprising:
a memory array to store states across one or more multilevel memory cells; and
a controller configured to:
encode said states according to a convolutional code to provide symbols representing said states;
read soft decisions associated with sensed threshold voltages from the multilevel memory cells;
access sets of periodic functions, wherein each set is associated with a different plurality of program distributions of stored states, wherein the stored states associated with a particular periodic function are periodic with respect to one another;
associate a soft decision with a particular set of the periodic functions; and
select a program distribution from the particular set of periodic functions based at least partly on a calculation of a most probable encoding path that had been followed for encoding to determine said encoded states using soft decision and convolutional decoding.
7. The memory device of claim 6 , wherein said soft decision comprises determining a most probable program state among a plurality of program states to determine said encoded states.
8. The memory device of claim 6 , further comprising:
a sensing circuit to determine said threshold voltages of said one or more multilevel memory cells; and
a converter to convert said values to a digital sequence of numbers indicating said threshold voltages of said one or more multilevel memory cells.
9. The memory device of claim 8 , wherein said sensing circuit comprises a voltage comparator having multiple inputs to receive a bit line voltage corresponding to one of said one or more multilevel memory cells and a reference voltage.
10. The memory device of claim 6 , wherein said multilevel memory cells are capable of storing two or more bits.
11. The memory device of claim 6 , further comprising:
a decoding unit to generate said values to quantify a probability that a sensed threshold voltage belongs to a particular program distribution of said one or more multilevel memory cells.
12. A system comprising:
a memory device comprising:
a memory array configured to store states across one or more multilevel memory cells, and
a controller configured to:
encode said states according to a convolutional code to provide symbols representing said states;
read soft decisions associated with sensed threshold voltages from the multilevel memory cells;
access sets of periodic functions, wherein each set is associated with a different plurality of program distributions of stored states, wherein the stored states associated with a particular periodic function are periodic with respect to one another;
associate a soft decision with a particular set of the periodic functions; and
select a program distribution from the particular set of periodic functions based at least partly on a calculation of a most probable encoding path that had been followed for encoding to determine said stored states using soft decision and convolutional decoding; and
a processor to host one or more applications and to initiate write and/or read commands to said memory device controller to provide access to said memory array.
13. The system of claim 12 , wherein said soft decision comprises determining a most probable program state among a plurality of program states to determine said stored states.
14. The system of claim 12 , further comprising:
a sensing circuit to determine said threshold voltages of said one or more multilevel memory cells; and
a converter to convert said values to a digital sequence of numbers indicating said threshold voltages of said one or more multilevel memory cells.
15. The system of claim 14 , wherein said sensing circuit comprises a voltage comparator having multiple inputs to receive a bit line voltage corresponding to one of said one or more multilevel memory cells and a reference voltage.
16. The system of claim 12 , wherein said convolutional encoding comprises Viterbi encoding.
17. The system of claim 12 , further comprising:
a decoding unit to generate said values to quantify a probability that a sensed threshold voltage belongs to a particular program distribution of said one or more multilevel memory cells.
18. The method of claim 1 , further comprising storing said values as multi-bit strings in said one or more multilevel memory cells.
19. The method of claim 18 , wherein said storing said values as multibit strings in said one or more multilevel memory cells further comprises storing said values as multi-bit strings in two or more of said multilevel memory cells.
20. The memory device of claim 6 , wherein said controller is further adapted to store said values as multi-bit strings in said one or more multilevel memory cells.
21. The method of claim 1 , wherein the sets of periodic functions are maintained in a lookup table.
22. The memory device of claim 6 , further comprising a look-up table configured to store the sets of periodic functions.
23. The system of claim 12 , further comprising a look-up table configured to store the sets of periodic functions.