Data storage device and method for using multiple models for predicting a read threshold
A data storage device can generate a recommended read threshold value using a model that was trained under a plurality of conditions. However, such a model may result in an undesirable bit error rate or programming latency. If that should occur, the data storage device can use a different model trained under a condition similar to a current condition of the data storage device. In addition to avoiding an undesirable bit error rate or programming latency, this can result in improved throughput, improved quality of service, and reduced power consumption.
1 . A data storage device comprising:
a memory; and
one or more processors configured to:
use a model trained under a plurality of conditions to generate a read threshold value;
track a parameter resulting from using the read threshold value to read the memory; and
in response to a value of the parameter exceeding a threshold:
use a different model that was trained under a condition similar to a current condition of the data storage device to generate a new read threshold value, wherein the different model is selected from a plurality of different models for having a similarity score that is highest among the plurality of different models, each of the plurality of different models being trained under a different condition; and
use the new read threshold value to read the memory;
wherein the one or more processors comprises (i) a plurality of processors configured individually or in combination or (ii) a single processor.
2 . The data storage device of claim 1 , wherein the parameter comprises bit error rate.
3 . The data storage device of claim 1 , wherein the parameter comprises program latency, power consumption, failed bit count, a syndrome weight value, and/or an activation of a high decoding mode.
4 . The data storage device of claim 1 , wherein the plurality of different models is stored in the memory of the data storage device.
5 . The data storage device of claim 1 , wherein the plurality of different models is stored in a host memory buffer in a host.
6 . The data storage device of claim 1 , wherein the similarity score is based on read temperature.
7 . The data storage device of claim 1 , wherein the similarity score is based on a number of program/erase cycles.
8 . The data storage device of claim 1 , wherein the similarity score is based on data retention quantization.
9 . The data storage device of claim 1 , wherein the plurality of conditions comprises at least two of: a beginning-of-life condition, a middle-of-life condition, an end-of-life condition.
10 . The data storage device of claim 1 , wherein the one or more processors are further configured to use different sizes of models for different values of the parameter and/or increasing sizes of models according to a previous indication of a high bit error rate, a high failed bit count, and/or a program-erase cycle value.
11 . The data storage device of claim 1 , wherein the one or more processors are further configured to use another one of the plurality of different models until the value of the parameter does not exceed the threshold.
12 . The data storage device of claim 1 , wherein the one or more processors are further configured to use another one of the plurality of different models after an elapsed period of time.
13 . The data storage device of claim 1 , wherein the memory comprises a three-dimensional memory.
14 . In a data storage device comprising a memory, a method comprising:
using a model trained under a plurality of conditions to generate a read threshold value;
tracking a parameter resulting from using the read threshold value to read the memory, and
in response to a value of the parameter exceeding a threshold:
using a different model that was trained under a condition similar to a current condition of the data storage device to generate a new read threshold value, wherein the different model is selected from a plurality of different models for having a similarity score that is highest among the plurality of different models, each of the plurality of different models being trained under a different condition; and
using the new read threshold value to read the memory.
15 . The method of claim 14 , wherein the different models comprise a model trained under a middle-of-life condition.
16 . The method of claim 14 , wherein the different models comprise a model trained under an end-of-life condition.
17 . The method of claim 14 , wherein the different model is selected in a predefined order.
18 . The method of claim 14 , wherein the different models comprise a model trained under a beginning-of-life condition.
19 . The method of claim 17 , wherein the memory comprises a three-dimensional memory.
20 . A data storage device comprising:
a memory; and
means for:
using a model trained under a plurality of conditions to generate a read threshold value;
tracking a parameter resulting from using the read threshold value to read the memory, and
in response to a value of the parameter exceeding a threshold:
using a different model that was trained under a condition similar to a current condition of the data storage device to generate a new read threshold value, wherein the different model is selected from a plurality of different models for having a similarity score that is highest among the plurality of different models, each of the plurality of different models being trained under a different condition; and
using the new read threshold value to read the memory.