Decoding optimization for channel mismatch
A memory system configured to decode a data set may pause a convergence process to update reliability metric values. The memory system may utilize a positive feedback system that updates the reliability metric values by analyzing current a posteriori reliability metric values to calculate average estimated reliability characteristic values associated with a memory error model. The updates to the reliability metric values may provide increased error correction capability and faster decoding.
1. A decoding method comprising:
starting, with a controller, a first portion of a convergence process, to decode a read data set using an initial set of reliability metric values generated based on a set of reliability bins and associating bits of the read data set with the set of reliability bins;
changing, with the controller, a configuration of the set of reliability bins thereby changing a total number of the reliability bins;
re-associating, with the controller, the bits of the read data set with the reliability bins according to the change in the configuration;
generating, with the controller, an updated set of reliability metric values according to the changed configuration of the set of reliability bins;
starting, with the controller, a second portion of the convergence process using the updated set of reliability metric values; and
completing, with the controller, the convergence process based on the updated set of reliability metric values.
2. The decoding method of claim 1 , wherein changing the total number of reliability bins comprises increasing the total number of reliability bins.
3. The decoding method of claim 1 , wherein changing the configuration of the set of reliability bins comprises changing, with the controller, the configuration of the set of reliability bins based on a change in reliability features used to determine the set of reliability bins.
4. The decoding method of claim 3 , wherein changing the configuration comprises increasing, with the controller, the total number of reliability bins according to an increase in a number of the reliability features used to determine the set of reliability bins.
5. The decoding method of claim 1 , wherein the convergence process is based on a memory error model and further including a step of updating the memory error model to an updated memory error model.
6. The decoding method of claim 1 , wherein the reliability metric values define magnitude components and sign components the sign components, correspond to hard bit representations of the read data set.
7. The decoding method of claim 6 , further including a step of calculating a set of estimated reliability characteristic values based on the magnitude components and sign components and wherein the set of estimated reliability characteristic values includes the reliability metric values, and wherein
the step of generating the updated set of reliability metric values is based on the set of estimated reliability characteristic values.
8. A decoding method comprising:
starting, with a controller, a first portion of a convergence process, to decode a read data set using an initial set of reliability metric values generated based on a set of reliability bins and associating bits of the read data set with the set of reliability bins, wherein the set of reliability bins is based on bits of the read data set being stored in good and bad storage locations;
changing, with the controller, a configuration of the set of reliability bins;
re-associating, with the controller, the bits of the read data set with the reliability bins according to the change in the configuration;
generating, with the controller, an updated set of reliability metric values according to the changed configuration of the set of reliability bins;
starting, with the controller, a second portion of the convergence process using the updated set of reliability metric values; and
completing, with the controller, the convergence process based on the updated set of reliability metric values.