Refresh, run, aggregate decoder recovery
A data processing system includes a likelihood input operable to receive encoded data, a decoder operable to apply a decoding algorithm to likelihood values for the received encoded data and to yield a decoded output, and a decoder input initialization circuit operable to generate new decoder input values based in part on the likelihood values for the received encoded data after the likelihood values for the received encoded data have failed to converge in the decoder.
1. A data processing system comprising:
a likelihood input operable to receive encoded data for a data recovery process;
a decoder operable to apply at least one decoding algorithm from a set of decoding algorithms to log likelihood ratio values for the encoded data and to yield at least one decoder output;
a decoder output aggregation circuit for aggregating, when applying the at least one decoding algorithm fails to converge, the at least one decoder output with one or more previous decoder outputs to generate an aggregate decoder output; and
a decoder input initialization circuit operable to generate new decoder input values based at least in part on the aggregated decoder output, the decoder applying an aggregation decoding algorithm to the new decoder input values, the set of decoding algorithms not including the aggregation decoding algorithm, the decoder terminating the data recovery process upon detecting convergence or upon satisfying an exit condition.
2. The data processing system of claim 1 , wherein the decoder output aggregation circuit is operable to average the collected decoder output from the plurality of decoder runs to compute the aggregate decoder output.
3. The data processing system of claim 1 , wherein the decoder output aggregation circuit is operable to compute the aggregate decoder output using simple majority selection of portions of the collected decoder output from the plurality of decoder runs.
4. The data processing system of claim 1 , wherein the decoder input initialization circuit is operable to generate the new decoder input values by log likelihood ratio magnitude replacement on the aggregate decoder output.
5. The data processing system of claim 1 , wherein the decoder input initialization circuit is operable to scale magnitudes of at least some of the new decoder input values for the encoded data by a scaling factor to generate the new decoder input values.
6. The data processing system of claim 1 , wherein generating the new decoder input values comprises copying sign values from the aggregated decoder output and applying a fixed value as magnitude values of the new decoder input values.
7. The data processing system of claim 6 , wherein at least one of the sign values is taken from the one or more previous decoder outputs.
8. The data processing system of claim 1 , wherein applying the at least one decoding algorithm from the set of decoding algorithms includes replacing magnitudes of at least some of the log likelihood ratio values for the encoded data with predetermined fixed values.
9. The data processing system of claim 1 , wherein satisfying the exit condition includes executing a set number of iterations or lapsing a time period, or both.
10. The data processing system of claim 1 , wherein satisfying the exit condition includes detecting when a difference between successive aggregated decoder outputs for two successive decoder runs is below a set threshold.
11. A method for a data recovery process, comprising:
receiving log likelihood ratio values for encoded data at a data input;
applying at least one decoding algorithm from a set of decoding algorithms to the log likelihood ratio values;
aggregating, when the at least one decoding algorithm fails to converge, a decoder output with one or more previous decoder outputs to generate an aggregated decoder output;
generating new decoder input values based at least in part on the aggregated decoder output;
applying an aggregation decoding algorithm to the new decoder input values, the set of decoding algorithms not including the aggregation decoding algorithm; and
terminating the data recovery process upon detecting convergence or upon satisfying an exit condition.
12. The method of claim 11 , wherein aggregating a decoder output with previous decoder outputs to generate the aggregated decoder output comprises averaging each of a plurality of previous decoder outputs.
13. The method of claim 11 , wherein aggregating a decoder output with previous decoder outputs to generate the aggregated decoder output comprises performing simple majority selection of portions of each of a plurality of previous decoder outputs.
14. The method of claim 11 , wherein generating the new decoder input values is based at least in part on log likelihood ratio magnitude replacement on the aggregate decoder output.
15. The method of claim 11 , wherein the decoder input initialization circuit is operable to scale magnitudes of at least some of the new decoder input values for the encoded data by a scaling factor to generate the new decoder input values.
16. The method of claim 11 , wherein generating the new decoder input values comprises copying sign values from the aggregated decoder output and applying a fixed value as magnitude values of the new decoder input values.
17. The method of claim 11 , wherein at least one of the sign values is taken from the one or more previous decoder outputs.
18. The method of claim 11 , wherein applying the at least one decoding algorithm from the set of decoding algorithms includes replacing magnitudes of at least some of the log likelihood ratio values for the encoded data with predetermined fixed values.
19. The method of claim 11 , wherein satisfying the exit condition includes at least one of executing a set number of iterations, lapsing a time period, detecting when a difference between successive aggregated decoder outputs for two successive decoder runs is below a set threshold, or any combination thereof.
20. A non-transitory computer-readable medium storing code for a data recovery process, the code comprising instructions executable by a processor to:
receive log likelihood ratio values for encoded data at a data input;
apply at least one decoding algorithm from a set of decoding algorithms to the log likelihood ratio values;
aggregate, when the at least one decoding algorithm failing to converge, a decoder output with one or more previous decoder outputs to generate an aggregated decoder output;
generate new decoder input values based at least in part on the aggregated decoder output;
apply an aggregation decoding algorithm to the new decoder input values, the set of decoding algorithms not including the aggregation decoding algorithm; and
terminate the data recovery process upon detecting convergence or upon satisfying an exit condition.